From zaidee rndtm.net.my Wed Jun 1 10:20:31 2005 From: zaidee rndtm.net.my (Ahmad Zaidee bin Abu) Date: Wed Jun 1 02:21:50 2005 Subject: [Mp4-tech] A doubt in calculation of PSNR value for Video Message-ID: <558E65614165F242AC17B5D316BFAB0F105629@mstmrndmx01.tmrnd.net.my> Dear Alexis. Thanks for the explanation, Silly me, I've never thought about the black and white. Regards, Zaidee. ________________________________ From: mp4-tech-bounces@lists.mpegif.org on behalf of Alexis Michael Tourapis Sent: Wed 6/1/2024 2:36 AM To: Ahmad Zaidee bin Abu; M Mahdi Ghandi; Tuukka Toivonen Cc: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] A doubt in calculation of PSNR value for Video Dear Ahmad, Your claim is not true. One may for example try to encode a completely black or a completely white image (i.e. consider fades to/from black and white). Other cases could also lead to 0 MSE => infinite PSNR. As Tuukka suggested, one way to avoid problems with the consideration of infinite frame PSNR is to consider the average MSE instead. Other solutions would be to penalize the MSE conditionally (as is done also in the JM) or not (i.e. by always adding a given penalty to the MSE or SSE) in an attempt to avoid infinite values. Obviouslyyou would need to use the same method when comparing two different methods. In general though you would need to be a bit cautious with the consideration of PSNR (or even MSE) as a measure of quality. It is quite possible for example that one may encode a sequence with algorithm/system A and have higher average PSNR than a scheme B, however perceptually B may look better due to fewer temporal or otherwise artifacts imposed by different characteristics of the encoding (normative and/or non normative). Frames that could lead to such "infinite" PSNRs could themselves also artificially spike your average numbers and essentially mask som! e "bad" PSNR cases that may appear within your sequence (use of average MSE for PSNR computation is more reliable in such cases, i.e. if you wish to experiment, assume you have 3 frames one with MSE equal to 1, a second with MSE equal to 100, and a third with MSE equal to 10. However, note also that averages do not take order of frames in account which could also affect subjective quality). Best regards, Alexis B -------------- Original message -------------- I believe that there is no way that MSE going to be 0. Unless if the compressed frame is identical to original image. Then why do we care calculate PSNR if it is totally the same thing? Am i correct ? ________________________________ From: mp4-tech-bounces@lists.mpegif.org on behalf of M Mahdi Ghandi Sent: Thu 5/26/2005 1:12 AM To: 'Tuukka Toivonen' Cc: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] A doubt in calculation of PSNR value for Video The main problem is when for example this happened for only one frame in a sequence. I will not include that infinite in average PSNR calculation or alternatively replace it with an assumed upper limit (e.g. 50 dB). -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Tuukka Toivonen Sent: 25 May 2024 08:31 To: T. Bheemarjuna Reddy Cc: mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] A doubt in calculation of PSNR value for Video On Tue, 24 May 2005, T. Bheemarjuna Reddy wrote: > I have one basic doubt in calculating PSNR value. > > PSNR (db) = 10 log {(2^n - 1)^2 / MSE} > > what happens to value of PSNR if MSE = 0 ? The PSNR is infinite then. You could check specifically for this case and return the lar! gest possible number. If using IEEE floating points, they can represent infinite exactly. However, normally this is very rare case so people often ignore it. _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php The information contained in this Internet message is confidential and intended only for the use of the individual or entity identified. If the reader of this message is not the intended recipient, any dissemination, distribution or copying of the information contained in this Internet message is strictly prohibited. If you received this message in error, please notify the sender immediately. The information contained in this Internet message is confidential and intended only for the use of the individual or entity identified. If the reader of this message is not the intended recipient, any dissemination, distribution or copying of the information contained in this Internet message is strictly prohibited. If you received this message in error, please notify the sender immediately. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/fe403416/attachment-0001.html From ying.s.zhang intel.com Wed Jun 1 10:33:13 2005 From: ying.s.zhang intel.com (Zhang, Ying S) Date: Wed Jun 1 02:27:55 2005 Subject: [Mp4-tech] convert ADTS format stream to mp4 Message-ID: <571ACEFD467F7749BC50E0A98C17CDD806E59618@pdsmsx403> Hi all, How to convert ADTS format stream to mp4 file format? Is there any available tool? Thanks! Best Regards, Zhang Ying iNet: 8-752-1572 Tel: +86-21-52574545-1572 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/1e52c4fc/attachment.html From jiazhaox mobilesoft.com.cn Wed Jun 1 11:28:36 2005 From: jiazhaox mobilesoft.com.cn (JazryXu) Date: Wed Jun 1 02:30:59 2005 Subject: [Mp4-tech][Audio] AAC File Format References: <1117440753.429acaf13a907@imapwww.epfl.ch> <429C02D5.8000809@neomagic.com> Message-ID: <006b01c56651$9d2a0cd0$7ca1a8c0@JazryXu> Hi, Expert: Does anyone tell me where can I get the AAC file format freely . Thanks a lot. Jazry From ugarg neomagic.com Wed Jun 1 12:07:15 2005 From: ugarg neomagic.com (Umang Garg) Date: Wed Jun 1 02:31:05 2005 Subject: [Mp4-tech][General]: Access to "http://mpeg.nist.gov/" Message-ID: <429D498B.6050409@neomagic.com> Dear List Members, It will be very helpful if someone could point out, which meeting's password is currently being used to access the latest developments on MPEG available at http://mpeg.nist.gov/ Best Regards, Umang Garg NeoMagic Corporation, based in Santa Clara, California, enables new generations of handheld systems with its Applications Processors that are designed to offer the lowest power, smallest form-factor and best multimedia features and performance. The company is a pioneer in the integration of complex logic, memory and analog circuits into single-chip solutions. Information on the company may be found on the World Wide Web at www.neomagic.com. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/fede5d32/attachment.html From ugarg neomagic.com Wed Jun 1 13:14:11 2005 From: ugarg neomagic.com (Umang Garg) Date: Wed Jun 1 05:26:43 2005 Subject: [Mp4-tech][Audio][Software]: Access to weekly snapshots of Reference Software Message-ID: <429D593B.2050207@neomagic.com> Dear List Members, I am out-of-touch with the developements on the MPEG-4 Audio for a year. It will be very helpful if someone could point out, which meeting's password is currently being used to access the latest development of MPEG-4 Audio Reference Software available at ftp://mpeg4vm@ftp.tnt.uni-hannover.de/incoming/cvs/ Best Regards, Umang Garg NeoMagic Corporation, based in Santa Clara, California, enables new generations of handheld systems with its Applications Processors that are designed to offer the lowest power, smallest form-factor and best multimedia features and performance. The company is a pioneer in the integration of complex logic, memory and analog circuits into single-chip solutions. Information on the company may be found on the World Wide Web at www.neomagic.com. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/7d8fe35c/attachment.html From rkoenen intertrust.com Wed Jun 1 03:20:42 2005 From: rkoenen intertrust.com (Rob Koenen) Date: Wed Jun 1 05:29:59 2005 Subject: [Mp4-tech][General]: Access to "http://mpeg.nist.gov/" Message-ID: You should acquire this password from your MPEG "Head of Delegation". It is not for public distribution, and please note that MPEG and MPEGIF are different organization. Let me know in a personal email email if you need help finding your HoD. Best, Rob ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Umang Garg Sent: Wednesday, June 01, 2024 07:37 To: MPEG Forum Cc: Umang Garg Subject: [Mp4-tech][General]: Access to "http://mpeg.nist.gov/" Dear List Members, It will be very helpful if someone could point out, which meeting's password is currently being used to access the latest developments on MPEG available at http://mpeg.nist.gov/ Best Regards, Umang Garg NeoMagic Corporation, based in Santa Clara, California, enables new generations of handheld systems with its Applications Processors that are designed to offer the lowest power, smallest form-factor and best multimedia features and performance. The company is a pioneer in the integration of complex logic, memory and analog circuits into single-chip solutions. Information on the company may be found on the World Wide Web at www.neomagic.com. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/3104a2cb/attachment.html From rkoenen intertrust.com Wed Jun 1 03:24:07 2005 From: rkoenen intertrust.com (Rob Koenen) Date: Wed Jun 1 06:10:32 2005 Subject: [Mp4-tech][Audio][Software]: Access to weekly snapshots of ReferenceSoftware Message-ID: Again, this is something that you should find through MPEG, not through the MPEG Industry Forum. (And again, let me know in a personal email if you need help finding the rightMPEG contacts.) Kind Regards, Rob ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Umang Garg Sent: Wednesday, June 01, 2024 08:44 To: mp4-tech@lists.mpegif.org Cc: Umang Garg Subject: [Mp4-tech][Audio][Software]: Access to weekly snapshots of ReferenceSoftware Dear List Members, I am out-of-touch with the developements on the MPEG-4 Audio for a year. It will be very helpful if someone could point out, which meeting's password is currently being used to access the latest development of MPEG-4 Audio Reference Software available at ftp://mpeg4vm@ftp.tnt.uni-hannover.de/incoming/cvs/ Best Regards, Umang Garg NeoMagic Corporation, based in Santa Clara, California, enables new generations of handheld systems with its Applications Processors that are designed to offer the lowest power, smallest form-factor and best multimedia features and performance. The company is a pioneer in the integration of complex logic, memory and analog circuits into single-chip solutions. Information on the company may be found on the World Wide Web at www.neomagic.com. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/c45ce22a/attachment.html From gripened gmail.com Wed Jun 1 20:45:34 2005 From: gripened gmail.com (Jayant Chauhan) Date: Wed Jun 1 07:50:32 2005 Subject: [Mp4-tech] Standalone H264 player Message-ID: <5b996acb05060103456b0426bf@mail.gmail.com> Hey fellas Is there any standalone player for H264 files?! with regards Jayant From rohan teneoris.com Wed Jun 1 18:28:22 2005 From: rohan teneoris.com (Rohan Kangralkar) Date: Wed Jun 1 08:29:13 2005 Subject: [Mp4-tech] Conformance testing of MPEG4 Advanced Simple profile (Video) Message-ID: <1117627102.3372.30.camel@rohan.teneoris.com> Hello, I'm looking for Conformance tests & bitstreams for MPEG-4 video for Advanced Simple profile. In this process, I purchased ISO Conformance document (ISO/IEC 14496-4) & the test-streams. 1. I found 13 bitstreams under the directory 'advanced_simple'. But the Conformance document does not have any description about these bitstreams. No Table given in the document has any mention on 'Advanced Simple profile'. 2. I came across a link : 'MPEG-4 Video Conformance Bitstreams available for download' in MPEG4 Industry forum (m4if.org). Under this link also, I found some bit-streams under the directory 'Advanced_Simple'. These streams are different from the ones which are distributed from ISO. Also, these appear to be newer than the ones from ISO (as evident from the timestamps of the files). 3. Eventhough ISO/IEC 14496-4 document does not seem to categorize any bitstreams under Advanced Simple profile, there are some bit-streams listed under Section 5.6 - 'Additional Conformance Testing' which have the features of Advanced Simple profile like Quarter-sample interpolation & GMC. In view of the above, I'm not quite clear about which of the sources I should be using (1, 2 or 3 mentioned above) for testing the features of my Advanced Simple profile decoder & Conformance. I'd appreciate if some one clarifies this. Regards, -- Rohan Kangralkar From singer apple.com Wed Jun 1 11:40:44 2005 From: singer apple.com (Dave Singer) Date: Thu Jun 2 08:24:15 2005 Subject: [Mp4-tech][Audio] AAC File Format In-Reply-To: <006b01c56651$9d2a0cd0$7ca1a8c0@JazryXu> References: <1117440753.429acaf13a907@imapwww.epfl.ch> <429C02D5.8000809@neomagic.com> <006b01c56651$9d2a0cd0$7ca1a8c0@JazryXu> Message-ID: At 10:28 +0800 1/06/05, JazryXu wrote: >Hi, Expert: > Does anyone tell me where can I get the AAC file format freely . It's not clear what you mean by the AAC file format. AAC streams are commonly found in * MP4 ('ISO') files, 14496-12 * ADTS format * ADIF format and, I think, packed into MPEG-2 environments. So, which do you need, and do you need the spec. or reference code, or example files, or...? > >Thanks a lot. > >Jazry > >_______________________________________________ >NOTE: Please use clear subject lines for your posts. Include [audio, >[video], [systems], [general] or another apppropriate identifier to >indicate the type of question you have. > >Note: Conduct on the mailing list is subject to the Antitrust >guidelines found at >http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -- David Singer Apple Computer/QuickTime From seblopez iuma.ulpgc.es Wed Jun 1 21:01:47 2005 From: seblopez iuma.ulpgc.es (Sebastian Lopez Suarez) Date: Thu Jun 2 08:30:21 2005 Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis Message-ID: Hi, Does anybody knows how to manage with the configuration file in order to avoid intra macroblocks in P frames? Thanks a lot in advance. Best! From rob.koenen mpegif.org Thu Jun 2 01:08:13 2005 From: rob.koenen mpegif.org (Rob Koenen (MPEGIF)) Date: Thu Jun 2 08:30:26 2005 Subject: FW: [Mp4-tech] Standalone H264 player Message-ID: <20050601220815.12F48191E0@olive.qinip.net> Forwarding auto-discarded message from non-member email address. Rob -----Original Message----- From: Ben Waggoner [mailto:ben@benwaggoner.com] Sent: Wednesday, June 01, 2024 19:32 To: Jayant Chauhan Cc: mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] Standalone H264 player Jayant, QuickTime 7 on Mac will play H.264 in .mp4 format. Windows version forthcoming VideoLAN Client (VLC) has some H.264 support in its current 0.8.1 version, although it's much improved in the version in CVS. It supports the codec in a variety of wrappers, including .mp4. There's a Moonlight player for Windows as well, but I haven't spent much time with it. Ben Waggoner, Interframe Media http://www.interframemedia.com Technology and Strategy for Digital Media My Book: http://www.cmpbooks.com/compression/ My Tutorials http://www.classondemand.net/benwaggoner/ My Stanford Class http://www.digitalmediaacademy.org/compression.html On Jun 1, 2005, at 3:45 AM, Jayant Chauhan wrote: > Is there any standalone player for H264 files?! > From harshahm tataelxsi.co.in Thu Jun 2 10:26:04 2005 From: harshahm tataelxsi.co.in (Harsha H M) Date: Thu Jun 2 08:30:31 2005 Subject: [Mp4-tech] Standalone H264 player In-Reply-To: <5b996acb05060103456b0426bf@mail.gmail.com> Message-ID: <000001c56727$004d7970$7722320a@telxsi.com> Refer to this link, u will get 21 days evaluation ver of H264 dec. http://moonlight.co.il/products/consumer/oneclick/ -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of Jayant Chauhan Sent: Wednesday, June 01, 2024 4:16 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Standalone H264 player Hey fellas Is there any standalone player for H264 files?! with regards Jayant _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From pankaj0iit yahoo.co.in Wed Jun 1 22:46:10 2005 From: pankaj0iit yahoo.co.in (pankaj bansal) Date: Thu Jun 2 08:30:36 2005 Subject: [Mp4-tech] doubt in H.264 ? Message-ID: <20050602044610.43104.qmail@web8402.mail.in.yahoo.com> Hi , Can anybody clear the difference in the following things in H.264? complementary field pair non-paired field paired field frames thanks & regards pankaj --------------------------------- Discover Yahoo! Get on-the-go sports scores, stock quotes, news & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050601/c9b82011/attachment.html From seblopez iuma.ulpgc.es Thu Jun 2 09:18:59 2005 From: seblopez iuma.ulpgc.es (Sebastian Lopez Suarez) Date: Thu Jun 2 08:30:41 2005 Subject: [Mp4-tech] [H.264]Mode decision by frame Message-ID: Hi, Does anyone knows how to manage with the configuration file in order to avoid the use of intra macroblocks in P frames? I was triying with the IntraDisableInterOnly but it does not seems to work .... Thanks a lot in advance. From samuel dc.fi.udc.es Thu Jun 2 11:13:43 2005 From: samuel dc.fi.udc.es (Samuel Rivas) Date: Thu Jun 2 08:30:46 2005 Subject: [Mp4-tech] MPEG-2 TS converter In-Reply-To: <84ac013050526032723a69ced@mail.gmail.com> References: <200505260142.j4Q1gkYA051353@smtp0.dreamwiz.com> <84ac013050526032723a69ced@mail.gmail.com> Message-ID: <20050602081343.GA7236@crusher.lfcia.pri> Alejandro Moya wrote: > Dear Ik.Hwan Cho, > converting MPEG-2 Program stream into one Transport Stream could notbe difficult if you just to mux one program into the TS without timinginformation: you should remove Pack and System Headers from ProgramStream and should divide the resulting Packets into 184 byte chunksadding an easy 4 byte header -> 0x47 0x0Y 0xYY 0x10 where "Y" are 13bits indicating your Program IDentification (you may choose it as youdesire). The last step would be to add two packets at the beginning ofthe file containing your new TS: one with Program Association Table(Tranport Stream which should have PID = 0) and Program Map Table(which will have the PID you defined in the Program Association Table)containing the PIDs of the transport stream packets carrying youroriginal video (i.e., the PID "Y" you choose above). It is is not so easy. TS packets need stuffing if payload is less than 184 bytes length, you must set the continuity counters and the payload_unit_start indicators, PAT and PMT should be inserted periodically, etc. An easy way to get a more or less correct transport stream is using Video Lan Client (http://www.videolan.org/vlc/) -- Samuel From vikas venera.co.in Thu Jun 2 17:33:05 2005 From: vikas venera.co.in (Vikas Singhal) Date: Thu Jun 2 08:30:50 2005 Subject: [Mp4-tech] AVC in transport stream References: <429D593B.2050207@neomagic.com> Message-ID: <005801c56762$a8bc8430$680ba8c0@MAYA> Hi, Can anyone point me to the location where I can download transport streams having AVC as video payload format. Thanks in Advance Vikas -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050602/cb5e5cea/attachment.html From andrewk vbrick.com Thu Jun 2 09:36:17 2005 From: andrewk vbrick.com (Andrew Krupiczka) Date: Thu Jun 2 17:39:04 2005 Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis Message-ID: Sebastian, You can try IntraDisableInterOnly = 1 # Disable Intra modes for Non I-Slices but based on my past brief testing I'm not sure it's working as it was intended to :-( Best regards, Andrew -----Original Message----- From: Sebastian Lopez Suarez [mailto:seblopez@iuma.ulpgc.es] Sent: Wednesday, June 01, 2024 3:02 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis Hi, Does anybody knows how to manage with the configuration file in order to avoid intra macroblocks in P frames? Thanks a lot in advance. Best! _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From kexu ee.cuhk.edu.hk Thu Jun 2 23:17:24 2005 From: kexu ee.cuhk.edu.hk (Xu Ke) Date: Thu Jun 2 17:39:09 2005 Subject: [Mp4-tech][Video][H.264]Doubts on coded_block_pattern table Message-ID: <200506021417.j52EHKtL010070@cuees8.ee.cuhk.edu.hk> Dear experts, Currently I have some doubts on Table9-4 "Assignment of codeNum to values of coded_block_pattern for macroblock prediction modes".This table is so irregular that I can't find any rule to faciliate my hardware decoding.In addition, why we need two seperate column for "Intra_4x4" and "Inter"? Combine them into single one maybe easier for both software/hardware decoding. Would anybody be kind enough to explain it for me? Best regards, XU Ke From almomo1 gmail.com Thu Jun 2 17:27:05 2005 From: almomo1 gmail.com (Alejandro Moya) Date: Thu Jun 2 17:39:14 2005 Subject: [Mp4-tech] MPEG-2 TS converter In-Reply-To: <20050602081343.GA7236@crusher.lfcia.pri> References: <200505260142.j4Q1gkYA051353@smtp0.dreamwiz.com> <84ac013050526032723a69ced@mail.gmail.com> <20050602081343.GA7236@crusher.lfcia.pri> Message-ID: <84ac01305060207274b79c196@mail.gmail.com> You needn't insert PMT or PAT periodically because what our friend Ik-Hwan asked was the way to convert one program stream into a TS for testing pourposes, not for streaming: when you broadacast a TS, even if it has only one program, you need to insert PMT or PAT so decoders can be able to get the program list and their associated PIDs; thatis not needed when using a TS inside a file. "continuity counters " aren't needed either and "payload_unit_start" bit is set in the byte sequence I told him to insert as transport stream header. The only thing I forgot to comment to make an easy test TS was that each transport packet should be 188 bytes length, so PMT and PAT packets should be padded with byte 0xFF. Regards. 2005/6/2, Samuel Rivas : > Alejandro Moya wrote: > > Dear Ik.Hwan Cho, > > converting MPEG-2 Program stream into one Transport Stream could notbe difficult if you just to mux one program into the TS without timinginformation: you should remove Pack and System Headers from ProgramStream and should divide the resulting Packets into 184 byte chunksadding an easy 4 byte header -> 0x47 0x0Y 0xYY 0x10 where "Y" are 13bits indicating your Program IDentification (you may choose it as youdesire). The last step would be to add two packets at the beginning ofthe file containing your new TS: one with Program Association Table(Tranport Stream which should have PID = 0) and Program Map Table(which will have the PID you defined in the Program Association Table)containing the PIDs of the transport stream packets carrying youroriginal video (i.e., the PID "Y" you choose above). > > It is is not so easy. TS packets need stuffing if payload is less > than 184 bytes length, you must set the continuity counters and the > payload_unit_start indicators, PAT and PMT should be inserted > periodically, etc. > > An easy way to get a more or less correct transport stream is using Video > Lan Client (http://www.videolan.org/vlc/) > > -- > Samuel > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > From Danijel.Domazet zg.htnet.hr Thu Jun 2 17:54:40 2005 From: Danijel.Domazet zg.htnet.hr (D.Domazet) Date: Thu Jun 2 17:39:19 2005 Subject: [Mp4-tech] [Audio] Master Frequency Band Table generation in HE AAC? Message-ID: <000e01c56783$0167e550$4800a8c0@FERGUSON> Hi all, I am looking at Figure 4.38 and Figure 4.39 pages 48 and 49 in ISO 14496-3:2001/Amd.1:2003 which is in fact HE AAC. >From this figures I am trying to understand the logic behind Master Frequency Band Table generation. What I see is a flowchart with bunch of formulas, more or less simple math, but there is very little explanation. Could someone (maybe from CodingTechnologies) briefly try to explain where do these formulas come from? Thanks very much, Daniel -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050602/70fa5db1/attachment.html From seblopez iuma.ulpgc.es Thu Jun 2 17:00:29 2005 From: seblopez iuma.ulpgc.es (Sebastian Lopez Suarez) Date: Thu Jun 2 17:39:24 2005 Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis In-Reply-To: References: Message-ID: I have experienced the same as you. Even when I define that each frame is composed by only one slice and I fix that IntraDisableInterOnly = 1 the the results indicate that some macroblocks are being coded as intra in P frames. ??????? Best! On Thu, 2 Jun 2005, Andrew Krupiczka wrote: > > Sebastian, > > You can try > > IntraDisableInterOnly = 1 # Disable Intra modes for Non I-Slices > > but based on my past brief testing I'm not sure it's working as it was > intended to :-( > > Best regards, > > Andrew > > > > > > -----Original Message----- > From: Sebastian Lopez Suarez [mailto:seblopez@iuma.ulpgc.es] > Sent: Wednesday, June 01, 2024 3:02 PM > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis > > Hi, > > Does anybody knows how to manage with the configuration file in order > to avoid intra macroblocks in P frames? Thanks a lot in advance. > > Best! > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to > indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > From vsathe multiratesystems.com Thu Jun 2 09:22:00 2005 From: vsathe multiratesystems.com (Vinay Sathe) Date: Thu Jun 2 17:39:28 2005 Subject: [Mp4-tech] MPEG-2 TS converter In-Reply-To: <20050602081343.GA7236@crusher.lfcia.pri> Message-ID: <200506021522.j52FM8sS011079@lists1.magma.ca> Hello Samuel, Why do you say "more or less correct"? In your experience, is videolan "less correct" in any aspect of creating a transport stream? Regards, Vinay -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Samuel Rivas Sent: Thursday, June 02, 2024 1:14 AM To: mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] MPEG-2 TS converter It is is not so easy. TS packets need stuffing if payload is less than 184 bytes length, you must set the continuity counters and the payload_unit_start indicators, PAT and PMT should be inserted periodically, etc. An easy way to get a more or less correct transport stream is using Video Lan Client (http://www.videolan.org/vlc/) -- Samuel _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From tomar_atul2000 yahoo.com Thu Jun 2 09:37:17 2005 From: tomar_atul2000 yahoo.com (atul tomar) Date: Thu Jun 2 17:39:33 2005 Subject: [Mp4-tech] frame_num, PreRefFrameNum in h.264.. Message-ID: <20050602153717.48316.qmail@web52301.mail.yahoo.com> Hi All, I am new to h.264 standard but working very hard to understand the ITU standard. I would appreciate if someone could clarify my doubts about how to determine frame_num, PrevRefFrameNum in Baseline profile. Let us say for I,P,P,P,P,I,P,P,P,P,I video sequence. I tried to read frame_num sementics in 7.4.3 but could not get it. Standard says that for 'I' IDR frame in my sequence these variable (frame_num,PreRefFrameNum) will be given values by (0,0). Now for next frame i.e. 'P' Non-IDR frame i am not able to get these two values. Please correct me, whether everytime 'I' frame of sequence will be having (0,0) values. If everytime 'I' frame_num is zero then how to assign MaxFrameNum. Looking for your response. Thanks, Atul --------------------------------- Discover Yahoo! Find restaurants, movies, travel & more fun for the weekend. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050602/658a5a93/attachment.html From tomar_atul2000 yahoo.com Thu Jun 2 09:39:40 2005 From: tomar_atul2000 yahoo.com (atul tomar) Date: Thu Jun 2 17:39:37 2005 Subject: [Mp4-tech] TopField and BottomField in Baseline proffile and POC Message-ID: <20050602153940.79745.qmail@web52305.mail.yahoo.com> Hi, As we know that we dont use interlacing in baseline proffile so Is it necessary to determine PictureOrderCount. I knew through standard that it is also used for decoder conformance checking but how Can we get TopField and BottomField field related variables if we are not dividing frame in fields then how can we work with above two variable while determing POC. As per my understanding, Is it correct if we consider only TopField variable thinking like a frame. Please give your inputs considering me as a newbie in this field. Many thanks in advance. Regards, Atul __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050602/adabe0ae/attachment.html From alexismt comcast.net Thu Jun 2 09:51:19 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Thu Jun 2 17:39:43 2005 Subject: [Mp4-tech] [H.264]Mode decision by frame In-Reply-To: Message-ID: <200506021551.j52FpNC6013101@lists1.magma.ca> Dear Sebastian, The IntraDisableInterOnly parameter works only in combination with the additional Intra mode disabling parameters in the config. This implies that even if you enable all of them, you may still see certain intra modes within inter slices (i.e. DC prediction). Currently, the only way to disable intra in inter slices is through the code. You should go in function init_enc_mb_params() (mode_decision.c) and modify the following lines : enc_mb->valid[I8MB] = input->Transform8x8Mode; enc_mb->valid[I4MB] = (input->Transform8x8Mode==2) ? 0:1; enc_mb->valid[I16MB] = 1; enc_mb->valid[IPCM] = (input->symbol_mode != CABAC && input->EnableIPCM); to the following : enc_mb->valid[I8MB] = intra ? input->Transform8x8Mode : 0; enc_mb->valid[I4MB] = intra ? (input->Transform8x8Mode==2) ? 0:1) : 0; enc_mb->valid[I16MB] = intra ? 1 : 0; enc_mb->valid[IPCM] = intra ? (input->symbol_mode != CABAC && input->EnableIPCM) : 0; Note that the code might still be doing some things for chroma intra, so you may wish to modify the code a bit further if you wish to avoid those computations. Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Sebastian Lopez Suarez Sent: Thursday, June 02, 2024 12:19 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264]Mode decision by frame Hi, Does anyone knows how to manage with the configuration file in order to avoid the use of intra macroblocks in P frames? I was triying with the IntraDisableInterOnly but it does not seems to work .... Thanks a lot in advance. _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From donald.koeleman hoppers.nl Fri Jun 3 00:27:19 2005 From: donald.koeleman hoppers.nl (Donald Koeleman) Date: Thu Jun 2 20:54:25 2005 Subject: [Mp4-tech] Standalone H264 player References: <20050601220815.12F48191E0@olive.qinip.net> Message-ID: <001301c567b9$dbb351a0$0202a8c0@rocklake> As for VLC, may I suggest downloading this one (8.0.2 first beta), instead of 8.0.1. Info: http://forum.videolan.org/viewtopic.php?t=9199 Direct download: windows: http://download2.videolan.org/pub/videolan/testing/vlc-0.8.2-test1/win32/vlc-0.8.2-test1-win32.exe osx: http://download2.videolan.org/pub/videolan/testing/vlc-0.8.2-test1/macosx/vlc-0.8.2-test1.dmg For linux and other platforms grab the sourcecode. I tried it today, and it seems to handle Quicktime produced h.264 files quite well. Including some of the Quicktime HD gallery stuff. I downloaded the 420P version of the Serenity promo, and I was able to play it back on a Dell p4 2.8c with 512 MByte of RAM. I could not get the demo files at http://www.fastvdo.com/H.264.html to play. Last time I tried the demo version of their Windows Media Plug-in repeatedly failed to install, so I couldn't try it with their own codec. A few weeks ago, I tried ffdshow and one or two players using the ffdshow library, but have never been able to get any of it to play h.264 files. Earlier in the month (oops that's May), I did download the Moonlight player/codec, I am not sure (belief I did, though) if I tried installing it and failed (perhaps I did not get all of the components and downloaded just some filters, or an older version, as there is an Israeli and a Russian site, and various downloadable stuff), or left installation for another day. Anyway, it never played any files on my pc, so I can't comment on its quality. For me, this new vlc beta worked right out of the box on the Quicktime 7 files, but not on the other files (using the .264 extension), I found online. Perhaps some-one can suggest some links to other non-qt files to try? cheers, Donald ----- Original Message ----- From: "Rob Koenen (MPEGIF)" To: Cc: Sent: Thursday, June 02, 2024 12:08 AM Subject: FW: [Mp4-tech] Standalone H264 player > Forwarding auto-discarded message from non-member email address. > > Rob > > -----Original Message----- > From: Ben Waggoner [mailto:ben@benwaggoner.com] > Sent: Wednesday, June 01, 2024 19:32 > To: Jayant Chauhan > Cc: mp4-tech@lists.mpegif.org > Subject: Re: [Mp4-tech] Standalone H264 player > > Jayant, > > QuickTime 7 on Mac will play H.264 in .mp4 format. Windows > version forthcoming > > VideoLAN Client (VLC) has some H.264 support in its current > 0.8.1 version, although it's much improved in the version in CVS. It > supports the codec in a variety of wrappers, including .mp4. > > There's a Moonlight player for Windows as well, but I haven't > spent much time with it. > > > Ben Waggoner, Interframe Media http://www.interframemedia.com > Technology and Strategy for Digital Media > > My Book: http://www.cmpbooks.com/compression/ > My Tutorials http://www.classondemand.net/benwaggoner/ > My Stanford Class http://www.digitalmediaacademy.org/compression.html > > On Jun 1, 2005, at 3:45 AM, Jayant Chauhan wrote: > > > Is there any standalone player for H264 files?! From garysull windows.microsoft.com Thu Jun 2 17:45:21 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Thu Jun 2 21:02:07 2005 Subject: [Mp4-tech] TopField and BottomField in Baseline proffile and POC Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460ED3CE50@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> There is nothing in the standard that prevents using the Baseline profile for interlaced video coding (although it will not be as efficient for interlaced video coding as other profiles would be). There is also nothing that forces the TopFieldOrderCnt and BottomFieldOrderCnt to be equal to each other when using the Baseline profile. However, I think I agree that it would probably be sufficient, when using the Baseline profile, to derive only the TopFieldOrderCnt in order to achieve decoder conformance. Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of atul tomar Sent: Thursday, June 02, 2024 8:40 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] TopField and BottomField in Baseline proffile and POC Hi, As we know that we dont use interlacing in baseline proffile so Is it necessary to determine PictureOrderCount. I knew through standard that it is also used for decoder conformance checking but how Can we get TopField and BottomField field related variables if we are not dividing frame in fields then how can we work with above two variable while determing POC. As per my understanding, Is it correct if we consider only TopField variable thinking like a frame. Please give your inputs considering me as a newbie in this field. Many thanks in advance. Regards, Atul __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050602/5753bf85/attachment.html From alexismt comcast.net Fri Jun 3 00:51:33 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Thu Jun 2 21:08:02 2005 Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis Message-ID: <060220052351.22579.429F9B850004ED200000583322058863609B039C07980A040E@comcast.net> Again, to clarify, this parameter was not added as a way to disable all intra modes within inter slices, but only the "more complicated" directional intra modes. I guess the confusion comes from the description in the config file (i.e. it states that the purpose is to "Disable Intra modes for Non I-Slices"), while in the JM reference manual (JVT-O017) the purpose of this parameter is clearly stated. I can understand however that we should not expect everyone to check the reference manual, and therefore a better comment should be added in the config file. Best regards, Alexis -------------- Original message -------------- > I have experienced the same as you. Even when I define that each frame is > composed by only one slice and I fix that IntraDisableInterOnly = 1 the > the results indicate that some macroblocks are being coded as intra > in P frames. > > ??????? > > Best! > > On Thu, 2 Jun 2005, Andrew Krupiczka wrote: > > > > > Sebastian, > > > > You can try > > > > IntraDisableInterOnly = 1 # Disable Intra modes for Non I-Slices > > > > but based on my past brief testing I'm not sure it's working as it was > > intended to :-( > > > > Best regards, > > > > Andrew > > > > > > > > > > > > -----Original Message----- > > From: Sebastian Lopez Suarez [mailto:seblopez@iuma.ulpgc.es] > > Sent: Wednesday, June 01, 2024 3:02 PM > > To: mp4-tech@lists.mpegif.org > > Subject: [Mp4-tech] [Video H.264] Mode decision on a frame basis > > > > Hi, > > > > Does anybody knows how to manage with the configuration file in order > > to avoid intra macroblocks in P frames? Thanks a lot in advance. > > > > Best! > > _______________________________________________ > > NOTE: Please use clear subject lines for your posts. Include [audio, > > [video], [systems], [general] or another apppropriate identifier to > > indicate the type of question you have. > > > > Note: Conduct on the mailing list is subject to the Antitrust guidelines > > found at > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], > [systems], [general] or another apppropriate identifier to indicate the type of > question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found > at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050602/2763d57e/attachment.html From samuel dc.fi.udc.es Fri Jun 3 09:40:15 2005 From: samuel dc.fi.udc.es (Samuel Rivas) Date: Fri Jun 3 15:03:02 2005 Subject: [Mp4-tech] MPEG-2 TS converter In-Reply-To: <84ac01305060207274b79c196@mail.gmail.com> References: <200505260142.j4Q1gkYA051353@smtp0.dreamwiz.com> <84ac013050526032723a69ced@mail.gmail.com> <20050602081343.GA7236@crusher.lfcia.pri> <84ac01305060207274b79c196@mail.gmail.com> Message-ID: <20050603064015.GA17262@crusher.lfcia.pri> Alejandro Moya wrote: > You needn't insert PMT or PAT periodically because what our friend > Ik-Hwan asked was the way to convert one program stream into a TS for > testing pourposes, not for streaming: when you broadacast a TS, even > if it has only one program, you need to insert PMT or PAT so decoders > can be able to get the program list and their associated PIDs; thatis > not needed when using a TS inside a file. > "continuity counters " aren't needed either and "payload_unit_start" > bit is set in the byte sequence I told him to insert as transport > stream header. > The only thing I forgot to comment to make an easy test TS was that > each transport packet should be 188 bytes length, so PMT and PAT > packets should be padded with byte 0xFF. Hi again, If I didn't misunderstood anything, Ik-Hwan wants to simulate a streaming application, not building something similar to a transport stream for any strange purpose. Anyway, your headers do not set the payload_unit_start_indicator bit. The first TS packet for each PES packet should be 0x47 [0x0Y | 0x40] 0xYY 0x1X being X the continuity counter (I am assuming that the "mpeg-2 file" Ik-Hwan is talking about is either a PS or a PES). I agree the continuity counter is useless in a stream stored in a file, but is needed if you are going perform some streaming with it (that is what the TS format is for). Second, the stuffing is not only needed for PAT or PMT. A PES length may not be multiple of 184, so the last TS packet for that PES should be stuffed to complete its length. And to do that you must insert an Adaptation Field (TS packet header is changed) Last, if you are going to test any streaming system, you need the clock references be inserted in the transport stream. For a program with only one stream it is an easy task, but must be done. In short, build a TS is not that easy as cutting up one stream in 184 byte length chunks and adding one simple header to those chunks. Regards -- Samuel From samuel dc.fi.udc.es Fri Jun 3 09:50:20 2005 From: samuel dc.fi.udc.es (Samuel Rivas) Date: Fri Jun 3 15:10:58 2005 Subject: [Mp4-tech] MPEG-2 TS converter In-Reply-To: <200506021522.j52FM8sS011079@lists1.magma.ca> References: <20050602081343.GA7236@crusher.lfcia.pri> <200506021522.j52FM8sS011079@lists1.magma.ca> Message-ID: <20050603065020.GA17565@crusher.lfcia.pri> Vinay Sathe wrote: > Why do you say "more or less correct"? In your experience, is videolan > "less > correct" in any aspect of creating a transport stream? Hi, It creates syntactically correct TS. Nevertheless, up to version 0.7.2, generated streams didn't complain with the system target decoder restrictions. I haven't tested later versions. However, the TS streams created by VLC are good enough for most of the cases (unless you use a decoder with a very short input buffer). In fact we used it extensively without problems. Regards. -- Samuel From abahari ed.ac.uk Fri Jun 3 11:32:45 2005 From: abahari ed.ac.uk (Asral Bahari) Date: Fri Jun 3 15:17:52 2005 Subject: [Mp4-tech] [Video-H.264] profiler Message-ID: <1117791164.42a023bd006ee@sms.ed.ac.uk> Hi, Where can I get the iprof - instruction level profiler written by Peter Kuhn? Thanks From almomo1 gmail.com Fri Jun 3 14:51:46 2005 From: almomo1 gmail.com (Alejandro Moya) Date: Fri Jun 3 15:24:13 2005 Subject: [Mp4-tech] MPEG-2 TS converter Message-ID: <84ac013050603045124b3e45f@mail.gmail.com> Hi, Yes, you are right: I forgot to say that payload_unit_indicator should be set; and there is something more I forgot: PMT and PAT will have their first byte in the payload to 0x00, as the pointer to the begining of the first sections in then transport packet. If you don't want to add Adapation Field for padding pourposes, you can make your PES packet's lenght multiple of 184, but I guessed there wasn't need to say it. And last, about timing, I am not sure if I am right, but I think you needn't it if you have only one program because PES packets have their own timing information but not the global system clock; I follow the steps I explained later to make my own easy TS, and I have tested them in Moonlight decoder and they worked. Obviously, it is a very simple way to create TS, but I think it is clear that if you want to fully test the whole range of possibilities a TS can offer, you can't do it "by hand". I said that, for doing that, you should pay for a TS encoder, as I don't know any freeware TS encoder or multiplexer. Greetings. From pankaj0iit yahoo.co.in Fri Jun 3 06:32:45 2005 From: pankaj0iit yahoo.co.in (pankaj bansal) Date: Fri Jun 3 15:29:42 2005 Subject: [Mp4-tech] One confusion in h.264 decoder Message-ID: <20050603123245.78621.qmail@web8408.mail.in.yahoo.com> Hi, I have one confusion regarding POC info. In H.264 decoder syntex elements(pic_oredr_cnt_lsb, delta_pic_order_cnt_bottom, delta_pic_oredr_cnt[0] and delta_pic_oredr_cnt[1]) used for POC comes in every slice, but POC information is derived for every frame, field(wheather decoded from a coded field or as a part of a decoded frame), or complementary field pair. Is these are redundant info? if not then how and where these are used for each slice? thanks pankaj --------------------------------- Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050603/299606d4/attachment.html From jmasuhr web.de Fri Jun 3 16:40:54 2005 From: jmasuhr web.de (Julian Masuhr) Date: Fri Jun 3 15:36:41 2005 Subject: [Mp4-tech] [Systems] [Video] Downloading latest reference software Message-ID: <009701c56841$e18ffc80$0200a8c0@jabba> Hi ! I'm a student from the University of Applied Sciences Cologne, working on my bachelor thesis. I hope to find some help where to download the LATEST FREE reference software [Systems] and [Video]. I am also wondering if there are existing binaries for the VOGUE-Software (from MoMuSys-Project) for WinXP or MAC OS. Thank you very much for any help! julian masuhr --- -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050603/15ef2b40/attachment.html From Marco.Miehe sci-worx.com Fri Jun 3 17:47:51 2005 From: Marco.Miehe sci-worx.com (Marco.Miehe@sci-worx.com) Date: Fri Jun 3 15:36:47 2005 Subject: [Mp4-tech] H.264-PrevRefFrameNum and gap handling Message-ID: <7E6219E8CD869D4186B919B0DCF127C0898D2E@hajse201.eu.infineon.com> Dear Members I have a question regarding the PrevRefFrameNum and the process for handling gap. Inside the H.264 standard (ITU-T, 3-2005) chapter 7.4.3, PrevRefFrameNum seems to be set only to frame_num values of reference frames (as the name of the variable also suggests). However, the check in section 8.2.5.2 seems to be executed for all types of frames. But this would lead to the condition to be true e.g in the case that 2 non-reference frames follow a reference frame (e.g. reference-frame has frame_num 0, 1st non-reference has frame_num 1, 2nd non-reference has fram_num 2 >> PrevRefFrameNum = 0 for 2nd non-reference frame, condition evaluates to true), although there is obviously no gap in the stream. The reference software seems to just update PrevRefFrameNum to the frame_num of the current frame after processing it, regardless of the frame type (reference / non-reference), which leads to a gap handling as I would expect it (to see if there are frames missing), but does not really seem to match with the behaviour described by the standard. I'd appreciate if some one clarifies this. Best regards, Marco Miehe From rob.koenen mpegif.org Fri Jun 3 23:36:00 2005 From: rob.koenen mpegif.org (Rob Koenen (MPEGIF)) Date: Fri Jun 3 16:50:38 2005 Subject: [Mp4-tech] FW: Standalone H264 player Message-ID: <20050603203557.0CF7C18233@olive.qinip.net> forward of bounced message from non-subscribed address. >> -----Original Message----- >> From: Pankaj Topiwala [mailto:pankaj@fastvdo.com] >> Sent: Friday, June 03, 2024 20:28 >> To: mp4-tech@lists.mpegif.org >> Cc: support@fastvdo.com >> Subject: Fw: Standalone H264 player >> >> FastVDO aims to provide a world-leading platform for H.264 playback. >> >> One common problem is that to install a plug-in properly, it >> requires root >> (admin) level access. >> Our codec is continually being improved. Please check for >> latest version of >> our codec. If you still have problems, please provide any >> bug reports to >> support@fastvdo.com. >> >> Rgds, >> Pankaj >> ************************************************************** >> *********** >> Dr. Pankaj Topiwala Voice: 410-309-6066 >> President/CEO FastVDO LLC Fax: 410-309-6554 >> 7150 Riverwood Dr., Mobile: 443-538-3782 NEW! >> Columbia, MD 21046-1245 USA Email: pankaj@fastvdo.com NEW! >> ************************************************************** >> *********** >> >> > -----Original Message----- >> > From: mp4-tech-bounces@lists.mpegif.org >> > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of >> Donald Koeleman >> > Sent: Thursday, June 02, 2024 5:27 PM >> > To: mp4-tech@lists.mpegif.org >> > Subject: Re: [Mp4-tech] Standalone H264 player >> > >> > As for VLC, may I suggest downloading this one (8.0.2 first >> beta), instead >> > of 8.0.1. >> > >> > Info: >> > http://forum.videolan.org/viewtopic.php?t=9199 >> > >> > Direct download: >> > windows: >> > >> http://download2.videolan.org/pub/videolan/testing/vlc-0.8.2-t >> est1/win32/vlc >> > -0.8.2-test1-win32.exe >> > >> > osx: >> > >> http://download2.videolan.org/pub/videolan/testing/vlc-0.8.2-t >> est1/macosx/vl >> > c-0.8.2-test1.dmg >> > >> > For linux and other platforms grab the sourcecode. >> > >> > >> > >> > I tried it today, and it seems to handle Quicktime produced >> h.264 files >> > quite well. Including some of the Quicktime HD gallery >> stuff. I downloaded >> > the 420P version of the Serenity promo, and I was able to >> play it back on >> > a >> > Dell p4 2.8c with 512 MByte of RAM. >> > >> > I could not get the demo files at >> http://www.fastvdo.com/H.264.html to >> > play. >> > Last time I tried the demo version of their Windows Media Plug-in >> > repeatedly >> > failed to install, so I couldn't try it with their own codec. >> > >> > A few weeks ago, I tried ffdshow and one or two players >> using the ffdshow >> > library, but have never been able to get any of it to play >> h.264 files. >> > >> > Earlier in the month (oops that's May), I did download the Moonlight >> > player/codec, I am not sure (belief I did, though) if I >> tried installing >> > it >> > and failed (perhaps I did not get all of the components and >> downloaded >> > just >> > some filters, or an older version, as there is an Israeli >> and a Russian >> > site, and various downloadable stuff), or left installation >> for another >> > day. >> > Anyway, it never played any files on my pc, so I can't >> comment on its >> > quality. >> > >> > For me, this new vlc beta worked right out of the box on >> the Quicktime 7 >> > files, but not on the other files (using the .264 >> extension), I found >> > online. Perhaps some-one can suggest some links to other >> non-qt files to >> > try? >> > >> > >> > >> > cheers, >> > >> > >> > >> > Donald >> > >> > ----- Original Message ----- >> > >> > From: "Rob Koenen (MPEGIF)" >> > To: >> > Cc: >> > Sent: Thursday, June 02, 2024 12:08 AM >> > Subject: FW: [Mp4-tech] Standalone H264 player >> > >> > >> >> Forwarding auto-discarded message from non-member email address. >> >> >> >> Rob >> >> >> >> -----Original Message----- >> >> From: Ben Waggoner [mailto:ben@benwaggoner.com] >> >> Sent: Wednesday, June 01, 2024 19:32 >> >> To: Jayant Chauhan >> >> Cc: mp4-tech@lists.mpegif.org >> >> Subject: Re: [Mp4-tech] Standalone H264 player >> >> >> >> Jayant, >> >> >> >> QuickTime 7 on Mac will play H.264 in .mp4 format. Windows >> >> version forthcoming >> >> >> >> VideoLAN Client (VLC) has some H.264 support in its current >> >> 0.8.1 version, although it's much improved in the version >> in CVS. It >> >> supports the codec in a variety of wrappers, including .mp4. >> >> >> >> There's a Moonlight player for Windows as well, but I haven't >> >> spent much time with it. >> >> >> >> >> >> Ben Waggoner, Interframe Media >> http://www.interframemedia.com >> >> Technology and Strategy for Digital Media >> >> >> >> My Book: http://www.cmpbooks.com/compression/ >> >> My Tutorials http://www.classondemand.net/benwaggoner/ >> >> My Stanford Class >> http://www.digitalmediaacademy.org/compression.html >> >> >> >> On Jun 1, 2005, at 3:45 AM, Jayant Chauhan wrote: >> >> >> >> > Is there any standalone player for H264 files?! >> > >> > >> > _______________________________________________ >> > NOTE: Please use clear subject lines for your posts. Include [audio, >> > [video], [systems], [general] or another apppropriate identifier to >> > indicate >> > the type of question you have. >> > >> > Note: Conduct on the mailing list is subject to the >> Antitrust guidelines >> > found at >> > >> http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti >> trust.php >> > >> > >> >> >> > > From ghughes broadbus.com Fri Jun 3 18:44:48 2005 From: ghughes broadbus.com (Gary Hughes) Date: Sat Jun 4 05:17:49 2005 Subject: [Mp4-tech] MPEG-2 TS converter Message-ID: <3B83C5E7EAB53F44A16CDFB70F9E4BF301248846@btiexch01.broadbus.com> > And last, about timing, I am not sure if I am > right, but I think you needn't it if you have only one > program because PES packets have their own timing information > but not the global system clock; I follow the steps I > explained later to make my own easy TS, and I have tested > them in Moonlight decoder and they worked. > You absolutely need a PCR for each program in the TS for it to be a legal transport stream. There is no notion of a 'global system clock' in a TS; nominally each program has its own PCR, although if you could slave all of the streams in a MPTS to a common clock you could conceivably share PCRs amongst programs. The spec does not prohibit this, but does not encourage it either. > Obviously, it is a very simple way to create TS, but I think > it is clear that if you want to fully test the whole range of > possibilities a TS can offer, you can't do it "by hand". I > said that, for doing that, you should pay for a TS encoder, > as I don't know any freeware TS encoder or multiplexer. If it doesn't meet all of the requirements specified for transport stream specified in 13818-1, it isn't a transport stream. The only freeware TS multiplexer I know of is VLC, which as has been already discussed does not follow the T-STD timing model. On the other hand, I've tried a few expensive multiplexers which didn't get that right either. I work primarily in cable VOD and many set tops in the cable space have the absolute minimum amount of buffer required by the standard so getting the timing right is critical for my applications. gary Gary Hughes Director of Media Engineering Broadbus Technologies, Inc. ghughes@broadbus.com v: 978 264 7919 f: 978 264 9108 -------------------------------------------------------- This email message and any files transmitted with it contain confidential information intended only for the person(s) to whom this email message is addressed. If you have received this email message in error, please notify the sender immediately by telephone or email and destroy the original message without making a copy. Thank you. From donald.koeleman hoppers.nl Sat Jun 4 01:21:59 2005 From: donald.koeleman hoppers.nl (Donald Koeleman) Date: Sat Jun 4 05:24:40 2005 Subject: [Mp4-tech] FW: Standalone H264 player References: <20050603203557.0CF7C18233@olive.qinip.net> Message-ID: <1e2301c5688a$a98bbc90$0202a8c0@rocklake> Dear Pankaj, Administrative rights weren't the problem on my side, it was the installer that balked on two different computers, returning the same error message. This was about 4 weeks ago, and I hadn't had a change to return to it, as explained in my post to this list. So, I just downloaded the 1.0 demo version (I believe the one that failed to install was also labelled 1.0, that's why I like it when software developers add clear release dates and build numbers on their websites, so one can directly see there is a newerversion out there, to fix any problems one may have with the existing version). Anyway, downloaded it, Installed it on one of the two pc I tried last time and opened one of your cif resolution demo clips I still had on my harddrive from last time. Of course wmp still popped-up a window, saying it didn't recognize the file extension, asking if it should try and play the file anyway. I went ahead and yes it does play the file just fine. I also tried the 720p highway clip, but noticed some motion shudder on the cars as they drove across the screen in a diagonal from the bottom left. The highway wasn't crowded with cars, so the segment wasn't a worst case demo, however I do suspect that my pc wasn't powerfull enough for smooth playback. I tried this on a p4 2.8c 512 MByte RAM Dell Dimension 8300 machine (Intel supplied I875P motherboard). So, perhaps others on this list could compare it using faster machines and or other codecs/players? I know it is advertised as a plug-in for wmp, but couldn't resist trying to see if it enabled playback in another dshow player, winamp, this failed for the obvious reason; it is a wmp plug-in. Playing back QT7 produced h.264 files did not work as wmp couldn't handle the .mov filestructure, and couldn't find the appropriate audiocodec. Pankaj, and all of you who aim to provide a world-leading platform for H.264 playback, success to you all, as 720P sure makes for a nice picture on my 17" tft monitor. Regards, Donald ----- Original Message ----- From: "Rob Koenen (MPEGIF)" To: Sent: Friday, June 03, 2024 10:36 PM Subject: [Mp4-tech] FW: Standalone H264 player > forward of bounced message from non-subscribed address. > > >> -----Original Message----- > >> From: Pankaj Topiwala [mailto:pankaj@fastvdo.com] > >> Sent: Friday, June 03, 2024 20:28 > >> To: mp4-tech@lists.mpegif.org > >> Cc: support@fastvdo.com > >> Subject: Fw: Standalone H264 player > >> > >> FastVDO aims to provide a world-leading platform for H.264 playback. > >> > >> One common problem is that to install a plug-in properly, it > >> requires root > >> (admin) level access. > >> Our codec is continually being improved. Please check for > >> latest version of > >> our codec. If you still have problems, please provide any > >> bug reports to > >> support@fastvdo.com. > >> > >> Rgds, > >> Pankaj > >> ************************************************************** > >> *********** > >> Dr. Pankaj Topiwala Voice: 410-309-6066 > >> President/CEO FastVDO LLC Fax: 410-309-6554 From samuel dc.fi.udc.es Sat Jun 4 12:41:05 2005 From: samuel dc.fi.udc.es (Samuel Rivas) Date: Sat Jun 4 14:00:01 2005 Subject: [Mp4-tech] MPEG-2 TS converter In-Reply-To: <84ac013050603045124b3e45f@mail.gmail.com> References: <84ac013050603045124b3e45f@mail.gmail.com> Message-ID: <20050604094105.GA3499@riker.dc.fi.udc.es> Alejandro Moya wrote: > Yes, you are right: I forgot to say that payload_unit_indicator should be > set; and there is something more I forgot: PMT and PAT will have their first > byte in the payload to 0x00, as the pointer to the begining of the first > sections in then transport packet. That is true, I did not realise that. > If you don't want to add Adapation Field for padding pourposes, you can make > your PES packet's lenght multiple of 184, but I guessed there wasn't need to > say it. Agreed, but then you have to pack an elementary stream or repack a PE stream. I assumed that the goal was transforming an already packetized PE stream in a TS stream. > And last, about timing, I am not sure if I am right, but I think you needn't > it if you have only one program because PES packets have their own > timing information but not the global system clock; The PES timing information are the timestamps for the access units. Theoretically, you still need the clock references to set the decoder clock. In streaming, those references are used to control the transfer rate. Thoso time references should avoid buffer overflows in the decoder, but sofware decoders usually have buffers much bigger than the suggested in the standard and they use their own mechanisms to control them. > I follow the steps I explained later to > make my own easy TS, and I have tested them in Moonlight decoder and they > worked. Many decoders ignore certain data of the TS streams. E.g. mplayer just trhows away all headers and decodes the elementary streams. So it depends on the tools you are using. > Obviously, it is a very simple way to create TS, but I think it is clear > that if you want to fully test the whole range of possibilities a TS can > offer, you can't do it "by hand". I said that, for doing that, you should > pay for a TS encoder, as I don't know any freeware TS encoder or > multiplexer. As I said before, VLC is a good tool to create TS. To multiplex different elementary streams ffmpeg and transcode are usefull too, but way more complex. Regards. -- Samuel From ksuehring web.de Sat Jun 4 13:12:18 2005 From: ksuehring web.de (Karsten Suehring) Date: Sat Jun 4 14:08:08 2005 Subject: [Mp4-tech] One confusion in h.264 decoder In-Reply-To: <20050603123245.78621.qmail@web8408.mail.in.yahoo.com> References: <20050603123245.78621.qmail@web8408.mail.in.yahoo.com> Message-ID: <42A17E82.6010906@web.de> Pankaj, since H.264/AVC contains no picture header concept there are some syntax elements which need to be repeated (with the same value) in every slice header of a picture. The POC elements are an example for that. The whole list of syntax elements which shall have the same value in all slice headers of a picture is contained in the beginning of section 7.4.3. Best regards, Karsten pankaj bansal wrote: > Hi, > I have one confusion regarding POC info. > In H.264 decoder syntex elements(pic_oredr_cnt_lsb, > delta_pic_order_cnt_bottom, delta_pic_oredr_cnt[0] and > delta_pic_oredr_cnt[1]) used for POC comes in every slice, but POC > information is derived for every frame, field(wheather decoded from a > coded field or as a part of a decoded frame), or complementary field pair. > Is these are redundant info? if not then how and where these are > used for each slice? > thanks > pankaj > > Discover Yahoo! > Stay in touch with email, IM, photo sharing & more. Check it out! > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From zhengyeliu gmail.com Sat Jun 4 08:03:21 2005 From: zhengyeliu gmail.com (Zhengye Liu) Date: Sat Jun 4 14:14:30 2005 Subject: [Mp4-tech] How to get Momusys v2.1 for Mpeg4 FGS? Message-ID: Hi, I am a graduate student in Polytechnic University, and I am doing some research on Mpeg4 FGS now. I checked that Mpeg4 FGS had been supported in Momusys v2.1. But how to get it. From the ISO website, I just got v1.0, which does not support FGS. Can anybody give me a link to download Momusys v2.1? Thanks in advance! Zhengye From rob.koenen mpegif.org Sat Jun 4 20:37:16 2005 From: rob.koenen mpegif.org (Rob Koenen (MPEGIF)) Date: Sat Jun 4 19:05:52 2005 Subject: FW: [Mp4-tech] H.264-PrevRefFrameNum and gap handling Message-ID: <20050604173716.0C487182E5@olive.qinip.net> Forward of email sent from non-member address -----Original Message----- From: Karsten Suehring [mailto:Karsten.Suehring@hhi.fraunhofer.de] Sent: Saturday, June 04, 2023 12:08 To: mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] H.264-PrevRefFrameNum and gap handling Marco, please read ahead in the frame_num semantics section: "If gaps_in_frame_num_value_allowed_flag is equal to 0, the value of frame_num for the current picture shall be equal to ( PrevRefFrameNum + 1 ) % MaxFrameNum." This means if you don't want a gap in frame_num, the second non-reference frame should also have frame_num equal to 1. Even the following reference frame would have frame_num equal to 1. If you continue with a non-reference frame frame_num would now be set to 2. Best regards, Karsten Marco.Miehe@sci-worx.com wrote: > Dear Members > > I have a question regarding the PrevRefFrameNum and the process for > handling gap. > > Inside the H.264 standard (ITU-T, 3-2005) chapter 7.4.3, PrevRefFrameNum > seems to be set only to frame_num values of reference frames (as the > name of the variable also suggests). However, the check in section > 8.2.5.2 seems to be executed for all types of frames. But this would > lead to the condition to be true e.g in the case that 2 non-reference > frames follow a reference frame (e.g. reference-frame has frame_num 0, > 1st non-reference has frame_num 1, 2nd non-reference has fram_num 2 >> > PrevRefFrameNum = 0 for 2nd non-reference frame, condition evaluates to > true), although there is obviously no gap in the stream. The reference > software seems to just update PrevRefFrameNum to the frame_num of the > current frame after processing it, regardless of the frame type > (reference / non-reference), which leads to a gap handling as I would > expect it (to see if there are frames missing), but does not really seem > to match with the behaviour described by the standard. > > I'd appreciate if some one clarifies this. > > > Best regards, > > Marco Miehe > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From amantrav gmail.com Sat Jun 4 19:52:01 2005 From: amantrav gmail.com (Anand Mantravadi) Date: Mon Jun 6 00:47:59 2005 Subject: [Mp4-tech] Problem with H.264 decoder Message-ID: Hello Everybody, I am using H.264 codec version JM 9.5 and I have encoded my video in the RTP packet format. I use some packet dropping scheme to drop the RTP packets as per a given Packet Loss Rate and feed the rest to the decoder. But for any case, I get an error which says...... 'duplicate frame in short term reference buffer' and the decoder stops. I am not sure why this happens. Can anyone please explain the reason and what can be done in this regard. It is really urgent for me and so I request you to kindly help me out with your suggestions..... Thanks and Regards... Anand From amantrav gmail.com Sat Jun 4 20:08:35 2005 From: amantrav gmail.com (Anand Mantravadi) Date: Mon Jun 6 00:48:05 2005 Subject: [Mp4-tech] urgent help needed Message-ID: Hello Everybody, I am using H.264 codec version JM 9.5 and I have encoded my video in the RTP packet format. I use some packet dropping scheme to drop the RTP packets as per a given Packet Loss Rate and feed the rest to the decoder. But for any case, I get an error which says...... 'duplicate frame in short term reference buffer' and the decoder stops. I am not sure why this happens. Can anyone please explain the reason and what can be done in this regard. It is really urgent for me and so I request you to kindly help me out with your suggestions..... Thanks and Regards... Anand From mpeg21 hotmail.com Mon Jun 6 10:49:41 2005 From: mpeg21 hotmail.com (Yung Hwa Jang) Date: Mon Jun 6 00:48:10 2005 Subject: [Mp4-tech] About dpb_combine_field_yuv in jm96 Message-ID: Hi, all Would your experts plz explain for me the following excerpt code from the function dpb_combine_field_yuv in jm96? Why does "fs->frame->frame_poc" have two assignments here? //---------------------------------------------------------------------------------------------------------------------------------- fs->poc=fs->frame->poc =fs->frame->frame_poc = min (fs->top_field->poc, fs->bottom_field->poc); fs->bottom_field->frame_poc=fs->top_field->frame_poc= fs->bottom_field->top_poc=fs->frame->frame_poc=fs->frame->top_poc=fs->top_field->poc; fs->top_field->bottom_poc=fs->bottom_field->poc; fs->frame->bottom_poc=fs->bottom_field->poc; //---------------------------------------------------------------------------------------------------------------------------------- Thanks in advance Best regards Jang _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ From vikass gmail.com Mon Jun 6 10:28:01 2005 From: vikass gmail.com (Vikas Singhal) Date: Mon Jun 6 00:48:15 2005 Subject: [Mp4-tech] AVC in transport stream In-Reply-To: <7E6219E8CD869D4186B919B0DCF127C0898D2E@hajse201.eu.infineon.com> References: <7E6219E8CD869D4186B919B0DCF127C0898D2E@hajse201.eu.infineon.com> Message-ID: Hi, Can anybody point me to location from where I can download the transport streams having AVC as the payload. Thanks in Advance Vikas From ksuehring web.de Mon Jun 6 09:56:05 2005 From: ksuehring web.de (Karsten Suehring) Date: Mon Jun 6 16:30:14 2005 Subject: [Mp4-tech] Problem with H.264 decoder In-Reply-To: References: Message-ID: <42A3F385.6040901@web.de> Anand, the JM reference software currently supports only loss of slices that are smaller than a whole frame (note: this includes single fields). If frame loss occurs, the DPB will have the wrong content that leads to an error condition (e.g. in your case the try to store a frame with the same frame_num like a reference frame that already exists in the DPB) I would also suggest not to rely on the RTP implementation. It has been implemented from a very early draft of the RTP RFC (which differed pretty much from the final version) and is currently not maintained. Best regards, Karsten Anand Mantravadi wrote: > Hello Everybody, > > I am using H.264 codec version JM 9.5 and I have encoded my > video in the RTP packet format. I use some packet dropping scheme to > drop the RTP packets as per a given Packet Loss Rate and feed the rest > to the decoder. But for any case, I get an error which says...... > > 'duplicate frame in short term reference buffer' > > and the decoder stops. I am not sure why this happens. Can anyone > please explain the reason and what can be done in this regard. It is > really urgent for me and so I request you to kindly help me out with > your suggestions..... > > Thanks and Regards... > Anand > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From pankaj0iit yahoo.co.in Mon Jun 6 03:20:40 2005 From: pankaj0iit yahoo.co.in (pankaj bansal) Date: Mon Jun 6 16:30:19 2005 Subject: [Mp4-tech] test case generation for h.264 decoder Message-ID: <20050606092040.76508.qmail@web8401.mail.in.yahoo.com> Hi all, Can anybody tell me how to generate the test stream with the help of reference JM software for setting the vaues of crop_left = 4, crop_right = 6, crop_top = 2,crop_bottom = 4. thank you pankaj --------------------------------- Yahoo! Mail Stay connected, organized, and protected. Take the tour -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050606/d49343a5/attachment.html From pankaj0iit yahoo.co.in Mon Jun 6 03:55:43 2005 From: pankaj0iit yahoo.co.in (pankaj bansal) Date: Mon Jun 6 16:30:24 2005 Subject: [Mp4-tech] raw data files for testing interlacing Message-ID: <20050606095543.71613.qmail@web8407.mail.in.yahoo.com> Hi All, Can you tell me the source of getting some of the following standard raw test streams. The size must be either 525 SD or 625 SD or SQCIF or VGA(640x480). (1)football (2)bus (3)ragby (4)tempete (5)mobile (6)coastguard (7)stefan (8)stefan_football (9)canoa thanks pankaj --------------------------------- Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050606/0a630083/attachment.html From amantrav gmail.com Mon Jun 6 10:26:34 2005 From: amantrav gmail.com (Anand Mantravadi) Date: Mon Jun 6 16:30:30 2005 Subject: [Mp4-tech] Problem with H.264 decoder In-Reply-To: <42A3F385.6040901@web.de> References: <42A3F385.6040901@web.de> Message-ID: Dear Dr.Karsten Thanks for the suggestion. I will be more specific in my question now. I am using the data partitioning mode in H.264 and I have 9 slices per frame. Now, if I drop large number of Partition A packets, I am able to decode, but i even I drop very few on Partition C, I am not able to decode and I get the error which I mentioned previously. ('duplecate frame number....') But this is opposite to the concept that Partition A is more important and so needs more protection. Can you please point out where is the mistake. I am using the NAL header to sort the partitions. Anand On 6/6/05, Karsten Suehring wrote: > Anand, > > the JM reference software currently supports only loss of slices that > are smaller than a whole frame (note: this includes single fields). If > frame loss occurs, the DPB will have the wrong content that leads to an > error condition (e.g. in your case the try to store a frame with the > same frame_num like a reference frame that already exists in the DPB) > > I would also suggest not to rely on the RTP implementation. It has been > implemented from a very early draft of the RTP RFC (which differed > pretty much from the final version) and is currently not maintained. > > Best regards, > Karsten > > > Anand Mantravadi wrote: > > Hello Everybody, > > > > I am using H.264 codec version JM 9.5 and I have encoded my > > video in the RTP packet format. I use some packet dropping scheme to > > drop the RTP packets as per a given Packet Loss Rate and feed the rest > > to the decoder. But for any case, I get an error which says...... > > > > 'duplicate frame in short term reference buffer' > > > > and the decoder stops. I am not sure why this happens. Can anyone > > please explain the reason and what can be done in this regard. It is > > really urgent for me and so I request you to kindly help me out with > > your suggestions..... > > > > Thanks and Regards... > > Anand > > > > _______________________________________________ > > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > -- Anand Mantravadi EE Graduate Student SUNY at Buffalo From ksuehring web.de Mon Jun 6 16:51:12 2005 From: ksuehring web.de (Karsten Suehring) Date: Mon Jun 6 16:30:35 2005 Subject: [Mp4-tech] Problem with H.264 decoder In-Reply-To: References: <42A3F385.6040901@web.de> Message-ID: <42A454D0.9090609@web.de> I think we didn't test slice-loss with data partitioning after re-enabling the error concealment features. I think the current decoder code starts decoding only after receiving all threee partitions. If one is missing it will probably mix up partitions of different pictures... There is currently also no code for decoding only from one or two of the three partitions, so the quick fix would be to discard the incomplete partitions which would be identical with encoding without DP. Best regards, Karsten Anand Mantravadi wrote: > Dear Dr.Karsten > > Thanks for the suggestion. I will be more specific in my question now. > I am using the data partitioning mode in H.264 and I have 9 slices per > frame. Now, if I drop large number of Partition A packets, I am able > to decode, but i even I drop very few on Partition C, I am not able > to decode and I get the error which I mentioned previously. > ('duplecate frame number....') > > But this is opposite to the concept that Partition A is more important > and so needs more protection. Can you please point out where is the > mistake. I am using the NAL header to sort the partitions. > > Anand > > > On 6/6/05, Karsten Suehring wrote: > >>Anand, >> >>the JM reference software currently supports only loss of slices that >>are smaller than a whole frame (note: this includes single fields). If >>frame loss occurs, the DPB will have the wrong content that leads to an >>error condition (e.g. in your case the try to store a frame with the >>same frame_num like a reference frame that already exists in the DPB) >> >>I would also suggest not to rely on the RTP implementation. It has been >>implemented from a very early draft of the RTP RFC (which differed >>pretty much from the final version) and is currently not maintained. >> >>Best regards, >>Karsten >> >> >>Anand Mantravadi wrote: >> >>>Hello Everybody, >>> >>> I am using H.264 codec version JM 9.5 and I have encoded my >>>video in the RTP packet format. I use some packet dropping scheme to >>>drop the RTP packets as per a given Packet Loss Rate and feed the rest >>>to the decoder. But for any case, I get an error which says...... >>> >>>'duplicate frame in short term reference buffer' >>> >>>and the decoder stops. I am not sure why this happens. Can anyone >>>please explain the reason and what can be done in this regard. It is >>>really urgent for me and so I request you to kindly help me out with >>>your suggestions..... >>> >>>Thanks and Regards... >>>Anand >>> >>>_______________________________________________ >>>NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. >>> >>>Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php >> >> > > From amantrav gmail.com Mon Jun 6 11:17:04 2005 From: amantrav gmail.com (Anand Mantravadi) Date: Mon Jun 6 16:30:40 2005 Subject: [Mp4-tech] Problem with H.264 decoder In-Reply-To: <42A454D0.9090609@web.de> References: <42A3F385.6040901@web.de> <42A454D0.9090609@web.de> Message-ID: Dear Dr.Karsten I guess you are right. But I have one question, if the decoder code is not present to handle losses of the partitions, then the idea of using unequal protection for the partitions to get better performance is not right. But there are many papers in this topic where a better performance is indicated by using DP over non DP cases. So, in a way all that work is not useful in the sense that since the decoder cannot handle partial partitions, there is no point working on using DP. I am also trying to give unequal protection and see the quality of the video from a non DP case. Do you suggest this is wrong in the light of the fact that DP is not properly supported at the decoder for packet loss channels. Anand On 6/6/05, Karsten Suehring wrote: > I think we didn't test slice-loss with data partitioning after > re-enabling the error concealment features. I think the current decoder > code starts decoding only after receiving all threee partitions. If one > is missing it will probably mix up partitions of different pictures... > > There is currently also no code for decoding only from one or two of the > three partitions, so the quick fix would be to discard the incomplete > partitions which would be identical with encoding without DP. > > Best regards, > Karsten > > Anand Mantravadi wrote: > > Dear Dr.Karsten > > > > Thanks for the suggestion. I will be more specific in my question now. > > I am using the data partitioning mode in H.264 and I have 9 slices per > > frame. Now, if I drop large number of Partition A packets, I am able > > to decode, but i even I drop very few on Partition C, I am not able > > to decode and I get the error which I mentioned previously. > > ('duplecate frame number....') > > > > But this is opposite to the concept that Partition A is more important > > and so needs more protection. Can you please point out where is the > > mistake. I am using the NAL header to sort the partitions. > > > > Anand > > > > > > On 6/6/05, Karsten Suehring wrote: > > > >>Anand, > >> > >>the JM reference software currently supports only loss of slices that > >>are smaller than a whole frame (note: this includes single fields). If > >>frame loss occurs, the DPB will have the wrong content that leads to an > >>error condition (e.g. in your case the try to store a frame with the > >>same frame_num like a reference frame that already exists in the DPB) > >> > >>I would also suggest not to rely on the RTP implementation. It has been > >>implemented from a very early draft of the RTP RFC (which differed > >>pretty much from the final version) and is currently not maintained. > >> > >>Best regards, > >>Karsten > >> > >> > >>Anand Mantravadi wrote: > >> > >>>Hello Everybody, > >>> > >>> I am using H.264 codec version JM 9.5 and I have encoded my > >>>video in the RTP packet format. I use some packet dropping scheme to > >>>drop the RTP packets as per a given Packet Loss Rate and feed the rest > >>>to the decoder. But for any case, I get an error which says...... > >>> > >>>'duplicate frame in short term reference buffer' > >>> > >>>and the decoder stops. I am not sure why this happens. Can anyone > >>>please explain the reason and what can be done in this regard. It is > >>>really urgent for me and so I request you to kindly help me out with > >>>your suggestions..... > >>> > >>>Thanks and Regards... > >>>Anand > >>> > >>>_______________________________________________ > >>>NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > >>> > >>>Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > >> > >> > > > > > > -- Anand Mantravadi EE Graduate Student SUNY at Buffalo From shakesp gmail.com Mon Jun 6 18:53:19 2005 From: shakesp gmail.com (shakesp) Date: Mon Jun 6 16:30:46 2005 Subject: [Mp4-tech] [h264] Message-ID: Dear experts, I'm using the JM96 software, how can I find the exact elements into the C slice data partition? From prashanthp gmail.com Tue Jun 7 10:46:28 2005 From: prashanthp gmail.com (Prashanth P) Date: Tue Jun 7 07:44:36 2005 Subject: [Mp4-tech] Tool -Video extractor Message-ID: <71c3c4ac05060621166839e42a@mail.gmail.com> Hi All, I needed immediately the Video Extractor from .AVI file in compressed format. I have Virtual dub which extracts only Video form .AVI file which will be in uncompressed format. could any body help regarding this. Rgds Prashanth From ksuehring web.de Tue Jun 7 11:50:32 2005 From: ksuehring web.de (Karsten Suehring) Date: Tue Jun 7 07:44:42 2005 Subject: [Mp4-tech] Problem with H.264 decoder In-Reply-To: References: <42A3F385.6040901@web.de> <42A454D0.9090609@web.de> Message-ID: <42A55FD8.2050903@web.de> Anand, please don't mix up the usefulness of an error resilience tool with the ability of a certain implementation to handle specific error cases. The target of the JM reference decoder is to (completely) implement the H.264/AVC decoding process. But the recommendation/standard describes only the decoding of conforming streams. If you introduce errors into the bitstream, you are outside the spec and you may implement whatever you think is useful. A number of studies have been accomplished on different error resilience tools of H.264. Most of them were done in a pretty early stage of the standardization process. Some of the error handling tools were included in the reference code while others have been implemented in seperate branches and have never been merged. When the development sped up we even needed to drop some of the features for being able to finish the decoding of conforming streams. I only recently reenabled the non-standardized error concelament for missing slices. But there is sill a large number of error cases which don't have proper handling (bit errors, frame loss, etc.) Volunteering work will always be appreciated. Anyway such an implementation could be only an example because, as I said before, the standard text doesn't contain error descriptions. Best regards, Karsten Anand Mantravadi wrote: > Dear Dr.Karsten > > I guess you are right. But I have one question, if the decoder > code is not present to handle losses of the partitions, then the idea > of using unequal protection for the partitions to get better > performance is not right. But there are many papers in this topic > where a better performance is indicated by using DP over non DP cases. > > So, in a way all that work is not useful in the sense that since > the decoder cannot handle partial partitions, there is no point > working on using DP. I am also trying to give unequal protection and > see the quality of the video from a non DP case. Do you suggest this > is wrong in the light of the fact that DP is not properly supported at > the decoder for packet loss channels. > > Anand > > On 6/6/05, Karsten Suehring wrote: > >>I think we didn't test slice-loss with data partitioning after >>re-enabling the error concealment features. I think the current decoder >>code starts decoding only after receiving all threee partitions. If one >>is missing it will probably mix up partitions of different pictures... >> >>There is currently also no code for decoding only from one or two of the >>three partitions, so the quick fix would be to discard the incomplete >>partitions which would be identical with encoding without DP. >> >>Best regards, >>Karsten >> >>Anand Mantravadi wrote: >> >>>Dear Dr.Karsten >>> >>>Thanks for the suggestion. I will be more specific in my question now. >>>I am using the data partitioning mode in H.264 and I have 9 slices per >>>frame. Now, if I drop large number of Partition A packets, I am able >>>to decode, but i even I drop very few on Partition C, I am not able >>>to decode and I get the error which I mentioned previously. >>>('duplecate frame number....') >>> >>>But this is opposite to the concept that Partition A is more important >>>and so needs more protection. Can you please point out where is the >>>mistake. I am using the NAL header to sort the partitions. >>> >>>Anand >>> >>> >>>On 6/6/05, Karsten Suehring wrote: >>> >>> >>>>Anand, >>>> >>>>the JM reference software currently supports only loss of slices that >>>>are smaller than a whole frame (note: this includes single fields). If >>>>frame loss occurs, the DPB will have the wrong content that leads to an >>>>error condition (e.g. in your case the try to store a frame with the >>>>same frame_num like a reference frame that already exists in the DPB) >>>> >>>>I would also suggest not to rely on the RTP implementation. It has been >>>>implemented from a very early draft of the RTP RFC (which differed >>>>pretty much from the final version) and is currently not maintained. >>>> >>>>Best regards, >>>>Karsten >>>> >>>> >>>>Anand Mantravadi wrote: >>>> >>>> >>>>>Hello Everybody, >>>>> >>>>> I am using H.264 codec version JM 9.5 and I have encoded my >>>>>video in the RTP packet format. I use some packet dropping scheme to >>>>>drop the RTP packets as per a given Packet Loss Rate and feed the rest >>>>>to the decoder. But for any case, I get an error which says...... >>>>> >>>>>'duplicate frame in short term reference buffer' >>>>> >>>>>and the decoder stops. I am not sure why this happens. Can anyone >>>>>please explain the reason and what can be done in this regard. It is >>>>>really urgent for me and so I request you to kindly help me out with >>>>>your suggestions..... >>>>> >>>>>Thanks and Regards... >>>>>Anand >>>>> >>>>>_______________________________________________ >>>>>NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. >>>>> >>>>>Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php >>>> >>>> >>> >> > > From janisjp naver.com Tue Jun 7 21:18:27 2005 From: janisjp naver.com (=?euc-kr?B?IrHowfi/tSI=?=) Date: Tue Jun 7 07:44:48 2005 Subject: [Mp4-tech] a few Qs about JM9.3 Message-ID: <42A58283.000001.21700@nhn430> Hi, experts! I've got some sample sequences like football_cif_ori90.yuv, akiyo_qcif_ori90.yuv, carphone_qcif_ori90.yuv, foreman_cif_ori90.yuv, and some others and testing the JM9.3 encoder for these sequences. I found something I could not understand. Please help me with these questions. 1. when i set 'Number of B coded frames inserted' as 1, the number of encoded frames is double the number i wrote in the 'encoder.cfg' file. i guess a frame which is not 'I frame' is encoded twice as 'P frame' and 'B frame?' how can i know which(P or B) is used for actual decoding for the current frame, or both are used? 2. Do the sequences like 'football_cif_ori90.yuv' have only one 'I frame?' Why i ask is, when 'IntraPeriod' is 0 in the file 'encoder.cfg', only the first frame is 'I frame' amd I don't know what number i have to put for every other sequences even if they have more 'I frames' than one. 3. 'test_rec.yuv' or 'test_dec.yuv' have half(? or less) frames than the original sequences(before encoded), I mean I wanted encode and decode all 90 frames for the sequence but the result turns out that only the frames much less than 90 are reconstructed or decoded at the end. why is this? 4. i am looking for the SAD value for the best mode in each block. is it 'min_cost' in the 'encode_one_macroblock()' function? -- Thanksss in advance, janis ------------------------------------------------------------------------ ??? :: ?! ???? ?? ??? ??? ?????. http://mail.naver.com/ From may_ank77 yahoo.com Tue Jun 7 08:00:04 2005 From: may_ank77 yahoo.com (mayank agarwal) Date: Tue Jun 7 11:10:31 2005 Subject: [Mp4-tech] playing encoded files Message-ID: <20050607140005.87204.qmail@web52801.mail.yahoo.com> Hi all, I am collected the encoded bitstream in one .m4v file and then trying to play the stream in other decoders. what is happening is that the file is getting played on one decoder which is of the company that has given the code while it is not playing in other decoderes. how is it possible.i want to know whether the generation of decoder.exe which play any encoded stream files is application specific or one decoder can play any other encoded stream provided that the format is supported. Regards, Mayank __________________________________ Discover Yahoo! Find restaurants, movies, travel and more fun for the weekend. Check it out! http://discover.yahoo.com/weekend.html From w_tong2000 yahoo.com Tue Jun 7 10:31:14 2005 From: w_tong2000 yahoo.com (Tong Wu) Date: Tue Jun 7 16:09:59 2005 Subject: [Mp4-tech] questions about the JM.94's direct MB's handling In-Reply-To: Message-ID: <20050607163114.72246.qmail@web50703.mail.yahoo.com> Dear experts, I don't understand why the Jm94 figures out the mv and ref_idx for a Direct partition in readMotionInfoFromNAL() and then does it again in decode_one_macroblock(). Can anyone help me on this? Thank you very much for your help. Tony __________________________________ Discover Yahoo! Use Yahoo! to plan a weekend, have fun online and more. Check it out! http://discover.yahoo.com/ From alexismt comcast.net Tue Jun 7 21:10:05 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Tue Jun 7 16:31:15 2005 Subject: [Mp4-tech] questions about the JM.94's direct MB's handling Message-ID: <060720052010.13033.42A5FF1C000A9DA2000032E922007481849B039C07980A040E@comcast.net> Dear Tong, You may notice that the computation of direct partitions in readMotionInfoFromNal() is only performed for the P8x8 mode. This is a necessary process for handling motion vector prediction between different submacroblock partitions. In decode_one_macroblock() the process is performed for all types of direct modes. Although you could probably avoid redoing this process for P8x8, we have left it for convenience. It is possible however that based on that implementation some "side" information is also generated within this process, although I have to admit this implementation was done way long ago (3 years back). Best regards, Alexis -------------- Original message -------------- > Dear experts, > > I don't understand why the Jm94 figures out the mv and > ref_idx for a Direct partition in > readMotionInfoFromNAL() and then does it again in > decode_one_macroblock(). Can anyone help me on this? > > Thank you very much for your help. > > Tony > > > > > __________________________________ > Discover Yahoo! > Use Yahoo! to plan a weekend, have fun online and more. Check it out! > http://discover.yahoo.com/ > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], > [systems], [general] or another apppropriate identifier to indicate the type of > question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found > at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050607/6b2ea9b6/attachment.html From ssamant iupui.edu Tue Jun 7 16:57:17 2005 From: ssamant iupui.edu (ssamant@iupui.edu) Date: Tue Jun 7 17:52:10 2005 Subject: [Mp4-tech] Transform used in H.264 Message-ID: <1118177837.42a60a2defd38@webmail.iu.edu> Hi I was wondering what type of transform is being used in H.264. Is Integer DCT being used? Is IntDCT the same as integer transform? If not what is the difference between the two? Thank You. Samant From w_tong2000 yahoo.com Tue Jun 7 15:29:46 2005 From: w_tong2000 yahoo.com (Tong Wu) Date: Tue Jun 7 17:58:28 2005 Subject: [Mp4-tech] questions about the JM.94's direct MB's handling In-Reply-To: <060720052010.13033.42A5FF1C000A9DA2000032E922007481849B039C07980A040E@comcast.net> Message-ID: <20050607212946.90223.qmail@web50702.mail.yahoo.com> Dear Alexis, Thanks a lot for your answer? But why is it necessary to handle P8x8 mode direct partitions in readMotionInfoFromNal()? Is it because of the CABAC mode MVD and refIdx context prediction? It seems that the Direct partition's MVD and refIdx are treated as zero when used for the CABAC context prediction. Thank you very much for your help. Tong --- Alexis Michael Tourapis wrote: > Dear Tong, > > You may notice that the computation of direct > partitions in readMotionInfoFromNal() is only > performed for the P8x8 mode. This is a necessary > process for handling motion vector prediction > between different submacroblock partitions. > > In decode_one_macroblock() the process is performed > for all types of direct modes. Although you could > probably avoid redoing this process for P8x8, we > have left it for convenience. It is possible however > that based on that implementation some "side" > information is also generated within this process, > although I have to admit this implementation was > done way long ago (3 years back). > > Best regards, > > Alexis > > -------------- Original message -------------- > > > Dear experts, > > > > I don't understand why the Jm94 figures out the mv > and > > ref_idx for a Direct partition in > > readMotionInfoFromNAL() and then does it again in > > decode_one_macroblock(). Can anyone help me on > this? > > > > Thank you very much for your help. > > > > Tony > > > > > > > > > > __________________________________ > > Discover Yahoo! > > Use Yahoo! to plan a weekend, have fun online and > more. Check it out! > > http://discover.yahoo.com/ > > _______________________________________________ > > NOTE: Please use clear subject lines for your > posts. Include [audio, [video], > > [systems], [general] or another apppropriate > identifier to indicate the type of > > question you have. > > > > Note: Conduct on the mailing list is subject to > the Antitrust guidelines found > > at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php __________________________________ Discover Yahoo! Use Yahoo! to plan a weekend, have fun online and more. Check it out! http://discover.yahoo.com/ From rkoenen intertrust.com Tue Jun 7 15:41:13 2005 From: rkoenen intertrust.com (Rob Koenen) Date: Wed Jun 8 17:30:18 2005 Subject: [Mp4-tech] FW: 10-bit HD data available Message-ID: I don't think this email ever made it to the list, even though it was intended to. Apologies to those who already received it through other lists. Rob -----Original Message----- From: Pankaj Topiwala [mailto:pankaj@fastvdo.com] Sent: Friday, 03 June, 2005 18:48 To: jvt-experts@lists.rwth-aachen.de; mpeg@itscj.ipsj.or.jp; mp4-tech@mpegif.org Cc: Gary Sullivan; Tom Mcmahon; rob.koenen@mpegif.org; hddata@fastvdo.com Subject: 10-bit HD data available Dear Video Experts, Due to demand for HD test data, FastVDO is pleased to provide a consolidated 10-bit HD data set (mostly 1080p) for the research community. Content includes a rich set of both film and non-film data. The data is from a variety of sources, which retain data rights; usage rights are limited to testing, research, standards development, and technical presentation. Please check the site below, where some preliminary information on this is available. www.fastvdo.com/hddata Included are: 1. some brief descriptions, including scene selections, provided by Dolby and FastVDO when this data was first made available to this community (JVT-J039 and JVT-J042), and 2. some instructions for obtaining this data (www.fastvdo.com/hddata/GetHDData.html). More information will be added shortly. Best regards, Pankaj ************************************************************************ * Dr. Pankaj Topiwala Voice: 410-309-6066 President/CEO FastVDO LLC Fax: 410-309-6554 7150 Riverwood Dr., Mobile: 443-538-3782 NEW! Columbia, MD 21046-1245 USA Email: pankaj@fastvdo.com NEW! ************************************************************************ * From alexismt comcast.net Tue Jun 7 23:38:24 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Wed Jun 8 17:30:24 2005 Subject: [Mp4-tech] questions about the JM.94's direct MB's handling Message-ID: <060720052238.18454.42A621E0000694B20000481622070215739B039C07980A040E@comcast.net> Dear Tong, It is important to clarify that many of the concepts existing in the software (and in general concepts in the standard) were added as incremental additions and at various stages, using limited resources and time. If you examine the code, this particular function not only decodes the differential mvs but also calculates the actual final mvs using the predictor motion vector and that is why direct modes were also implemented in this fashion (i.e. code was already there). If I also recall, the CABAC constraints did not exist at the time, at least within the software. This does not mean that this is the best way to implement this (as you have also probably observed), but is instead a "working" implementation that was at the time constrained by the design of the reference software (which could say "a lot" about this design). Best regards, Alexis -------------- Original message -------------- > Dear Alexis, > > Thanks a lot for your answer? But why is it necessary > to handle P8x8 mode direct partitions in > readMotionInfoFromNal()? Is it because of the CABAC > mode MVD and refIdx context prediction? It seems that > the Direct partition's MVD and refIdx are treated as > zero when used for the CABAC context prediction. > > Thank you very much for your help. > > Tong > > --- Alexis Michael Tourapis > wrote: > > > Dear Tong, > > > > You may notice that the computation of direct > > partitions in readMotionInfoFromNal() is only > > performed for the P8x8 mode. This is a necessary > > process for handling motion vector prediction > > between different submacroblock partitions. > > > > In decode_one_macroblock() the process is performed > > for all types of direct modes. Although you could > > probably avoid redoing this process for P8x8, we > > have left it for convenience. It is possible however > > that based on that implementation some "side" > > information is also generated within this process, > > although I have to admit this implementation was > > done way long ago (3 years back). > > > > Best regards, > > > > Alexis > > > > -------------- Original message -------------- > > > > > Dear experts, > > > > > > I don't understand why the Jm94 figures out the mv > > and > > > ref_idx for a Direct partition in > > > readMotionInfoFromNAL() and then does it again in > > > decode_one_macroblock(). Can anyone help me on > > this? > > > > > > Thank you very much for your help. > > > > > > Tony > > > > > > > > > > > > > > > __________________________________ > > > Discover Yahoo! > > > Use Yahoo! to plan a weekend, have fun online and > > more. Check it out! > > > http://discover.yahoo.com/ > > > _______________________________________________ > > > NOTE: Please use clear subject lines for your > > posts. Include [audio, [video], > > > [systems], [general] or another apppropriate > > identifier to indicate the type of > > > question you have. > > > > > > Note: Conduct on the mailing list is subject to > > the Antitrust guidelines found > > > at > > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > > > > > __________________________________ > Discover Yahoo! > Use Yahoo! to plan a weekend, have fun online and more. Check it out! > http://discover.yahoo.com/ > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], > [systems], [general] or another apppropriate identifier to indicate the type of > question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found > at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050607/9be4712a/attachment.html From vikass gmail.com Wed Jun 8 10:30:24 2005 From: vikass gmail.com (Vikas Singhal) Date: Wed Jun 8 17:30:29 2005 Subject: [Mp4-tech] playing encoded files In-Reply-To: <20050607140005.87204.qmail@web52801.mail.yahoo.com> References: <20050607140005.87204.qmail@web52801.mail.yahoo.com> Message-ID: Hi Mayank, You can check the conformity of the encoded file (If it is MPEG-4 Part 2 encoded) using the freeware provided by Venera. You can download it from http://www.venera.co.in/dhruv.asp . If you need more indepth analysis then you can go for one of the many Analyzers available. Regards Vikas On 6/7/05, mayank agarwal wrote: > Hi all, > I am collected the encoded bitstream in one .m4v > file > and then trying to play the stream in other decoders. > what is happening is that the file is getting played > on one decoder which is of the company that has given > the code while it is not playing in other decoderes. > how is it possible.i want to know whether the > generation of decoder.exe which play any encoded > stream > files is application specific or one decoder can play > any other encoded stream provided that the format is > supported. > > Regards, > Mayank > > > > > __________________________________ > Discover Yahoo! > Find restaurants, movies, travel and more fun for the weekend. Check it out! > http://discover.yahoo.com/weekend.html > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > From tuukkat ee.oulu.fi Wed Jun 8 12:48:18 2005 From: tuukkat ee.oulu.fi (Tuukka Toivonen) Date: Wed Jun 8 17:30:35 2005 Subject: [Mp4-tech] Transform used in H.264 In-Reply-To: <1118177837.42a60a2defd38@webmail.iu.edu> References: <1118177837.42a60a2defd38@webmail.iu.edu> Message-ID: On Tue, 7 Jun 2023 ssamant@iupui.edu wrote: > I was wondering what type of transform is being used in H.264. Is Integer DCT > being used? Is IntDCT the same as integer transform? If not what is the > difference between the two? Yes, 4x4 DCT approximated by integers. See http://www.vcodex.com/ in particular http://www.vcodex.com/h264_transform.pdf The standard itself could also be helpful, although I wouldn't recommend it as a starting point. From dmitriy graphics.cs.msu.ru Thu Jun 9 00:26:07 2005 From: dmitriy graphics.cs.msu.ru (Dmitriy Vatolin) Date: Wed Jun 8 17:30:41 2005 Subject: [Mp4-tech] Codec comparison tool Message-ID: <1725310658.20050608232607@graphics.cs.msu.ru> Hello! We (MSU G&M Lab) release video quality metric program: MSU Video Quality Measurement BETA 0.71 http://www.compression.ru/video/quality_measure/video_measurement_tool_en.html Metric supports different input formats (AVI, YUV, BMP) and metrics * PSNR, * Delta, * VQM, * SSIM, * MSU Blurring/Blocking Measures. Main features: results are saved in *.CSV (can be opened everywhere), all metrics are visualized, comparative analysis (comparison of three files: original, and two other, for example processed by different codecs). Program allows to answer on questions: 1) One codec has more blurring than another. On what frames? Average mark? 2) One codec has more blocking than another. On what frames? Average mark? 3) One codec has more lower quality than another. On what frames? Average mark in PSNR, VQM, SSIM? Correspondent "bad frames" are saved automatically into BMP files. Metric home page: http://www.compression.ru/video/quality_measure/video_measurement_tool_en.html Direct link to program: http://www.compression.ru/video/quality_measure/src/msu_vmt.zip ANY COMMENT AND IDEAS ABOUT FUTURE METRIC DEVELOPMENT ARE WELCOME! :) -- Best regards, Dmitriy mailto:dmitriy@graphics.cs.msu.ru From hkl0220 gmail.com Wed Jun 8 20:36:54 2005 From: hkl0220 gmail.com (hkl@GMail) Date: Thu Jun 9 09:25:37 2005 Subject: [Mp4-tech] [H.264] Reference Frames for P-frame and B-frame Message-ID: <98251ba05060817367b1e18fe@mail.gmail.com> Dear Experts, Could anyone kindly tell me how does h.264 decide the reference frames for P-frames and B-frames? Best, From w_tong2000 yahoo.com Wed Jun 8 19:43:14 2005 From: w_tong2000 yahoo.com (Tong Wu) Date: Thu Jun 9 09:30:18 2005 Subject: [Mp4-tech] questions about the JM.94's direct MB's handling In-Reply-To: <060720052238.18454.42A621E0000694B20000481622070215739B039C07980A040E@comcast.net> Message-ID: <20050609014314.12974.qmail@web50708.mail.yahoo.com> Dear Alexis, I have another question regarding the Direct MB. Please see table 8-8. Since mbAddrCol3 is in an AFRM field with the same parity as the current field picture (and current MB), and the current field picture's list 1 ref_idx 0 may be a AFRM field with different parity as the current field picture, is it possible that the mbAddrCol3 actually is not in the AFRM field that is used for the pixel interpolation? Thanks. Tong --- Alexis Michael Tourapis wrote: > Dear Tong, > > It is important to clarify that many of the concepts > existing in the software (and in general concepts in > the standard) were added as incremental additions > and at various stages, using limited resources and > time. If you examine the code, this particular > function not only decodes the differential mvs but > also calculates the actual final mvs using the > predictor motion vector and that is why direct modes > were also implemented in this fashion (i.e. code was > already there). If I also recall, the CABAC > constraints did not exist at the time, at least > within the software. This does not mean that this is > the best way to implement this (as you have also > probably observed), but is instead a "working" > implementation that was at the time constrained by > the design of the reference software (which could > say "a lot" about this design). > > Best regards, > > Alexis > > > -------------- Original message -------------- > > > Dear Alexis, > > > > Thanks a lot for your answer? But why is it > necessary > > to handle P8x8 mode direct partitions in > > readMotionInfoFromNal()? Is it because of the > CABAC > > mode MVD and refIdx context prediction? It seems > that > > the Direct partition's MVD and refIdx are treated > as > > zero when used for the CABAC context prediction. > > > > Thank you very much for your help. > > > > Tong > > > > --- Alexis Michael Tourapis > > wrote: > > > > > Dear Tong, > > > > > > You may notice that the computation of direct > > > partitions in readMotionInfoFromNal() is only > > > performed for the P8x8 mode. This is a necessary > > > > process for handling motion vector prediction > > > between different submacroblock partitions. > > > > > > In decode_one_macroblock() the process is > performed > > > for all types of direct modes. Although you > could > > > probably avoid redoing this process for P8x8, we > > > > have left it for convenience. It is possible > however > > > that based on that implementation some "side" > > > information is also generated within this > process, > > > although I have to admit this implementation was > > > > done way long ago (3 years back). > > > > > > Best regards, > > > > > > Alexis > > > > > > -------------- Original message -------------- > > > > > > > Dear experts, > > > > > > > > I don't understand why the Jm94 figures out > the mv > > > and > > > > ref_idx for a Direct partition in > > > > readMotionInfoFromNAL() and then does it again > in > > > > decode_one_macroblock(). Can anyone help me on > > > > this? > > > > > > > > Thank you very much for your help. > > > > > > > > Tony > > > > > > > > > > > > > > > > > > > > __________________________________ > > > > Discover Yahoo! > > > > Use Yahoo! to plan a weekend, have fun online > and > > > more. Check it out! > > > > http://discover.yahoo.com/ > > > > > _______________________________________________ > > > > NOTE: Please use clear subject lines for your > > > posts. Include [audio, [video], > > > > [systems], [general] or another apppropriate > > > identifier to indicate the type of > > > > question you have. > > > > > > > > Note: Conduct on the mailing list is subject > to > > > the Antitrust guidelines found > > > > at > > > > > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > > > > > > > > > > > > __________________________________ > > Discover Yahoo! > > Use Yahoo! to plan a weekend, have fun online and > more. Check it out! > > http://discover.yahoo.com/ > > _______________________________________________ > > NOTE: Please use clear subject lines for your > posts. Include [audio, [video], > > [systems], [general] or another apppropriate > identifier to indicate the type of > > question you have. > > > > Note: Conduct on the mailing list is subject to > the Antitrust guidelines found > > at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > _______________________________________________ > NOTE: Please use clear subject lines for your posts. > Include [audio, [video], [systems], [general] or > another apppropriate identifier to indicate the type > of question you have. > > Note: Conduct on the mailing list is subject to the > Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From Gilberto ensilica.com Thu Jun 9 13:28:58 2005 From: Gilberto ensilica.com (Gilberto Rodriguez) Date: Thu Jun 9 09:30:24 2005 Subject: [Mp4-tech] [MPEG-4] [h264] reverence model from videolan or from http://ip.hhi.de/organisation.htm Message-ID: Hi all, Can anybody tell me what are the differences on the reference code that you can get from videolan and HHI? Is any of them more reliable than the other? Thank you, Gilberto ====================================== Gilberto Rodriguez - IC Designer EnSilica Limited 2nd floor Cyberhouse Molly Millar's Lane Mobile: +44(0)7736 619 511 / +34 619 529 825 Wokingham Tel: +44(0)1189 798 159 Berkshire Fax: +44(0)1189 798 160 RG41 2PX www.ensilica.com ====================================== -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050609/2678e9e9/attachment.html From pankaj0iit yahoo.co.in Thu Jun 9 22:05:28 2005 From: pankaj0iit yahoo.co.in (pankaj bansal) Date: Fri Jun 10 12:12:58 2005 Subject: [Mp4-tech] test streams for h264 decoder Message-ID: <20050610040529.72105.qmail@web8408.mail.in.yahoo.com> Hi, I am trying to generate some test vectors with the help of Reference JM Encoder. When I make pic_interlace value = 2, it's allways encoding the pictures as frame, but i want that it's should encode some pictures as frame and some as fields. I have tried some test input as football, stefan, tempete. Could you please tell me that whitch test input is required for generating that perticular test stream? and where are they avaiable? or if possible you can send me those kind of encoded test streams. thank you pankaj bansal --------------------------------- Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050609/967c6531/attachment-0001.html From sandipkr dgbmicro.com Fri Jun 10 11:21:50 2005 From: sandipkr dgbmicro.com (Sandip Ray) Date: Fri Jun 10 12:20:35 2005 Subject: [Mp4-tech] [video]Test sequence Message-ID: Hi, I need Akiyo 4:2:0 YUV video test sequence(300 frames). Can anyone let me know the link where I can find it ? Regards, Sandip -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050610/c1751ccb/attachment.html From abdessamad.falhi epfl.ch Fri Jun 10 12:34:29 2005 From: abdessamad.falhi epfl.ch (abdessamad.falhi@epfl.ch) Date: Fri Jun 10 12:27:06 2005 Subject: [Mp4-tech] AAC decoder with fixed point Message-ID: <1118396069.42a95ea5333ce@imapwww.epfl.ch> Hi, Has anybody implemted a fixed point version of an AAC decoder? What are the sensitve parts where we must be careful when we want to skip from float to int? with regards, Abdessamad From alexismt comcast.net Fri Jun 10 10:04:04 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Sat Jun 11 04:00:57 2005 Subject: [Mp4-tech] test streams for h264 decoder In-Reply-To: <20050610040529.72105.qmail@web8408.mail.in.yahoo.com> Message-ID: <200506101604.j5AG483A011675@lists1.magma.ca> Dear Pankaj Is your content in interlace format? You may be able to find some interlace sequences (although in 4:2:2 format) on the VQEG site. Alexis _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of pankaj bansal Sent: Thursday, June 09, 2023 9:05 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] test streams for h264 decoder Hi, I am trying to generate some test vectors with the help of Reference JM Encoder. When I make pic_interlace value = 2, it's allways encoding the pictures as frame, but i want that it's should encode some pictures as frame and some as fields. I have tried some test input as football, stefan, tempete. Could you please tell me that whitch test input is required for generating that perticular test stream? and where are they avaiable? or if possible you can send me those kind of encoded test streams. thank you pankaj bansal _____ Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050610/75d67d5a/attachment.html From alexismt comcast.net Fri Jun 10 10:07:00 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Sat Jun 11 04:08:56 2005 Subject: [Mp4-tech] questions about the JM.94's direct MB's handling In-Reply-To: <20050609014314.12974.qmail@web50708.mail.yahoo.com> Message-ID: <200506101607.j5AG743A011979@lists1.magma.ca> Dear Tong, Yes you are correct on your observation. Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Tong Wu Sent: Wednesday, June 08, 2023 6:43 PM To: Alexis Michael Tourapis; mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] questions about the JM.94's direct MB's handling Dear Alexis, I have another question regarding the Direct MB. Please see table 8-8. Since mbAddrCol3 is in an AFRM field with the same parity as the current field picture (and current MB), and the current field picture's list 1 ref_idx 0 may be a AFRM field with different parity as the current field picture, is it possible that the mbAddrCol3 actually is not in the AFRM field that is used for the pixel interpolation? Thanks. Tong --- Alexis Michael Tourapis wrote: > Dear Tong, > > It is important to clarify that many of the concepts existing in the > software (and in general concepts in the standard) were added as > incremental additions and at various stages, using limited resources > and time. If you examine the code, this particular function not only > decodes the differential mvs but also calculates the actual final mvs > using the predictor motion vector and that is why direct modes were > also implemented in this fashion (i.e. code was already there). If I > also recall, the CABAC constraints did not exist at the time, at least > within the software. This does not mean that this is the best way to > implement this (as you have also probably observed), but is instead a > "working" > implementation that was at the time constrained by the design of the > reference software (which could say "a lot" about this design). > > Best regards, > > Alexis > > > -------------- Original message -------------- > > > Dear Alexis, > > > > Thanks a lot for your answer? But why is it > necessary > > to handle P8x8 mode direct partitions in readMotionInfoFromNal()? Is > > it because of the > CABAC > > mode MVD and refIdx context prediction? It seems > that > > the Direct partition's MVD and refIdx are treated > as > > zero when used for the CABAC context prediction. > > > > Thank you very much for your help. > > > > Tong > > > > --- Alexis Michael Tourapis > > wrote: > > > > > Dear Tong, > > > > > > You may notice that the computation of direct partitions in > > > readMotionInfoFromNal() is only performed for the P8x8 mode. This > > > is a necessary > > > > process for handling motion vector prediction between different > > > submacroblock partitions. > > > > > > In decode_one_macroblock() the process is > performed > > > for all types of direct modes. Although you > could > > > probably avoid redoing this process for P8x8, we > > > > have left it for convenience. It is possible > however > > > that based on that implementation some "side" > > > information is also generated within this > process, > > > although I have to admit this implementation was > > > > done way long ago (3 years back). > > > > > > Best regards, > > > > > > Alexis > > > > > > -------------- Original message -------------- > > > > > > > Dear experts, > > > > > > > > I don't understand why the Jm94 figures out > the mv > > > and > > > > ref_idx for a Direct partition in > > > > readMotionInfoFromNAL() and then does it again > in > > > > decode_one_macroblock(). Can anyone help me on > > > > this? > > > > > > > > Thank you very much for your help. > > > > > > > > Tony > > > > > > > > > > > > > > > > > > > > __________________________________ > > > > Discover Yahoo! > > > > Use Yahoo! to plan a weekend, have fun online > and > > > more. Check it out! > > > > http://discover.yahoo.com/ > > > > > _______________________________________________ > > > > NOTE: Please use clear subject lines for your > > > posts. Include [audio, [video], > > > > [systems], [general] or another apppropriate > > > identifier to indicate the type of > > > > question you have. > > > > > > > > Note: Conduct on the mailing list is subject > to > > > the Antitrust guidelines found > > > > at > > > > > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > > > > > > > > > > > > __________________________________ > > Discover Yahoo! > > Use Yahoo! to plan a weekend, have fun online and > more. Check it out! > > http://discover.yahoo.com/ > > _______________________________________________ > > NOTE: Please use clear subject lines for your > posts. Include [audio, [video], > > [systems], [general] or another apppropriate > identifier to indicate the type of > > question you have. > > > > Note: Conduct on the mailing list is subject to > the Antitrust guidelines found > > at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > _______________________________________________ > NOTE: Please use clear subject lines for your posts. > Include [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From harish.vasudeva amd.com Fri Jun 10 16:49:06 2005 From: harish.vasudeva amd.com (Vasudeva, Harish) Date: Sat Jun 11 04:15:37 2005 Subject: [Mp4-tech] Number Of B-Frames Message-ID: <4A77E4BC05A4EF4D9DAA4C8CB5EEE51CBA3363@SAUSEXMB1.amd.com> Hi, Is there a way to tell (at the VOL level) the consecutive number of B-frames in a sequence? Best Regards, HARISH V -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050610/891d2222/attachment.html From lists diversedev.com Fri Jun 10 15:49:13 2005 From: lists diversedev.com (Justin Todd) Date: Sat Jun 11 04:22:13 2005 Subject: [Mp4-tech] What encoding information do I need to decode an MPEG4 stream References: <1725310658.20050608232607@graphics.cs.msu.ru> Message-ID: <002101c56e06$4593a410$6601a8c0@jtlaptop> Hello. I'm still very new to MPEG4; my knowledge is very limited, so please excuse my vagueness over-simplification of the problem. The problem: We have a hardware MPEG4 encoder with a matching decoder unit. We would like to be able to connect a PC with our own custom software (which emulates the decoder) to the encoder unit. Ive figured out a lot of the control channel packets and I'm able to get the encoder to send me an MPEG4 stream, but i am not entirely sure how to enterpret it. In the encoders setting menu the following options can be configured: Minimum frame distance, Encoder bitrate, I quantizer, P quantizer, I frame distance, I update distance, frame size (res), width, height If an decoder (lets say xvid) knows these parameters, can it safely decode a stream provided the encoder adheres to the standard? From gaetanwillems gmail.com Sun Jun 12 01:16:50 2005 From: gaetanwillems gmail.com (=?ISO-8859-1?Q?Ga=EBtan_Willems?=) Date: Sun Jun 12 19:24:11 2005 Subject: [Mp4-tech] [video] motion estimation Message-ID: <56c09c340506111516525d2ca7@mail.gmail.com> Hello experts, I'm a beginner in the video domain and i'm working on slow motion. I need to create a program that interpolate new frames between the existing frames of the video in order to obtain slow motion. To do that, I'm thinking about processing the best motion estimation possible from the original frames and then use the motion vectors to interpolate the new frames. I downloaded several codec sources: MPEG-1 (Berkeley) MPEG-2 (MSSG) H.264 (x264, JM 9.5, JVT 8.2) but I'm really lost in all the files. Could you tell me which codec is best to look at for motion estimation and if possible which files should I consider? Is there any place where I could find exclusively the motion estimation part of the code? Do you think that working like that on motion vectors is a good idea or would you go a different way? Thanks a lot in advance Best regards Ga?tan From gthm159 yahoo.co.in Mon Jun 13 07:22:05 2005 From: gthm159 yahoo.co.in (Googie) Date: Mon Jun 13 16:14:11 2005 Subject: [Mp4-tech] MPEG4 resync marker In-Reply-To: <200506111611.j5BG5fYQ003144@lists1.magma.ca> Message-ID: <20050613052205.91727.qmail@web53704.mail.yahoo.com> Hi, I have learnt from a friend that the resync marker in case of MPEG4 streams is, 0000 0000 0000 0000 1xxx xxxx Is this correct? Also, is the resync marker byte-aligned. Right now, my objective is to packetize a long MPEG4 bitstream into multiple RTP packets. Therefore, I thought it would be wise to split the input stream at resync markers. Hence, I would appreciate it if you could confirm that the resync marker I'm using is the right one. Thanks in advance, Googie --------------------------------- Too much spam in your inbox? Yahoo! Mail gives you the best spam protection for FREE! http://in.mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050613/2177a5eb/attachment.html From tuukkat ee.oulu.fi Mon Jun 13 12:11:09 2005 From: tuukkat ee.oulu.fi (Tuukka Toivonen) Date: Mon Jun 13 16:20:42 2005 Subject: [Mp4-tech] [video]Test sequence In-Reply-To: References: Message-ID: On Fri, 10 Jun 2005, Sandip Ray wrote: > I need Akiyo 4:2:0 YUV video test sequence(300 frames). Can anyone let > me know the link where I can find it ? Here is a number of test sequences: http://media.xiph.org/video/derf/ From sagar sarayusoftech.com Mon Jun 13 15:08:23 2005 From: sagar sarayusoftech.com (sagar) Date: Mon Jun 13 16:26:08 2005 Subject: [Mp4-tech] RE: Welcome to the "Mp4-tech" mailing list In-Reply-To: Message-ID: Hi, I am new to this list. I am working on MPEG 4. I want some m4v player. If anybody can even send the same as an attachment, that will be still better. Regards, Sagar -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of mp4-tech-request@lists.mpegif.org Sent: Monday, June 13, 2023 1:05 PM To: sagar@sarayusoftech.com Subject: Welcome to the "Mp4-tech" mailing list Welcome to the Mp4-tech@lists.mpegif.org mailing list! To post to this list, send your email to: mp4-tech@lists.mpegif.org General information about the mailing list is at: http://lists.mpegif.org/mailman/listinfo/mp4-tech If you ever want to unsubscribe or change your options (eg, switch to or from digest mode, change your password, etc.), visit your subscription page at: http://lists.mpegif.org/mailman/options/mp4-tech/sagar%40sarayusoftech.com You can also make such adjustments via email by sending a message to: Mp4-tech-request@lists.mpegif.org with the word `help' in the subject or body (don't include the quotes), and you will get back a message with instructions. You must know your password to change your options (including changing the password, itself) or to unsubscribe. It is: sag007 Normally, Mailman will remind you of your lists.mpegif.org mailing list passwords once every month, although you can disable this if you prefer. This reminder will also include instructions on how to unsubscribe or change your account options. There is also a button on your options page that will email your current password to you. From kaustubh.patankar vsnl.net Mon Jun 13 15:26:34 2005 From: kaustubh.patankar vsnl.net (kaustubh.patankar@vsnl.net) Date: Mon Jun 13 16:30:39 2005 Subject: [Mp4-tech] [video] motion estimation Message-ID: <303c57f303b286.303b286303c57f@vsnl.net> Dear Gaetan, As per my view it is better to look into MPEG2 code of encoder for motion estimation. And decoder for motion compensation. Also there are lots of Text books available and the information on web reagrding the motion estimation. For your reference the motion estimation in MPEG 2 encoder is as per the full search algorithm. For implementation initially you can go for 3, step, Nearest neighbourhood algorithm. with regards Kaustubh ----- Original Message ----- From: Ga?tan Willems Date: Sunday, June 12, 2023 3:46 am Subject: [Mp4-tech] [video] motion estimation > Hello experts, > > I'm a beginner in the video domain and i'm working on slow motion. I > need to create a program that interpolate new frames between the > existing frames of the video in order to obtain slow motion. > > To do that, I'm thinking about processing the best motion estimation > possible from the original frames and then use the motion vectors to > interpolate the new frames. > > I downloaded several codec sources: MPEG-1 (Berkeley) MPEG-2 (MSSG) > H.264 (x264, JM 9.5, JVT 8.2) but I'm really lost in all the files. > Could you tell me which codec is best to look at for motion estimation > and if possible which files should I consider? > Is there any place where I could find exclusively the motion > estimation part of the code? > Do you think that working like that on motion vectors is a good idea > or would you go a different way? > > Thanks a lot in advance > > Best regards > > Ga?tan > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042- > Antitrust.php From andrewk vbrick.com Mon Jun 13 09:05:05 2005 From: andrewk vbrick.com (Andrew Krupiczka) Date: Mon Jun 13 16:30:45 2005 Subject: [Mp4-tech] [video] motion estimation Message-ID: Gaetan, I'm not sure whether looking at standardized video codecs motion estimation technique is your best first shot for what you intend to do. Please note, that mpeg family codecs are based on so called "block matching" techniques and for such motion vectors describe displacements of entire blocks of pixels (in the case of H.264 the smallest size is 4x4 block of pixels) and not a motion of single pixels (or at least 2x2 blocks) I'd be interested in if exploring temporal interpolation subject. In addition, motion vectors from video codecs try to minimize not only SADs of entire blocks but so called total motion cost (which also considers a cost of encoding for particular MV) and thus not always represent a "true motion" which I think is relevant for quality temporal interpolation. My suggestions would be: 1) you might want to research rather video deinterlacing techniques (e.g. pel-recursive, optical-flow, phase correlation) incl. also ones used by chip vendors like lsi, zoran, genesis, siliconimage instead of video codec motion estimation techniques 2) for a nice coverage of subject you might want to study book of G. de Haan, `Video processing for multimedia systems', ISBN: 90-9014015-8, Eindhoven Sep. 2000. (The book comes with great demo software with an easy to use Windows interface, enabling you to process video sequences, on your PC, with the algorithms described in the book) http://www.ics.ele.tue.nl/~dehaan/publications.html 3) some more reference links of interest http://www.hometheaterhifi.com/volume_7_4/dvd-benchmark-part-5-progressive-10-2000.html http://www.teranex.com/resources/whitepapers.htm http://www.semiconductors.philips.com/pip/SAA4994.html Best regards, Andrew -----Original Message----- From: Ga?tan Willems [mailto:gaetanwillems@gmail.com] Sent: Saturday, June 11, 2023 6:17 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [video] motion estimation Hello experts, I'm a beginner in the video domain and i'm working on slow motion. I need to create a program that interpolate new frames between the existing frames of the video in order to obtain slow motion. To do that, I'm thinking about processing the best motion estimation possible from the original frames and then use the motion vectors to interpolate the new frames. I downloaded several codec sources: MPEG-1 (Berkeley) MPEG-2 (MSSG) H.264 (x264, JM 9.5, JVT 8.2) but I'm really lost in all the files. Could you tell me which codec is best to look at for motion estimation and if possible which files should I consider? Is there any place where I could find exclusively the motion estimation part of the code? Do you think that working like that on motion vectors is a good idea or would you go a different way? Thanks a lot in advance Best regards Ga?tan _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From koen.dewolf ugent.be Mon Jun 13 15:20:45 2005 From: koen.dewolf ugent.be (Koen De Wolf) Date: Mon Jun 13 16:30:50 2005 Subject: [Mp4-tech] [video] motion estimation In-Reply-To: <56c09c340506111516525d2ca7@mail.gmail.com> Message-ID: <20050613121738.B55F62BAD8E@mail.elis.UGent.be> Hi Ga?tan, I'm not familiar with slow motion, but I doubt that motion vectors are really useful to interpolate frames as motion vectors are used to minimize the prediction error and not to model the motion of objects. However, a motion estimation algorithm that is somehow capable of generating aligned motion vectors, might do the trick. Please do not forget that you also need the residuals in order to reconstruct a frame, and you probably need to interpolate them as well. Best regards, Koen De Wolf -- Multimedia Lab Ghent University - ELIS - IBBT URL http://multimedialab.elis.ugent.be > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Ga?tan Willems > Sent: zondag 12 juni 2005 0:17 > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] [video] motion estimation > > Hello experts, > > I'm a beginner in the video domain and i'm working on slow > motion. I need to create a program that interpolate new > frames between the existing frames of the video in order to > obtain slow motion. > > To do that, I'm thinking about processing the best motion > estimation possible from the original frames and then use the > motion vectors to interpolate the new frames. > > I downloaded several codec sources: MPEG-1 (Berkeley) MPEG-2 (MSSG) > H.264 (x264, JM 9.5, JVT 8.2) but I'm really lost in all the files. > Could you tell me which codec is best to look at for motion > estimation and if possible which files should I consider? > Is there any place where I could find exclusively the motion > estimation part of the code? > Do you think that working like that on motion vectors is a > good idea or would you go a different way? > > Thanks a lot in advance > > Best regards > > Ga?tan > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti > trust.php > From shakesp gmail.com Mon Jun 13 15:19:19 2005 From: shakesp gmail.com (shakesp) Date: Mon Jun 13 16:30:56 2005 Subject: [Mp4-tech] [H.264] Data Partitioning Message-ID: Hi, I have a question for you. I'm using the H.264 JM96 coder. If I want to simulate in the decoder the loss of all data information in the partition C, how can I do? Is there a flag that I can use to communicate to the decoder the loss of data partition C? Thanks in advance From vinay.chiniwar iap-online.net Tue Jun 14 13:23:53 2005 From: vinay.chiniwar iap-online.net (Vinay Chiniwar) Date: Tue Jun 14 01:41:42 2005 Subject: [Mp4-tech] Media File --> Raw File Convertion ( AVI or other media files --> YUV420/422 or M4V) Message-ID: Hi, I have player running on my system supports only YUV420 and YUV422 format and both of them are raw video formats. My player is able to play attached m4v files which is also a raw format. * How m4v format is different from YUV420/YUV422 format * How can I create YUV420/YUV422 format files from other media formats like wma,avi,vob,mov,mp4 ... Now I want add more clips to my system either m4v or YUV420/422 .. I did convert the mp4 format files to m4v format using mp4UI tool (http://mp4ui.sourceforge.net/) but then my player is not able to play converted m4v format files returns with invalid file format. * How can I check the header information of the converted m4v file ( fps,bit rate, video size..etc) * Is there any other tool which converts into m4v / yuv format * How Can I test play m4v or YUV format files under Windows System Please guide me to right direction... Regards, Vinay From smithajo gmail.com Tue Jun 14 10:52:01 2005 From: smithajo gmail.com (Smitha Joseph) Date: Tue Jun 14 01:41:48 2005 Subject: [Mp4-tech] [H.264]Regarding Spatial direct luma motion vector Message-ID: <7bca48b405061321224435840e@mail.gmail.com> Hi All, I have a doubt regarding the derivation of spatial direct luma motion vector. The decoding process specifies that for both B_Direct_16x16 and B_Direct_8x8 the neighboring motion vectors mvLXN and reference indices refIdxLXN with N being replaced by A, B, C are derived with MbpartIdx = 0 and subMbpartIdx = 0. This would mean that for the B_Direct_8x8 sub macroblock type, we are not using the reference indices and motion vectors of the neighboring sub-macroblock partitions but that of the macroblocks around the macroblock to which this B_direct _8x8 macroblock belongs to. Could you please explain the reason for this. Thanks, Smitha Joseph From vikass gmail.com Tue Jun 14 11:52:28 2005 From: vikass gmail.com (Vikas Singhal) Date: Tue Jun 14 04:51:57 2005 Subject: [Mp4-tech] RE: Welcome to the "Mp4-tech" mailing list In-Reply-To: References: Message-ID: Try Dhruv freeware from www.venera.co.in. It is capable of playback of elementary stream with some basic analysis. Vikas On 6/13/05, sagar wrote: > Hi, > > I am new to this list. I am working on MPEG 4. > I want some m4v player. > > If anybody can even send the same as an attachment, that will be still > better. > > > Regards, > Sagar > > > > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of > mp4-tech-request@lists.mpegif.org > Sent: Monday, June 13, 2023 1:05 PM > To: sagar@sarayusoftech.com > Subject: Welcome to the "Mp4-tech" mailing list > > > Welcome to the Mp4-tech@lists.mpegif.org mailing list! > > To post to this list, send your email to: > > mp4-tech@lists.mpegif.org > > General information about the mailing list is at: > > http://lists.mpegif.org/mailman/listinfo/mp4-tech > > If you ever want to unsubscribe or change your options (eg, switch to > or from digest mode, change your password, etc.), visit your > subscription page at: > > http://lists.mpegif.org/mailman/options/mp4-tech/sagar%40sarayusoftech.com > > > You can also make such adjustments via email by sending a message to: > > Mp4-tech-request@lists.mpegif.org > > with the word `help' in the subject or body (don't include the > quotes), and you will get back a message with instructions. > > You must know your password to change your options (including changing > the password, itself) or to unsubscribe. It is: > > sag007 > > Normally, Mailman will remind you of your lists.mpegif.org mailing > list passwords once every month, although you can disable this if you > prefer. This reminder will also include instructions on how to > unsubscribe or change your account options. There is also a button on > your options page that will email your current password to you. > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > From vishalwc rediffmail.com Tue Jun 14 14:06:50 2005 From: vishalwc rediffmail.com (vishal ) Date: Tue Jun 14 16:10:49 2005 Subject: [Mp4-tech] ISO/IEC AAC- LC vs AAC core of 3gpp enhanced AAC plus Message-ID: <20050614130822.17599.qmail@webmail7.rediffmail.com> ?Hi All, I am currently studying the Enhanced aacPlus encoder reference source code from 3gpp . I am finding diffrences in ISO/IEC 13818-7:2003 (MPEG-2 AAC)encoder as compare to 3gpp (enhanced AAC plus V1 )AAC core . I have an existing ISO/IEC AAC-LC running on my platform . what will be the issue related to licensing and quality when i reuse ISO/IEC AAC-LC as 3gpp enhanced AAC plus V1 AAC core . Thanks & Regards Vishal -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050614/9d2d3e3f/attachment.html From choquito_tf hotmail.com Wed Jun 15 14:34:06 2005 From: choquito_tf hotmail.com (james jami) Date: Thu Jun 16 08:30:21 2005 Subject: [Mp4-tech] signed QP in the H264 reference code? Message-ID: Hello, I have realized that in the function biari_init_context of the binaridecod.c file (JM93), the Quantification Parameter qp is used as a signed integer (due the use of the max(0,img->qp) ). But in the standard qp only can be a number between 0 and 51. Is there a hidden reason for this max operation or I can remove it? Thanks in advance for the answer, Carlos From pankaj0iit yahoo.co.in Wed Jun 15 07:13:07 2005 From: pankaj0iit yahoo.co.in (pankaj bansal) Date: Thu Jun 16 08:30:28 2005 Subject: [Mp4-tech] problem with H.264 Encoder JM.95 Message-ID: <20050615131307.57028.qmail@web8409.mail.in.yahoo.com> Hi, Please see the Table A-3 and A-4 in the standard Draft ITU-T Rec.H.264(2002E), 8th meeting :Geneva, Switzerland, 23-27 May, 2003. I have strong feeling that JM.95 Encoder doesn't consider of generating errors when level is (below 2.1 or above 4.1) and Pic_interlace flag (in encoder config) is 1 or 2 in main and extended profiles. I mean that If i set the values for profile == Main or Extended , level = 1.2(<2.1 or >4.1) and pic_interlace = 1 or 2; then Encoder should give error. But currently it's not happening. Please share your views. thanks pankaj bansal --------------------------------- Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050615/abfcaab5/attachment.html From gops98 yahoo.com Wed Jun 15 08:03:48 2005 From: gops98 yahoo.com (Gopalakrishnan Srinivasan) Date: Thu Jun 16 08:30:34 2005 Subject: [Mp4-tech] YUV422 compressed data Message-ID: <20050615140348.25929.qmail@web54204.mail.yahoo.com> Hi all, I am looking for YUV422 uncompressed video data. May I know where can we get that. Thanks, SGK __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050615/24da1191/attachment.html From gops98 yahoo.com Wed Jun 15 08:03:55 2005 From: gops98 yahoo.com (Gopalakrishnan Srinivasan) Date: Thu Jun 16 08:30:39 2005 Subject: [Mp4-tech] YUV422 uncompressed data Message-ID: <20050615140355.44378.qmail@web54202.mail.yahoo.com> Hi all, I am looking for YUV422 uncompressed video data. May I know where can we get that. Thanks, SGK --------------------------------- Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050615/a4189f8f/attachment.html From tomar_atul2000 yahoo.com Thu Jun 16 08:15:22 2005 From: tomar_atul2000 yahoo.com (atul tomar) Date: Fri Jun 17 06:54:19 2005 Subject: [Mp4-tech] About map unit to slice group map. Message-ID: <20050616141522.70065.qmail@web52310.mail.yahoo.com> Hi All, Can anyone please explain me what standard mean by " map unit to slice group map". I read ITU standard many times to understand this but could not get it clear. Please expalin what does "map unit" represent. Thanks in advance. Reg, Atul --------------------------------- Discover Yahoo! Find restaurants, movies, travel & more fun for the weekend. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050616/95a87e1d/attachment.html From g.lakshmankumar gmail.com Thu Jun 16 21:05:42 2005 From: g.lakshmankumar gmail.com (Lakshman Kumar) Date: Fri Jun 17 07:02:05 2005 Subject: [Mp4-tech] Re: [Mp4-tech ]Video encoding h.264 In-Reply-To: <20050615140355.44378.qmail@web54202.mail.yahoo.com> References: <20050615140355.44378.qmail@web54202.mail.yahoo.com> Message-ID: <138aa52f0506160735289919d2@mail.gmail.com> Hi all during h.264 video eccoding prefix and suffix variables are used . can i know them their use with respect to implementatiom On 6/15/05, Gopalakrishnan Srinivasan wrote: > Hi all, > I am looking for YUV422 uncompressed video data. May I know where > can we get that. > > Thanks, > SGK > > ________________________________ > Discover Yahoo! > Stay in touch with email, IM, photo sharing & more. Check it out! > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > -- lakshman kumar From alexismt comcast.net Thu Jun 16 10:03:32 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Fri Jun 17 07:07:59 2005 Subject: [Mp4-tech] problem with H.264 Encoder JM.95 In-Reply-To: <20050615131307.57028.qmail@web8409.mail.in.yahoo.com> Message-ID: <200506161603.j5GG3cT9025121@lists1.magma.ca> Dear Pankaj, You are correct, the checking of levels and profiles is still not fully properly handled in the reference software. Will see if we could add something on that soon. Best regards Alexis _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of pankaj bansal Sent: Wednesday, June 15, 2023 6:13 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] problem with H.264 Encoder JM.95 Hi, Please see the Table A-3 and A-4 in the standard Draft ITU-T Rec.H.264(2002E), 8th meeting :Geneva, Switzerland, 23-27 May, 2003. I have strong feeling that JM.95 Encoder doesn't consider of generating errors when level is (below 2.1 or above 4.1) and Pic_interlace flag (in encoder config) is 1 or 2 in main and extended profiles. I mean that If i set the values for profile == Main or Extended , level = 1.2(<2.1 or >4.1) and pic_interlace = 1 or 2; then Encoder should give error. But currently it's not happening. Please share your views. thanks pankaj bansal _____ Discover Yahoo! Stay in touch with email, IM, photo sharing & more. Check it out! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050616/f2ad651d/attachment.html From gops98 yahoo.com Thu Jun 16 16:39:19 2005 From: gops98 yahoo.com (Gopalakrishnan Srinivasan) Date: Fri Jun 17 07:13:59 2005 Subject: [Mp4-tech] y4m video format In-Reply-To: <20050615140355.44378.qmail@web54202.mail.yahoo.com> Message-ID: <20050616223919.82571.qmail@web54208.mail.yahoo.com> Hi all, I am to trying to get how data is stored in *.y4m uncompressed data. May I know where can I get documentation of this video format. Thanks & Regards, SGK --------------------------------- Do you Yahoo!? Yahoo! Mail - Find what you need with new enhanced search. Learn more. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050616/494e9b26/attachment.html From magarwal neomagic.com Fri Jun 17 14:16:16 2005 From: magarwal neomagic.com (Mohit Agarwal) Date: Fri Jun 17 07:50:52 2005 Subject: [Mp4-tech] [Mpeg4+AAC] Message-ID: <42B27FC8.4020308@neomagic.com> Hi, Can anybody in the group guide me on how to write mpeg4 stream and AAC stream in an AVI file. Is there a freely downloadable tool available to do the same. Best Regards Mohit From tuukkat ee.oulu.fi Fri Jun 17 16:06:50 2005 From: tuukkat ee.oulu.fi (Tuukka Toivonen) Date: Fri Jun 17 08:30:20 2005 Subject: [Mp4-tech] y4m video format In-Reply-To: <20050616223919.82571.qmail@web54208.mail.yahoo.com> References: <20050616223919.82571.qmail@web54208.mail.yahoo.com> Message-ID: On Thu, 16 Jun 2005, Gopalakrishnan Srinivasan wrote: > I am to trying to get how data is stored in *.y4m uncompressed >data. May I know where can I get documentation of this video format. (Maybe you're using the streams from Xiph's site http://media.xiph.org/) As the above page says, the format is used by mjpegtools. http://mjpeg.sourceforge.net/ Maybe there is some documentation, or if not, at least code for reading the files. As far as I know, .y4m files have a short header in ASCII, followed by uncompressed 4:2:0 sampled frames. Between the frames there's probably an ASCII header as well. You could convert the .y4m files into uncompressed headerless yuv files with Mencoder: mencoder /video/stefan_sif.y4m -ovc raw -of rawvideo -vf format=i420 -o /tmp/xx.yuv http://www.mplayerhq.hu From ywang ee.columbia.edu Fri Jun 17 12:49:00 2005 From: ywang ee.columbia.edu (Yong Wang@EE) Date: Fri Jun 17 13:00:36 2005 Subject: [Mp4-tech] Media File --> Raw File Convertion ( AVI or other mediafiles --> YUV420/422 or M4V) References: Message-ID: <01cb01c57354$14d5a500$64423b80@ee.columbia.edu> Hi Vinay, Based on your requirement, I suggest you have a look at VisGenie, one software package I wrote: http://www.ee.columbia.edu/~ywang/Research/VisGenie Using VisGenie, you can: 1. Convert media file into YUV420/YUV422 format (Specifically, IYUV and YUY2) using the built-in tool Frame Ripper. Support any video types as long as your Windows Media Player recognizes them. 2. Render YUV files in Windows Media Player. 3. Building demos and video processing applications. You can check the URL yourself. BTW, you might want to check the open source package "ffdshow" to get rich media support. Hope these help. Regards, Yong ----- Original Message ----- From: "Vinay Chiniwar" To: Sent: Monday, June 13, 2023 11:23 PM Subject: [Mp4-tech] Media File --> Raw File Convertion ( AVI or other mediafiles --> YUV420/422 or M4V) > Hi, > > I have player running on my system supports only YUV420 and YUV422 format > and both of them are raw video formats. > My player is able to play attached m4v files which is also a raw format. > > * How m4v format is different from YUV420/YUV422 format > * How can I create YUV420/YUV422 format files from other media formats like > wma,avi,vob,mov,mp4 ... > > Now I want add more clips to my system either m4v or YUV420/422 .. > > I did convert the mp4 format files to m4v format using mp4UI tool > (http://mp4ui.sourceforge.net/) but then my player is not able to play > converted m4v format files returns with invalid file format. > > * How can I check the header information of the converted m4v file ( > fps,bit rate, video size..etc) > * Is there any other tool which converts into m4v / yuv format > * How Can I test play m4v or YUV format files under Windows System > > Please guide me to right direction... > > Regards, > Vinay > > > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [vide o], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From nicko via-telecom.com Fri Jun 17 11:01:19 2005 From: nicko via-telecom.com (Nobu Okuyama) Date: Fri Jun 17 13:17:43 2005 Subject: [Mp4-tech][Audio][AAC] AAC vs. WMA Message-ID: <22DB07950E378646922C61A74FCEB6DF0112BA93@exchsd1> Dear all, I would like to have expert comment on comparison between aac and wma in terms of summary of encoding distinctives and performance. Would there be any site that discusses these aspects somewhere? I also would like to know where one may obtain some technical presentation on wma..... Thank you, -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050617/dc0f7888/attachment.html From cmjoshi dlink.co.in Sat Jun 18 16:33:48 2005 From: cmjoshi dlink.co.in (Joshi) Date: Sat Jun 18 08:59:14 2005 Subject: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream Message-ID: <000101c573ed$057eb300$1a0aa8c0@CMJoshi> Hi, I do have Mpeg 4 Transport Stream data reading from an Encoder (Video ONLY - VES). I am reading (188 * 10) = 1880 bytes in a buffer. I need to parse this Data, in order to know the end of Frame(s) in it. In other words i need to know End of one frame and beginning of next frame. This is very much important for me because, I need to Fragment this data (corresponding to single frame) into Smaller chunks of length equal to multiple of 188 bytes. For the Last chunk I need to assign the Marker Bit as '1' because I use RTP as my Transport protocol. Note: Out of 188 bytes, 4 bytes will be header. Can i get information about end Start of New Frame in it? Does this 4 Byte Header Pattern vary across the other formats like MPEG-2 TS, H.263 TS, H.264 TS. Pls Comment. Regards, Joshi. Dlink India Limited, Bangalore. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050618/5d7b1e97/attachment.html From cmjoshi dlink.co.in Sun Jun 19 13:00:01 2005 From: cmjoshi dlink.co.in (Joshi) Date: Sun Jun 19 05:50:05 2005 Subject: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream In-Reply-To: Message-ID: <000001c57498$53448e60$1a0aa8c0@CMJoshi> Dear Mr. Peter, Thanks for your reply. I have few more doubts. 1. Whether MPEG-4 use the same header format as that of MPEG-2 Transport Stream. 2. According to my understanding, every frame of data starts with a SYNC Byte, which is indicated by 0x47. The following data structure kind of thing I am using typedef struct { unsigned int sync_byte :8; unsigned int transport_err_inc :1; unsigned int payload_start :1; unsigned int transport_prio :1; unsigned int pid :13; unsigned int transport_scrambling_control :2; unsigned int adap_field_control :2; unsigned int continuity_counter :4; }mpeg_ts_header; The payload_start, indicates whether the Payload is starting (Start of Frame) is present. Further subsequent headers need not have the payload_start bit as '1', till the completion of the current frame, and start of the next frame. I am parsing the buffer, byte by byte in order to know the occurrence of sync_byte (0x47) and corresponding payload_start bit as '1'. I am getting the pattern which is shown below This is for MPEG-2 Transport Stream. Note: Here Starting Index is the point where 0x47 (Sync Byte) is found and corresponding payload_start is also '1'. Buffer length was 10 KB. Frame starts Starting Index = 623 Frame starts Starting Index = 1033 Difference = 410 Frame starts Starting Index = 1304 Difference = 271 Frame starts Starting Index = 3125 Difference = 1821 Frame starts Starting Index = 4067 Difference = 942 Frame starts Starting Index = 5195 Difference = 1128 Frame starts Starting Index = 5972 Difference = 777 Frame starts Starting Index = 6032 Difference = 60 Frame starts Starting Index = 6103 Difference = 71 Frame starts Starting Index = 6125 Difference = 22 Frame starts Starting Index = 6407 Difference = 282 Frame starts Starting Index = 6545 Difference = 138 Frame starts Starting Index = 6566 Difference = 21 Frame starts Starting Index = 7121 Difference = 555 Frame starts Starting Index = 8108 Difference = 987 Frame starts Starting Index = 8834 Difference = 726 Frame starts Starting Index = 8970 Difference = 136 Frame starts Starting Index = 9078 Difference = 108 Frame starts Starting Index = 9174 Difference = 96 Frame starts Starting Index = 9193 Difference = 19 Frame starts Starting Index = 9208 Difference = 15 Frame starts Starting Index = 9221 Difference = 13 Frame starts Starting Index = 9311 Difference = 90 Frame starts Starting Index = 9393 Difference = 82 Frame starts Starting Index = 9501 Difference = 108 Frame starts Starting Index = 9902 Difference = 401 My Doubt is why the Difference is not multiple of 188 bytes?. On what logic I need to fragment this? Can any of you pls share me the Frame formats for MPEG-4, H.263 Transport Stream Header if it is different from the MPEG-2 Transport Stream System. Pls let me know. Pls comment. Regards, Joshi -----Original Message----- From: Peter Schirling [mailto:schirlin@us.ibm.com] Sent: Saturday, June 18, 2023 6:30 PM To: Joshi Subject: Re: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream Joshi, What you have is MPEG- 4 video over MPEG-2 Transport Stream. In order to demux you need to obtain the MPEG-2 Systems Standards ISO/IEC 13818-1 from: India (BIS) Address Bureau of Indian Standards Manak Bhavan 9 Bahadur Shah Zafar Marg IN-New Delhi 110002 Tel +91 11 23 23 79 91 Fax +91 11 23 23 93 99 E-mail ird@bis.org.in Web http://www.bis.org.in Pete Schirling Digital Media Standards IBM Research Office: +1 802 769 6123/Mobile: +1 802 238 2036/E-Fax: +1 802 769 7362 Internet e-mail: schirlin@us.ibm.com "Joshi" To Sent by: "MPEG4" mp4-tech-bounces@ cc lists.mpegif.org Subject [Mp4-tech] Need Information about 06/18/2005 07:03 Parsing of MPEG-4 Transport Stream PM Hi, I do have Mpeg 4 Transport Stream data reading from an Encoder (Video ONLY - VES). I am reading (188 * 10) = 1880 bytes in a buffer. I need to parse this Data, in order to know the end of Frame(s) in it. In other words i need to know End of one frame and beginning of next frame. This is very much important for me because, I need to Fragment this data (corresponding to single frame) into Smaller chunks of length equal to multiple of 188 bytes. For the Last chunk I need to assign the Marker Bit as '1' because I use RTP as my Transport protocol. Note: Out of 188 bytes, 4 bytes will be header. Can i get information about end Start of New Frame in it? Does this 4 Byte Header Pattern vary across the other formats like MPEG-2 TS, H.263 TS, H.264 TS. Pls Comment. Regards, Joshi. Dlink India Limited, Bangalore. _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From ysramkumar yahoo.com Sun Jun 19 12:20:29 2005 From: ysramkumar yahoo.com (Srinivas Ramkumar Yadavalli) Date: Mon Jun 20 03:30:17 2005 Subject: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream In-Reply-To: <000001c57498$53448e60$1a0aa8c0@CMJoshi> Message-ID: <20050619182029.73986.qmail@web54501.mail.yahoo.com> Hello, Even for h.264/mpeg4 use the same transport packet stream as was used for mpeg2. The maximum size of TS packet is 188. Regards, Ramkumar --- Joshi wrote: > Dear Mr. Peter, > > Thanks for your reply. > I have few more doubts. > > 1. Whether MPEG-4 use the same header format as > that of MPEG-2 > Transport Stream. > 2. According to my understanding, every frame of > data starts with a > SYNC Byte, which is indicated by 0x47. The > following data structure > kind of thing I am using > typedef struct { > unsigned int sync_byte :8; > unsigned int transport_err_inc :1; > unsigned int payload_start :1; > unsigned int transport_prio :1; > unsigned int pid :13; > unsigned int > transport_scrambling_control :2; > unsigned int adap_field_control :2; > unsigned int continuity_counter :4; > }mpeg_ts_header; > > The payload_start, indicates whether the Payload is > starting (Start of > Frame) is present. Further subsequent headers need > not have the > payload_start bit as '1', till the completion of the > current frame, and > start of the next frame. > > I am parsing the buffer, byte by byte in order to > know the occurrence of > sync_byte (0x47) and corresponding payload_start bit > as '1'. I am > getting the pattern which is shown below > This is for MPEG-2 Transport Stream. > > Note: Here Starting Index is the point where 0x47 > (Sync Byte) is found > and corresponding payload_start is also '1'. Buffer > length was 10 KB. > > Frame starts Starting Index = 623 > Frame starts Starting Index = 1033 Difference = 410 > Frame starts Starting Index = 1304 Difference = 271 > Frame starts Starting Index = 3125 Difference = > 1821 > Frame starts Starting Index = 4067 Difference = 942 > Frame starts Starting Index = 5195 Difference = > 1128 > Frame starts Starting Index = 5972 Difference = 777 > Frame starts Starting Index = 6032 Difference = 60 > Frame starts Starting Index = 6103 Difference = 71 > Frame starts Starting Index = 6125 Difference = 22 > Frame starts Starting Index = 6407 Difference = 282 > Frame starts Starting Index = 6545 Difference = 138 > Frame starts Starting Index = 6566 Difference = 21 > Frame starts Starting Index = 7121 Difference = 555 > Frame starts Starting Index = 8108 Difference = 987 > Frame starts Starting Index = 8834 Difference = 726 > Frame starts Starting Index = 8970 Difference = 136 > Frame starts Starting Index = 9078 Difference = 108 > Frame starts Starting Index = 9174 Difference = 96 > Frame starts Starting Index = 9193 Difference = 19 > Frame starts Starting Index = 9208 Difference = 15 > Frame starts Starting Index = 9221 Difference = 13 > Frame starts Starting Index = 9311 Difference = 90 > Frame starts Starting Index = 9393 Difference = 82 > Frame starts Starting Index = 9501 Difference = 108 > Frame starts Starting Index = 9902 Difference = 401 > > My Doubt is why the Difference is not multiple of > 188 bytes?. On what > logic I need to fragment this? > > Can any of you pls share me the Frame formats for > MPEG-4, H.263 > Transport Stream Header if it is different from the > MPEG-2 Transport > Stream System. Pls let me know. > > Pls comment. > > Regards, > Joshi > > -----Original Message----- > From: Peter Schirling [mailto:schirlin@us.ibm.com] > Sent: Saturday, June 18, 2023 6:30 PM > To: Joshi > Subject: Re: [Mp4-tech] Need Information about > Parsing of MPEG-4 > Transport Stream > > Joshi, > > What you have is MPEG- 4 video over MPEG-2 Transport > Stream. In order to > demux you need to obtain the MPEG-2 Systems > Standards > > ISO/IEC 13818-1 from: > > > > > > > > > India (BIS) > > > > > > > > > > > > > > Address > > Bureau of Indian Standards > > Manak Bhavan > > 9 Bahadur Shah Zafar Marg > > > > IN-New Delhi 110002 > > > > > > Tel > > +91 11 23 23 79 91 > > Fax > > +91 11 23 23 93 99 > > E-mail > > ird@bis.org.in > > Web > > http://www.bis.org.in > > > > > > > > > > Pete Schirling > > Digital Media Standards > IBM Research > Office: +1 802 769 6123/Mobile: +1 802 238 > 2036/E-Fax: +1 802 769 7362 > Internet e-mail: schirlin@us.ibm.com > > > > > > "Joshi" > > > .in> > To > Sent by: "MPEG4" > > mp4-tech-bounces@ > cc > lists.mpegif.org > > > Subject > [Mp4-tech] > Need Information about > > 06/18/2005 07:03 Parsing of > MPEG-4 Transport > Stream > PM > > > > > > > > > > > > > > > > Hi, > > > > > > I do have Mpeg 4 Transport Stream data reading from > an Encoder (Video > ONLY > - VES). I am reading (188 * 10) = 1880 bytes in a > buffer. I need to > parse > this Data, in order to know the end of Frame(s) in > it. In other words i > need to know End of one frame and beginning of next > frame. > > > > > > This is very much important for me because, I need > to Fragment this data > (corresponding to single frame) into Smaller chunks > of length equal to > multiple of 188 bytes. For the Last chunk I need > to assign the Marker > Bit > as '1' because I use RTP as my Transport protocol. > > > > > > Note: > > > > > > Out of 188 bytes, 4 bytes will be header. Can i get > information about > end > Start of New Frame in it? > > > > > > Does this 4 Byte Header Pattern vary across the > other formats like > MPEG-2 > TS, H.263 TS, H.264 TS. > > > > > > Pls Comment. > > > > > > Regards, > > > Joshi. > > > Dlink India Limited, Bangalore. > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. > Include [audio, > [video], [systems], [general] or another > apppropriate identifier to > indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the > Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. > Include [audio, [video], [systems], [general] or > another apppropriate identifier to indicate the type > of question you have. > > Note: Conduct on the mailing list is subject to the > Antitrust guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > ____________________________________________________ Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football http://football.fantasysports.yahoo.com From bharatsoni gmail.com Mon Jun 20 10:22:19 2005 From: bharatsoni gmail.com (Bharat Soni) Date: Mon Jun 20 03:30:23 2005 Subject: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream In-Reply-To: <000001c57498$53448e60$1a0aa8c0@CMJoshi> References: <000001c57498$53448e60$1a0aa8c0@CMJoshi> Message-ID: Dear Mr. Joshi. > I am parsing the buffer, byte by byte in order to know the occurrence of > sync_byte (0x47) and corresponding payload_start bit as '1'. I am > getting the pattern which is shown below > This is for MPEG-2 Transport Stream. > > Note: Here Starting Index is the point where 0x47 (Sync Byte) is found > and corresponding payload_start is also '1'. Buffer length was 10 KB. 0x47 is a sync word representing the start of a new transport packet. This is repeated every 188 bytes. For your need, you need to first search for this sync word. Once you have detected this, then serch for the required pattern in those 188 bytes. For example if you are looking for the start of MPEG2 video frame, then you have to search for a start code 0x00000100 (4 bytes). You have to take care of the situation when this startcode is splitted between transport packet. Now in the log that you have generated, it seems that it is reflecting the difference between two consecutive frame start since you have mentioned that the offset values are taken when payload start indicator == 1. If this is true, then this difference need not be multiple of 188 bytes. Regards Bharat From schirlin us.ibm.com Mon Jun 20 14:05:37 2005 From: schirlin us.ibm.com (Peter Schirling) Date: Mon Jun 20 03:30:29 2005 Subject: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream In-Reply-To: <000001c57498$53448e60$1a0aa8c0@CMJoshi> Message-ID: Joshi, Transport Stream syntax is the same whether MPEG-2 or AVC is carried in the multiplex. Likewise, PES packets headers the same. What changes are codes that identify the various MPEG video and audio standards. As I indicated, in order to handle all the information correctly you will need to obtain a copy of ISO/IEC 13818-1. An email reflector is not the place for this discussion. Pete Schirling Digital Media Standards IBM Research Office: +1 802 769 6123/Mobile: +1 802 238 2036/E-Fax: +1 802 769 7362 Internet e-mail: schirlin@us.ibm.com "Joshi" To Peter 06/19/2005 03:30 Schirling/Burlington/IBM@IBMUS, PM "MPEG4" cc Subject RE: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream Dear Mr. Peter, Thanks for your reply. I have few more doubts. 1. Whether MPEG-4 use the same header format as that of MPEG-2 Transport Stream. 2. According to my understanding, every frame of data starts with a SYNC Byte, which is indicated by 0x47. The following data structure kind of thing I am using typedef struct { unsigned int sync_byte :8; unsigned int transport_err_inc :1; unsigned int payload_start :1; unsigned int transport_prio :1; unsigned int pid :13; unsigned int transport_scrambling_control :2; unsigned int adap_field_control :2; unsigned int continuity_counter :4; }mpeg_ts_header; The payload_start, indicates whether the Payload is starting (Start of Frame) is present. Further subsequent headers need not have the payload_start bit as '1', till the completion of the current frame, and start of the next frame. I am parsing the buffer, byte by byte in order to know the occurrence of sync_byte (0x47) and corresponding payload_start bit as '1'. I am getting the pattern which is shown below This is for MPEG-2 Transport Stream. Note: Here Starting Index is the point where 0x47 (Sync Byte) is found and corresponding payload_start is also '1'. Buffer length was 10 KB. Frame starts Starting Index = 623 Frame starts Starting Index = 1033 Difference = 410 Frame starts Starting Index = 1304 Difference = 271 Frame starts Starting Index = 3125 Difference = 1821 Frame starts Starting Index = 4067 Difference = 942 Frame starts Starting Index = 5195 Difference = 1128 Frame starts Starting Index = 5972 Difference = 777 Frame starts Starting Index = 6032 Difference = 60 Frame starts Starting Index = 6103 Difference = 71 Frame starts Starting Index = 6125 Difference = 22 Frame starts Starting Index = 6407 Difference = 282 Frame starts Starting Index = 6545 Difference = 138 Frame starts Starting Index = 6566 Difference = 21 Frame starts Starting Index = 7121 Difference = 555 Frame starts Starting Index = 8108 Difference = 987 Frame starts Starting Index = 8834 Difference = 726 Frame starts Starting Index = 8970 Difference = 136 Frame starts Starting Index = 9078 Difference = 108 Frame starts Starting Index = 9174 Difference = 96 Frame starts Starting Index = 9193 Difference = 19 Frame starts Starting Index = 9208 Difference = 15 Frame starts Starting Index = 9221 Difference = 13 Frame starts Starting Index = 9311 Difference = 90 Frame starts Starting Index = 9393 Difference = 82 Frame starts Starting Index = 9501 Difference = 108 Frame starts Starting Index = 9902 Difference = 401 My Doubt is why the Difference is not multiple of 188 bytes?. On what logic I need to fragment this? Can any of you pls share me the Frame formats for MPEG-4, H.263 Transport Stream Header if it is different from the MPEG-2 Transport Stream System. Pls let me know. Pls comment. Regards, Joshi -----Original Message----- From: Peter Schirling [mailto:schirlin@us.ibm.com] Sent: Saturday, June 18, 2023 6:30 PM To: Joshi Subject: Re: [Mp4-tech] Need Information about Parsing of MPEG-4 Transport Stream Joshi, What you have is MPEG- 4 video over MPEG-2 Transport Stream. In order to demux you need to obtain the MPEG-2 Systems Standards ISO/IEC 13818-1 from: India (BIS) Address Bureau of Indian Standards Manak Bhavan 9 Bahadur Shah Zafar Marg IN-New Delhi 110002 Tel +91 11 23 23 79 91 Fax +91 11 23 23 93 99 E-mail ird@bis.org.in Web http://www.bis.org.in Pete Schirling Digital Media Standards IBM Research Office: +1 802 769 6123/Mobile: +1 802 238 2036/E-Fax: +1 802 769 7362 Internet e-mail: schirlin@us.ibm.com "Joshi" To Sent by: "MPEG4" mp4-tech-bounces@ cc lists.mpegif.org Subject [Mp4-tech] Need Information about 06/18/2005 07:03 Parsing of MPEG-4 Transport Stream PM Hi, I do have Mpeg 4 Transport Stream data reading from an Encoder (Video ONLY - VES). I am reading (188 * 10) = 1880 bytes in a buffer. I need to parse this Data, in order to know the end of Frame(s) in it. In other words i need to know End of one frame and beginning of next frame. This is very much important for me because, I need to Fragment this data (corresponding to single frame) into Smaller chunks of length equal to multiple of 188 bytes. For the Last chunk I need to assign the Marker Bit as '1' because I use RTP as my Transport protocol. Note: Out of 188 bytes, 4 bytes will be header. Can i get information about end Start of New Frame in it? Does this 4 Byte Header Pattern vary across the other formats like MPEG-2 TS, H.263 TS, H.264 TS. Pls Comment. Regards, Joshi. Dlink India Limited, Bangalore. _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From ssingh cedt.iisc.ernet.in Mon Jun 20 22:50:33 2005 From: ssingh cedt.iisc.ernet.in (Svatantra Singh) Date: Mon Jun 20 16:37:23 2005 Subject: [Mp4-tech] Doubt about u(v) descriptor Message-ID: <9343BE359FA4224CA72A94D8CC362C9C49CCAB@edpc500.cedt.iisc.ernet.in> Hello all. In the syntax table I found a descriptor represented as u(v); it is written that parsing process for this descriptor is depends on the "other" syntax elements. So number of bits returned by u(v) may vary. Can any one please tell what is implied in word "other syntax elements." In particular in syntax of slice header frame_num parameter is described by this descriptor. How do I find that parsing the bit stream using this descriptor. Thanks and regards, svatantra Svatantra Singh M.E.(Microelectronics Systems) Centre for Electronics Design & Technmology (CEDT) Indian Institute of Science(IISc), Bangalore-560012 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050620/a8942b13/attachment.html From garysull windows.microsoft.com Mon Jun 20 14:54:23 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Tue Jun 21 17:30:14 2005 Subject: [Mp4-tech] Doubt about u(v) descriptor Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F1987CB@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Svatantra Singh et al, Each such syntax element has a corresponding semantics section that tells you how many bits is used to represent it in the bitstream. For example, the semantics section for the slice header (subclause 7.4.3) says that frame_num "shall be represented by log2_max_frame_num_minus4 + 4 bits in the bitstream". Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Svatantra Singh Sent: Monday, June 20, 2023 9:21 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Doubt about u(v) descriptor Hello all. In the syntax table I found a descriptor represented as u(v); it is written that parsing process for this descriptor is depends on the "other" syntax elements. So number of bits returned by u(v) may vary. Can any one please tell what is implied in word "other syntax elements." In particular in syntax of slice header frame_num parameter is described by this descriptor. How do I find that parsing the bit stream using this descriptor. Thanks and regards, svatantra Svatantra Singh M.E.(Microelectronics Systems) Centre for Electronics Design & Technmology (CEDT) Indian Institute of Science(IISc), Bangalore-560012 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050620/e7c6b0ee/attachment.html From mail.tobi gmx.net Mon Jun 20 21:12:15 2005 From: mail.tobi gmx.net (Tobias Wolff) Date: Tue Jun 21 17:30:20 2005 Subject: [Mp4-tech] H.264: Quantizationparameter, Intraperiod, artifacts Message-ID: <6.2.1.2.0.20050620201017.0275cae8@pop.gmx.net> Dear Experts, I am new to videocoding and currently doing some experiments with H.264 in the baseline Profile. However, there are some things which I don?t quite understand... Is the Quantizationparameter assigned to the whole frame or to each Macroblock (which I thought is the case)? In case it is assigned to each MB, then why does the Encoder give me one number (for the Quantizationparameter QP) per Frame in its comandline output? How is the Intraperiod handled in a reasonable way, depending on the bitrate? Is there a reason for chosing "only the first frame as an Intraframe" as the default setting? What would be a reason for chosing a small Intraperiod at low bitrates? And what is concidered to be a very small or very large Intraperiod? Since P-Frames allow for intracoded blocks and the Encoder decides on its own how to code each block (within a P-Frame), I don?t really see a reason to force the encoder to use expensive I-Frames in a fixed period, exept for Random Access. Can anybody help me with that? Last but not least, I?d really appreciate any hint to a survey about the artifacts exhibited by H.264. Does anybody know to what extend this has been investigated? Thanks a lot in advance, Best regards, Tobias From Ye-Kui.Wang nokia.com Tue Jun 21 10:01:17 2005 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Tue Jun 21 17:30:25 2005 Subject: [Mp4-tech] About map unit to slice group map. Message-ID: <1C1F3D15859526459B4DD0A7A9B2268BA38B60@trebe101.NOE.Nokia.com> Hi Atul, "map unit" may either be a macroblock or a macroblock pair if MBAFF (macroblock adaptive frame field coding) is in use. "map unit to slice group map" tells the informaiton to which slice group each macroblock or macroblock pair belongs. BR, YK -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext atul tomar Sent: 16 June, 2005 17:15 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] About map unit to slice group map. Hi All, Can anyone please explain me what standard mean by " map unit to slice group map". I read ITU standard many times to understand this but could not get it clear. Please expalin what does "map unit" represent. Thanks in advance. Reg, Atul _____ Discover Yahoo! Find restaurants, movies, travel & more fun for the weekend. Check it out! From Ye-Kui.Wang nokia.com Tue Jun 21 10:05:43 2005 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Tue Jun 21 17:30:31 2005 Subject: [Mp4-tech] Doubt about u(v) descriptor Message-ID: <1C1F3D15859526459B4DD0A7A9B2268BA38B61@trebe101.NOE.Nokia.com> Hi Svatantra, If u(v) is used, the parsing dependency is given in the semantics of the syntax element. For example, for frame_num, it is said that frame_num "shall be represented by log2_max_frame_num_minus4 + 4 bits in the bitstream. ". BR, YK -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext Svatantra Singh Sent: 20 June, 2005 19:21 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Doubt about u(v) descriptor Hello all. In the syntax table I found a descriptor represented as u(v); it is written that parsing process for this descriptor is depends on the "other" syntax elements. So number of bits returned by u(v) may vary. Can any one please tell what is implied in word "other syntax elements." In particular in syntax of slice header frame_num parameter is described by this descriptor. How do I find that parsing the bit stream using this descriptor. Thanks and regards, svatantra Svatantra Singh M.E.(Microelectronics Systems) Centre for Electronics Design & Technmology (CEDT) Indian Institute of Science(IISc), Bangalore-560012 From amni_gupta yahoo.co.in Tue Jun 21 09:10:36 2005 From: amni_gupta yahoo.co.in (Amni Gupta) Date: Tue Jun 21 17:30:36 2005 Subject: [Mp4-tech] Help-Urgent Message-ID: <20050621071036.80316.qmail@web8610.mail.in.yahoo.com> Hi; Please any one can clarify that is there support of Field prediction and Frame prediction both for MPEG-4 Natural Video Simple Profile. Simple profile doesn't support Interlace, but supports P pictures. Can I get the free C code for only MPEG-4 natural video simple profile. Regards Amni. --------------------------------- Too much spam in your inbox? Yahoo! Mail gives you the best spam protection for FREE! http://in.mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050621/f5b5809a/attachment.html From A.Thomson indigovision.com Tue Jun 21 11:20:43 2005 From: A.Thomson indigovision.com (Alan Thomson) Date: Tue Jun 21 17:30:42 2005 Subject: [Mp4-tech] Doubt about u(v) descriptor Message-ID: <67336BF4D5D23C4AA30EC72DFCFCA6DF012BF8A1@peebles.indigovision.com> Hi - The "other" syntax element depends on which particular syntax element you are considering. In the case of frame_num, the length is governed by the value of log2_max_frame_num_minus4, from the sequence parameter set. BR, Alan -----Original Message----- From: Svatantra Singh [mailto:ssingh@cedt.iisc.ernet.in] Sent: 20 June 2023 17:21 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Doubt about u(v) descriptor Hello all. In the syntax table I found a descriptor represented as u(v); it is written that parsing process for this descriptor is depends on the "other" syntax elements. So number of bits returned by u(v) may vary. Can any one please tell what is implied in word "other syntax elements." In particular in syntax of slice header frame_num parameter is described by this descriptor. How do I find that parsing the bit stream using this descriptor. Thanks and regards, svatantra Svatantra Singh M.E.(Microelectronics Systems) Centre for Electronics Design & Technmology (CEDT) Indian Institute of Science(IISc), Bangalore-560012 _____ scanned for viruses and spam by indigovision If you consider this email spam, please forward to spam@emailfiltering.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050621/73545527/attachment.html From soria lsi.usp.br Tue Jun 21 16:15:07 2005 From: soria lsi.usp.br (Sergio Rodríguez Soria) Date: Tue Jun 21 17:30:47 2005 Subject: [Mp4-tech] MP4FF Library Message-ID: Dear all, This was a request posted in 2004: Here is a quote from MPEG-4 audio decoder reference software readme file: .\audio\natural\mp4AudVm_Rewrite\win32\readme.txt: "...the MPEG-4 Version 2 Decoder supports the MPEG-4 File format. For this, the MPEG-4 File Format reference software library provided by ISO is required. These two libraries are not part of the MPEG-4 Audio Reference Software and have to be obtained externally..." .. Please, does anyone know how can I obtain the MPEG-4 File Format (MP4FF) reference software library? And this was the answer provided by Mr. Schuyler Quackenbush: The ISOFF library is available from Apple for MPEG members: ftp: index.apple.com user: sc29wg11 pswd: <66th MPEG meeting password> Now, I have the same problem, I need those libraries but I can't access to the Apple ftp server with these user & password. Could anyone help me please? Thanks in advance Sergio -- Sergio Rodríguez Soria LSI-EPUSP AAC Consortium - SBTVD From lobue medialab.pa.icar.cnr.it Tue Jun 21 21:06:16 2005 From: lobue medialab.pa.icar.cnr.it (Lo BUE Antonino) Date: Tue Jun 21 17:30:53 2005 Subject: [Mp4-tech] ServerCommand usage Message-ID: <1809.194.119.214.112.1119377176.squirrel@medialab.pa.icar.cnr.it> Hi, is possible to use ServerCommand to deal with server sending event and receive data (strings) like coordinates to update the scene? How work MPEG-4 for update from server? there is a polling mode? Thanks From mail.tobi gmx.net Tue Jun 21 12:05:36 2005 From: mail.tobi gmx.net (Tobias Wolff) Date: Tue Jun 21 17:30:59 2005 Subject: [Mp4-tech] help Message-ID: <6.2.1.2.0.20050621103641.02757a38@pop.gmx.net> hi, yesterday I posted my first message to the MP4-Tech list, after subscribing to it. Eventhough I have checked the option to receive a copy of my postings, I did not receive any copy... Now I?m unsure wether my message has ben posted at all, or if I should post it again? this was my subject : H.264: Quantizationparameter, Intraperiod, artifacts Thanks for your help, Best Regards, Tobias From garysull windows.microsoft.com Tue Jun 21 17:14:43 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Wed Jun 22 09:53:30 2005 Subject: [Mp4-tech] About map unit to slice group map. Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F20DE2C@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> To be more precise, see the following excerpts from the specification itself. Note that there is another case besides the MBAFF case in which the map units are not individual macroblocks (the last case listed below). 3.82 map unit to slice group map: A means of mapping slice group map units of a picture into slice groups. The map unit to slice group map consists of a list of numbers, one for each slice group map unit, specifying the slice group to which each coded slice group map unit belongs. 3.139 slice group map units: The units of the map unit to slice group map. 7.4.2.2 Slice group map units are specified as follows. - If frame_mbs_only_flag is equal to 0 and mb_adaptive_frame_field_flag is equal to 1 and the coded picture is a frame, the slice group map units are macroblock pair units. - Otherwise, if frame_mbs_only_flag is equal to 1 or a coded picture is a field, the slice group map units are units of macroblocks. - Otherwise (frame_mbs_only_flag is equal to 0 and mb_adaptive_frame_field_flag is equal to 0 and the coded picture is a frame), the slice group map units are units of two macroblocks that are vertically contiguous as in a frame macroblock pair of an MBAFF frame. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of +> Ye-Kui.Wang@nokia.com +> Sent: Monday, June 20, 2023 11:01 PM +> To: tomar_atul2000@yahoo.com; mp4-tech@lists.mpegif.org +> Subject: RE: [Mp4-tech] About map unit to slice group map. +> +> Hi Atul, +> +> "map unit" may either be a macroblock or a macroblock pair +> if MBAFF (macroblock adaptive frame field coding) is in use. +> "map unit to slice group map" tells the informaiton to which +> slice group each macroblock or macroblock pair belongs. +> +> BR, YK +> +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext atul tomar +> Sent: 16 June, 2005 17:15 +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] About map unit to slice group map. +> +> +> Hi All, +> +> Can anyone please explain me what standard mean by " map +> unit to slice group map". I read ITU standard many times to +> understand this but could not get it clear. +> Please expalin what does "map unit" represent. +> +> Thanks in advance. +> +> Reg, +> Atul +> +> +> +> _____ +> +> Discover Yahoo! +> Find restaurants, movies, travel & more fun for the weekend. +> Check +> weekend.html> it out! +> +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From singer apple.com Tue Jun 21 19:03:22 2005 From: singer apple.com (Dave Singer) Date: Wed Jun 22 10:01:41 2005 Subject: [Mp4-tech] MP4FF Library In-Reply-To: References: Message-ID: At 15:15 +0000 21/06/05, Sergio =?UNKNOWN?Q?Rodr=EDguez?= Soria wrote: >Dear all, > >This was a request posted in 2004: > >Here is a quote from MPEG-4 audio decoder reference software readme file: >.\audio\natural\mp4AudVm_Rewrite\win32\readme.txt: > >"...the MPEG-4 Version 2 Decoder supports the MPEG-4 File format. For >this, the MPEG-4 File Format reference software library provided by ISO is >required. These two libraries are not part of the MPEG-4 Audio Reference >Software and have to be obtained externally..." > >.. > >Please, does anyone know how can I obtain the MPEG-4 File Format (MP4FF) >reference software library? > >And this was the answer provided by Mr. Schuyler Quackenbush: > >The ISOFF library is available from Apple for MPEG members: > >ftp: index.apple.com >user: sc29wg11 >pswd: <66th MPEG meeting password> > >Now, I have the same problem, I need those libraries but I can't access to >the Apple ftp server with these user & password. > >Could anyone help me please? > Sure. one only needs ask the mp4ra (www.mp4ra.org) which is me, and I'll be happy to mail it. -- David Singer Apple Computer/QuickTime From everain castis.com Wed Jun 22 12:31:25 2005 From: everain castis.com (???) Date: Wed Jun 22 10:08:31 2005 Subject: [Mp4-tech] H.264 : Extract only I frames - Not Playable In-Reply-To: <200506220115.j5M1DMf1029826@lists1.magma.ca> Message-ID: Hi, I'm trying to implement a trick-like play by extracting all I frames from a H.264 byte stream and playing them. But I've found that the extracted stream is NOT played smoothly. 1) first, during the playback, there comes pause(of 1~2 sec) time to time. 2) Especially during the period when IDR frames are in serial sequence ( ... - IDR- IDR - ... - IDR - ... ), only the last IDR frame of that sequence is played during that period. It seems I've got somthing to modify syntax elements when doing the job, but I can't figure it out. Would you please let me know what I did wrong ? Thanks in advance, Young-Woo. ____________ Unicure?! 1UAI.??o ?E>c 1oA?: AVK 15.0.5568, 2023.06.19 1UAI.??o ?o?o: www.geot.com From ugarg neomagic.com Wed Jun 22 13:16:49 2005 From: ugarg neomagic.com (Umang Garg) Date: Wed Jun 22 10:16:38 2005 Subject: [Mp4-tech] MP4FF Library In-Reply-To: References: Message-ID: <42B90959.4040907@neomagic.com> Hello Sergio, Yes, there seems to be problem. The site no longer accepts the password for 66th MPEG meeting. I think that a new meting password has been given to the site. Could someone please identify which meeting's password is currently active on index.apple.com ? Best Regards, Umang Garg NeoMagic Corporation Sergio Rodr?guez Soria wrote: >Dear all, > >This was a request posted in 2004: > >Here is a quote from MPEG-4 audio decoder reference software readme file: >.\audio\natural\mp4AudVm_Rewrite\win32\readme.txt: > >"...the MPEG-4 Version 2 Decoder supports the MPEG-4 File format. For >this, the MPEG-4 File Format reference software library provided by ISO is >required. These two libraries are not part of the MPEG-4 Audio Reference >Software and have to be obtained externally..." > >.. > >Please, does anyone know how can I obtain the MPEG-4 File Format (MP4FF) >reference software library? > >And this was the answer provided by Mr. Schuyler Quackenbush: > >The ISOFF library is available from Apple for MPEG members: > >ftp: index.apple.com >user: sc29wg11 >pswd: <66th MPEG meeting password> > >Now, I have the same problem, I need those libraries but I can't access to >the Apple ftp server with these user & password. > >Could anyone help me please? > >Thanks in advance > >Sergio > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/20498958/attachment.html From Ye-Kui.Wang nokia.com Wed Jun 22 11:28:58 2005 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Wed Jun 22 10:23:33 2005 Subject: [Mp4-tech] MP4FF Library Message-ID: <1C1F3D15859526459B4DD0A7A9B2268BA38B6B@trebe101.NOE.Nokia.com> Hi Sergio, You may check www.mpeg4ip.net. BR, YK > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext > Sergio Rodr?guez Soria > Sent: 21 June, 2005 18:15 > To: mp4-tech@lists.mpegif.org > Cc: sergio@lps.usp.br > Subject: [Mp4-tech] MP4FF Library > > > > Dear all, > > This was a request posted in 2004: > > Here is a quote from MPEG-4 audio decoder reference software > readme file: > .\audio\natural\mp4AudVm_Rewrite\win32\readme.txt: > > "...the MPEG-4 Version 2 Decoder supports the MPEG-4 File format. For > this, the MPEG-4 File Format reference software library > provided by ISO is > required. These two libraries are not part of the MPEG-4 > Audio Reference > Software and have to be obtained externally..." > > .. > > Please, does anyone know how can I obtain the MPEG-4 File > Format (MP4FF) > reference software library? > > And this was the answer provided by Mr. Schuyler Quackenbush: > > The ISOFF library is available from Apple for MPEG members: > > ftp: index.apple.com > user: sc29wg11 > pswd: <66th MPEG meeting password> > > Now, I have the same problem, I need those libraries but I > can't access to > the Apple ftp server with these user & password. > > Could anyone help me please? > > Thanks in advance > > Sergio > > -- > Sergio Rodr?guez Soria > LSI-EPUSP > AAC Consortium - SBTVD > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php From JelyZhu innosis.com.cn Wed Jun 22 18:32:07 2005 From: JelyZhu innosis.com.cn (JelyZhu@innosis.com.cn) Date: Wed Jun 22 10:30:13 2005 Subject: [Mp4-tech] about horizontal motion vector component range in baseline profile! Message-ID: hi experts: The vertical motion vector component range has a corresponding limit for each level In Annex A.3.1. However, there is no a specific range for horizontal motion vecor but one sentence "Horizontal motion vector range does not exceed the range of -2048 to 2047.75, inclusive, in units of luma samples". I wonder whether this means horizontal motion vector for all level should obey this sentence? If the size of input file is just 176x144, the range is still right? best regards Jelly From tomar_atul2000 yahoo.com Wed Jun 22 08:57:49 2005 From: tomar_atul2000 yahoo.com (atul tomar) Date: Wed Jun 22 17:48:45 2005 Subject: [Mp4-tech] About intra prediction Message-ID: <20050622145749.31392.qmail@web52301.mail.yahoo.com> Dear all, I have a question on intra prediction in h.264 standard. Would you please explain it to me. If the Intra4x4 subMB is the first subMB in the first MB(upper-left of frame) of an IDR-picture, and I want to decide predIntra4x4PredMode for this subMB. How to decide this beacuse it does not have any above and left SubMBs or any samples? If it is set to Mode2(DC prediction mode) even this mode require above and left samples to get Mean of these samples for prediction. Same case is with Intra16x16 prediction for first MB of an IDR or I picture. Thanks in advance. Regards, Atul --------------------------------- Do you Yahoo!? Make Yahoo! your home page -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/c4c2fa04/attachment.html From gops98 yahoo.com Wed Jun 22 09:10:40 2005 From: gops98 yahoo.com (Krithikh Malini) Date: Wed Jun 22 17:48:51 2005 Subject: [Mp4-tech] YUV data Message-ID: <20050622151041.38891.qmail@web54202.mail.yahoo.com> Hi all, I am looking for YUV raw video data of 60frames/sec. Could you pls send me some pointer where can I get that. Thanks & Regards, SGK --------------------------------- Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/d3f80aeb/attachment.html From iamsangu gmail.com Wed Jun 22 23:47:57 2005 From: iamsangu gmail.com (sangeeth) Date: Wed Jun 22 17:48:58 2005 Subject: [Mp4-tech] DRM Message-ID: <42B99D45.7010608@gmail.com> hi I want to know about Digital rights management. (DRM). I am new to this technology. can u please tell me 1. from which site i can get the documents about DRM 2. from where i can download source code for DRM thanks in advance lisa sangs From garysull windows.microsoft.com Wed Jun 22 13:01:30 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Wed Jun 22 17:49:05 2005 Subject: [Mp4-tech] about horizontal motion vector component range inbaseline profile! Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F20E67A@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> I think that is the correct interpretation. -Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of +> JelyZhu@innosis.com.cn +> Sent: Wednesday, June 22, 2023 2:32 AM +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] about horizontal motion vector component +> range inbaseline profile! +> +> hi experts: +> +> The vertical motion vector component range has a +> corresponding limit +> for each level In Annex A.3.1. However, there is no a +> specific range for +> horizontal motion vecor but one sentence "Horizontal motion +> vector range +> does not exceed the range of -2048 to 2047.75, inclusive, in +> units of luma +> samples". I wonder whether this means horizontal motion +> vector for all +> level should obey this sentence? If the size of input file +> is just 176x144, +> the range is still right? +> +> best regards +> Jelly +> +> +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From jingli888ca 163.com Wed Jun 22 11:06:45 2005 From: jingli888ca 163.com (=?gb2312?B?SmluZyBMaQ==?=) Date: Thu Jun 23 04:26:09 2005 Subject: [Mp4-tech] How to make inter prediction and de-blocking work in parallel? Message-ID: <42B8C7B5.000008.29259@m230.163.com> Hello, Experts: When making the baseline profile FMO, did people consider the following scenario: When inter predicting an MB in the current frame, the referred pixels has NOT been de-blocked. Then inter prediction has to wait for de-blocking to catch up. In a worst case, if the referred pixels are 100 MB ahead of the MB being de-blocked, then it will cause the system to wait for a significantly long time. Is there a way that inter prediction and de-blocking work in parallel? Thanks. Jing -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/44e754d1/attachment.html From singer apple.com Wed Jun 22 16:39:20 2005 From: singer apple.com (Dave Singer) Date: Thu Jun 23 04:33:17 2005 Subject: [Mp4-tech] MP4FF Library In-Reply-To: <42B90959.4040907@neomagic.com> References: <42B90959.4040907@neomagic.com> Message-ID: At 12:16 +0530 22/06/05, Umang Garg wrote: >Hello Sergio, > >Yes, there seems to be problem. The site no >longer accepts the password for 66th MPEG >meeting. I think that a new meting password has >been given to the site. Could someone please >identify which meeting's password is currently >active on index.apple.com ? > >Best Regards, >Umang Garg > >NeoMagic Corporation I can't send you mpeg passwords, but contacting the mp4ra.org (me) will get the library to you on request. > >Sergio Rodr?guez Soria wrote: > >>Dear all, >> >>This was a request posted in 2004: >> >>Here is a quote from MPEG-4 audio decoder reference software readme file: >>.\audio\natural\mp4AudVm_Rewrite\win32\readme.txt: >> >>"...the MPEG-4 Version 2 Decoder supports the MPEG-4 File format. For >>this, the MPEG-4 File Format reference software library provided by ISO is >>required. These two libraries are not part of the MPEG-4 Audio Reference >>Software and have to be obtained externally..." >> >>.. >> >>Please, does anyone know how can I obtain the MPEG-4 File Format (MP4FF) >>reference software library? >> >>And this was the answer provided by Mr. Schuyler Quackenbush: >> >>The ISOFF library is available from Apple for MPEG members: >> >>ftp: index.apple.com >>user: sc29wg11 >>pswd: <66th MPEG meeting password> >> >>Now, I have the same problem, I need those libraries but I can't access to >>the Apple ftp server with these user & password. >> >>Could anyone help me please? >> >>Thanks in advance >> >>Sergio >> >> >> > >_______________________________________________ >NOTE: Please use clear subject lines for your >posts. Include [audio, [video], [systems], >[general] or another apppropriate identifier to >indicate the type of question you have. > >Note: Conduct on the mailing list is subject to >the Antitrust guidelines found at >http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -- David Singer Apple Computer/QuickTime -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/9b6a8be5/attachment.html From singer apple.com Wed Jun 22 16:40:05 2005 From: singer apple.com (Dave Singer) Date: Thu Jun 23 04:33:23 2005 Subject: [Mp4-tech] MP4FF Library In-Reply-To: <1C1F3D15859526459B4DD0A7A9B2268BA38B6B@trebe101.NOE.Nokia.com> References: <1C1F3D15859526459B4DD0A7A9B2268BA38B6B@trebe101.NOE.Nokia.com> Message-ID: At 10:28 +0300 22/06/05, Ye-Kui.Wang@nokia.com wrote: >Hi Sergio, > >You may check www.mpeg4ip.net. > >BR, YK Yes, I think that's a different implementation. I believe that there are a number of open-source implementations (the virtues of open standards -- you get to choose!). > >> -----Original Message----- >> From: mp4-tech-bounces@lists.mpegif.org >> [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext >> Sergio Rodr?guez Soria >> Sent: 21 June, 2005 18:15 >> To: mp4-tech@lists.mpegif.org >> Cc: sergio@lps.usp.br >> Subject: [Mp4-tech] MP4FF Library >> >> >> >> Dear all, >> >> This was a request posted in 2004: >> >> Here is a quote from MPEG-4 audio decoder reference software >> readme file: >> .\audio\natural\mp4AudVm_Rewrite\win32\readme.txt: >> >> "...the MPEG-4 Version 2 Decoder supports the MPEG-4 File format. For >> this, the MPEG-4 File Format reference software library >> provided by ISO is >> required. These two libraries are not part of the MPEG-4 >> Audio Reference >> Software and have to be obtained externally..." >> >> .. >> >> Please, does anyone know how can I obtain the MPEG-4 File >> Format (MP4FF) >> reference software library? >> >> And this was the answer provided by Mr. Schuyler Quackenbush: >> >> The ISOFF library is available from Apple for MPEG members: >> >> ftp: index.apple.com >> user: sc29wg11 >> pswd: <66th MPEG meeting password> >> >> Now, I have the same problem, I need those libraries but I >> can't access to >> the Apple ftp server with these user & password. >> >> Could anyone help me please? >> >> Thanks in advance >> >> Sergio >> >> -- >> Sergio Rodr?guez Soria >> LSI-EPUSP >> AAC Consortium - SBTVD >> _______________________________________________ >> NOTE: Please use clear subject lines for your posts. Include >> [audio, [video], [systems], [general] or another apppropriate >> identifier to indicate the type of question you have. >> >> Note: Conduct on the mailing list is subject to the Antitrust >> guidelines found at >> http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti >trust.php > >_______________________________________________ >NOTE: Please use clear subject lines for your >posts. Include [audio, [video], [systems], >[general] or another apppropriate identifier to >indicate the type of question you have. > >Note: Conduct on the mailing list is subject to >the Antitrust guidelines found at >http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -- David Singer Apple Computer/QuickTime From alexismt comcast.net Thu Jun 23 00:05:50 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Thu Jun 23 04:33:30 2005 Subject: [Mp4-tech] About intra prediction Message-ID: <062220052305.22021.42B9EECE0004A2170000560522007621949B039C07980A040E@comcast.net> Dear Atul If you read section 8.3.1.2.3 carefully you will note the following statement : =========================== - Otherwise (some samples p[ x, -1 ], with x = 0..3 and some samples p[ -1, y ], with y = 0..3 are marked as “not available for Intra_4x4 prediction”), the values of the prediction samples pred4x4L[ x, y ], with x, y = 0..3 are derived by pred4x4L[ x, y ] = ( 1 << ( BitDepthY – 1 ) ) (8-50) NOTE – A 4x4 luma block can always be predicted using this mode. =========================== The above suggests that if samples are not available for prediction, then prediction is set to a fixed value according to the BitDepth format of the video sequence. i.e. assuming that you are using 8 bits, then prediction is equal to 128. You will notice a similar consideration for the Intra_8x8, Intra_16x16, and even chroma prediction cases. Best regards, Alexis -------------- Original message -------------- Dear all, I have a question on intra prediction in h.264 standard. Would you please explain it to me. If the Intra4x4 subMB is the first subMB in the first MB(upper-left of frame) of an IDR-picture, and I want to decide predIntra4x4PredMode for this subMB. How to decide this beacuse it does not have any above and left SubMBs or any samples? If it is set to Mode2(DC prediction mode) even this mode require above and left samples to get Mean of these samples for prediction. Same case is with Intra16x16 prediction for first MB of an IDR or I picture. Thanks in advance. Regards, Atul Do you Yahoo!? Make Yahoo! your home page -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/d3d5d18f/attachment-0001.html From kexu ee.cuhk.edu.hk Thu Jun 23 10:31:31 2005 From: kexu ee.cuhk.edu.hk (Xu Ke) Date: Thu Jun 23 04:33:37 2005 Subject: [Mp4-tech] About intra prediction Message-ID: <200506230131.j5N1VItL023114@cuees8.ee.cuhk.edu.hk> Hi Atul, Yes,you are right,it will be predicted by Mode2(DC).Just look careful at the standard: -- Otherwise (some samples p[ x, -1 ], with x = 0..3 and some samples p[ -1, y ], with y = 0..3 are marked as Ą°not available for Intra_4x4 predictionĄą), the values of the prediction samples pred4x4L[ x, y ], with x, y = 0..3 are derived by:pred4x4L[ x, y ] = 128. So all the predicted values of this Intra4x4blk will be set to 128. Best regards, XU Ke ======= 2023-06-22 22:57:49 ================= >Dear all, > >I have a question on intra prediction in h.264 standard. Would you please explain it to me. > >If the Intra4x4 subMB is the first subMB in the first MB(upper-left of frame) of an IDR-picture, and I want to decide predIntra4x4PredMode for this subMB. How to decide this beacuse it does not have any above and left SubMBs or any samples? > >If it is set to Mode2(DC prediction mode) even this mode require above and left samples to get Mean of these samples for prediction. > >Same case is with Intra16x16 prediction for first MB of an IDR or I picture. > >Thanks in advance. >Regards, >Atul > > >--------------------------------- >Do you Yahoo!? > Make Yahoo! your home page >_______________________________________________ >NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > >Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php = = = = = = = = = = = = = = = = = = = = From alexismt comcast.net Wed Jun 22 20:11:42 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Thu Jun 23 04:33:44 2005 Subject: [Mp4-tech] H.264: Quantizationparameter, Intraperiod, artifacts In-Reply-To: <6.2.1.2.0.20050620201017.0275cae8@pop.gmx.net> Message-ID: <200506230211.j5N2Bvge010973@lists1.magma.ca> Dear Tobias, The H.264 standard definitely allows Quantization (QP) variations at the Macroblock level. This can be done basically through the use of the mb_qp_delta element. Macroblock level but even Frame level adaptations of the QP however require certain decision mechanisms, i.e. through the consideration of a rate or quality control mechanism, which are beyond the standard itself (i.e. these are considered as non-normative concepts). To give you an example, one may analyze a scene and assign higher quality (smaller QP values) to regions which are considered more important (i.e. have higher subjective impact). Considering the overhead that could be involved with transmitting this extra information however (especially at very low bitrates), in many applications fixed QP, or QP adaptation at the frame level, can be sufficient. The JM reference software does provide an "example" rate control (please note that the keyword here is example since it is nowhere claimed that this rate control is the best rate control or that one should follow this implementation) which could potentially lead to QP variations at the MB level. In terms of QP statistics however, displaying all QP values for each MB may not be as intuitive (i.e. even for a CIF resolution sequence you will have to display 396 values). Therefore, and for readability, the encoder would only display the last Macroblock qp used for that frame (note that this argument is also true if a frame has been coded as two separate fields using different qp values). If you would like to see more refined statistics about the QP values used, or even other MB/block level statistics, you may want to enable trace file generation (see #define TRACE in defines.h) or consider the use of a bitstream analyzer. The selection of the Intra period strongly relates to the application you have in mind and not just bitrate. Such could affect random access and trick mode capabilities (this also includes channel change), error robustness etc. Progressive refresh (i.e. enforce coding of certain regions in Inter slices using "constrained intra predicted" MBs) could be used as an alternative. However, note that both methods have different benefits or drawbacks, such as in terms of Rate Control design/Rate fluctuations, positive or negative impact on Rate/Distortion performance (i.e. progressive refresh could result in "motion field gaps" while you may have a bit less efficient intra MB mode type signaling within inter slices), subjective artifacts for certain applications (i.e. during channel change some users may not feel as comfortable with the impact of Progressive refresh on the decoded stream) etc. Performance could vary also depending on content (low/high motion, texture etc), presence of frequent scene changes etc. Some more recent methods actually propose the use of a secondary lower quality, usually intra only, bitstream as an alternative which introduces some small drift during a "relaxation" period. In terms of error robustness, note also that the actual bitrate could also play significant role (i.e. if for the same quality you could further compress a sequence, then impact of a given error rate might be different). In any case, I guess you have to figure out yourself how important intra refresh is for your application, since also as you have observed the use of frequent intra slices could result in significant performance loss, which can be even greater at lower rates. In terms of artifacts, some things that have been observed in the past include detail loss, flickering in "all intra" sequences but also when transitioning from inter to intra frames, lower chroma quality etc. However some of these problems were related to non-normative decisions (fixed quantization, thresholding and rounding of coefficients during quantization, motion estimation/mode decision, quantization values used for the encoding of chroma components, deblocking parameters etc), while others were partially resolved by the consideration of the 8x8 transform and Quantization matrices in the FRExt profiles (these tools are not however supported in baseline, main, and extended profiles). Pre-processing strategies could also have an important impact in terms of subjective quality, while their design could impact differently different codecs. In terms of flickering you may find some analysis (it does not consider some of the above conditions that I have mentioned, while use of B slices could severely reduce flickering artifacts) in http://ftp3.itu.int/av-arch/jvt-site/2002_10_Geneva/JVT-E070r2.doc. I hope the above helps. Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Tobias Wolff Sent: Monday, June 20, 2023 8:12 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] H.264: Quantizationparameter, Intraperiod, artifacts Dear Experts, I am new to videocoding and currently doing some experiments with H.264 in the baseline Profile. However, there are some things which I don?t quite understand... Is the Quantizationparameter assigned to the whole frame or to each Macroblock (which I thought is the case)? In case it is assigned to each MB, then why does the Encoder give me one number (for the Quantizationparameter QP) per Frame in its comandline output? How is the Intraperiod handled in a reasonable way, depending on the bitrate? Is there a reason for chosing "only the first frame as an Intraframe" as the default setting? What would be a reason for chosing a small Intraperiod at low bitrates? And what is concidered to be a very small or very large Intraperiod? Since P-Frames allow for intracoded blocks and the Encoder decides on its own how to code each block (within a P-Frame), I don?t really see a reason to force the encoder to use expensive I-Frames in a fixed period, exept for Random Access. Can anybody help me with that? Last but not least, I?d really appreciate any hint to a survey about the artifacts exhibited by H.264. Does anybody know to what extend this has been investigated? Thanks a lot in advance, Best regards, Tobias _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From jingli888ca 163.com Thu Jun 23 11:34:37 2005 From: jingli888ca 163.com (=?gb2312?B?SmluZyBMaQ==?=) Date: Thu Jun 23 04:33:50 2005 Subject: [Mp4-tech] Inter prediction and de-blocking work in parallel?? Message-ID: <42BA1FBD.000082.17311@m0201.mail.163.com> Hello, Experts: I have a question on making inter prediction and de-blocking work in parallel in baselie profile FMO. When inter prediction processes an MB in the current frame, it refers to pixels in previous frame, and these pixels have to be de-blocked already. Since de-blocking processes the MBs in the last frame, if the current MB refers to the pixels 100 MB ahead of the MB being de-blocked, then inter prediction cannot start until de-blocking catches up. This will cause the system performance wait for a long time. Is this designed intentionly in baseline profile FMO? Is there a way to make inter prediction and de-blocking work in parallel? Thanks. Jing -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/9b37fb8b/attachment.html From mouliiniitkgp yahoo.co.in Thu Jun 23 06:48:13 2005 From: mouliiniitkgp yahoo.co.in (chandra mouli) Date: Thu Jun 23 04:33:57 2005 Subject: [Mp4-tech] input data to h.264 filter Message-ID: <20050623044813.32273.qmail@web8310.mail.in.yahoo.com> hi, experts Sorry in advance for asking a foolish question. I want to work on h.264 deblocking filter. I worked on Mpeg-4 filter but tis one is new for me.I want to know what all values the h.264 standard sends for the deblocking filter for filtering to take place . I read filtering depends on boundary strength which depends on whether the vop is intracoded and macroblock or not etc. But i want to know what all values/data it receives for filtering to take place.Thanx in advance for reply best regards, Mouli --------------------------------- Free antispam, antivirus and 1GB to save all your messages Only in Yahoo! Mail: http://in.mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/edf47096/attachment.html From sdey pace.stpp.soft.net Thu Jun 23 11:38:38 2005 From: sdey pace.stpp.soft.net (Soumen Kumar Dey) Date: Thu Jun 23 04:34:04 2005 Subject: [Mp4-tech] About intra prediction References: <20050622145749.31392.qmail@web52301.mail.yahoo.com> Message-ID: <004501c577b1$9dc07640$8564a8c0@psil> In DC mode, we can predict all the 4x4 and 16x16 samples with a predicted block having all the elements 128 if left and above block is not available. best regards, Soumen ----- Original Message ----- From: atul tomar To: mp4-tech@lists.mpegif.org Sent: Wednesday, June 22, 2023 8:27 PM Subject: [Mp4-tech] About intra prediction Dear all, I have a question on intra prediction in h.264 standard. Would you please explain it to me. If the Intra4x4 subMB is the first subMB in the first MB(upper-left of frame) of an IDR-picture, and I want to decide predIntra4x4PredMode for this subMB. How to decide this beacuse it does not have any above and left SubMBs or any samples? If it is set to Mode2(DC prediction mode) even this mode require above and left samples to get Mean of these samples for prediction. Same case is with Intra16x16 prediction for first MB of an IDR or I picture. Thanks in advance. Regards, Atul ------------------------------------------------------------------------------ Do you Yahoo!? Make Yahoo! your home page ------------------------------------------------------------------------------ _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/ba7f276d/attachment.html From amni_gupta yahoo.co.in Thu Jun 23 07:34:02 2005 From: amni_gupta yahoo.co.in (Amni Gupta) Date: Thu Jun 23 04:34:10 2005 Subject: [Mp4-tech] Doubt- Please write Message-ID: <20050623053402.21931.qmail@web8604.mail.in.yahoo.com> Hi; I have a doubt. Please clarify. Does the simple profile of MPEg-4 natural video support Overlapped motion compensation? I also have the same doubt for the Horizontal and Vertical padding sometimes done before motion compensation. Since in Simple profile the shape is rectangular and the full frame is a VOP. Is padding required before Motion compensation? What is the exact use of this padding? Please can any one clarify. Regards Amni. --------------------------------- Too much spam in your inbox? Yahoo! Mail gives you the best spam protection for FREE! http://in.mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/aa3af002/attachment.html From robin94539 yahoo.com Wed Jun 22 23:40:59 2005 From: robin94539 yahoo.com (Robin Zoo) Date: Thu Jun 23 04:34:18 2005 Subject: [MP4-tech] intra8x8 mode Message-ID: <20050623054059.6298.qmail@web30205.mail.mud.yahoo.com> Dear experts, I noticed that transform_8x8_mode_flag "shall not be present in picture parameter sets" for Baseline/Main/Extended profiles, which means that 8x8 transform decoding process is not in use. I'd like to know whether intra8x8 prediction mode is used in the profiles or not. If yes, is the 4x4 based inverse transform being used accordingly to get the reconstructed picture? thanks in advance! Robin __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From iamsangu gmail.com Thu Jun 23 12:16:52 2005 From: iamsangu gmail.com (lisa sangs) Date: Thu Jun 23 04:34:25 2005 Subject: [Mp4-tech] DRM In-Reply-To: <5F09D220B62F79418461A978CA0921BD36A898@pslexc01.psl.local> References: <5F09D220B62F79418461A978CA0921BD36A898@pslexc01.psl.local> Message-ID: <42BA4CCC.7070706@gmail.com> Thanks for your immediate reply. But i want to know about DRM in audio domain(like DRM in MPEG4 audio, WMA) so can u please tell me can u please tell me 1. from which site i can get the documents about DRM 2. from where i can download source code for DRM thanks in advance lisa sangs Zhong Yang Huang wrote: >Hi, > >MPEG-4 System has defined a section about an interoperable and flexible DRM solution (IPMP - Intellectual Property Management and Protection). You could find the final standard specification on IPMP extension in N5282 (MPEG Shanghai meeting) and FAQ in N5790 (MPEG Trondheim meeting). MPEG-4 IPMPX also has its reference software implementation in IM1, which is located in MPEG CVS server. > >Best regards, >Huang Zhongyang >Panasonic > > > > >>-----Original Message----- >>From: mp4-tech-bounces@lists.mpegif.org >>[mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of sangeeth >>Sent: 2005?6?23? 1:18 >>To: mp4-tech@lists.mpegif.org; ralph.sperschneider@iis.fraunhofer.de >>Subject: [Mp4-tech] DRM >> >>hi >> I want to know about Digital rights management. (DRM). I >>am new to this technology. >>can u please tell me >>1. from which site i can get the documents about DRM 2. from >>where i can download source code for DRM >> >>thanks in advance >>lisa sangs >> >> >>_______________________________________________ >>NOTE: Please use clear subject lines for your posts. Include >>[audio, [video], [systems], [general] or another apppropriate >>identifier to indicate the type of question you have. >> >>Note: Conduct on the mailing list is subject to the Antitrust >>guidelines found at >>http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti >>trust.php >> >> >> > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/f36fca21/attachment.html From tomar_atul2000 yahoo.com Thu Jun 23 00:54:54 2005 From: tomar_atul2000 yahoo.com (atul tomar) Date: Thu Jun 23 04:34:31 2005 Subject: [Mp4-tech] About map unit to slice group map. In-Reply-To: <91D7F2CEE3425A4A9D11311D09FCE2460F20DE2C@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Message-ID: <20050623065454.54442.qmail@web52305.mail.yahoo.com> Hi, Thanks for pointing me to exact information. As in 7.4.2.2 there are three options: I have doubt in third last option, where (frame_mbs_only_flag is equal to 0 and mb_adaptive_frame_field_flag is equal to 0 and the coded picture is a frame), the slice group map units are units of two macroblocks that are vertically contiguous as in a frame macroblock pair of an MBAFF frame. My question is here, " How can we say that coded picture is frame?", if frame_mbs_only_flag is equal to 0 which says that macroblock units are taken from picture which is complete frame not filed. Second please correct me if I am wrong, In baseline profile we set frame_mbs_only_flag equal to 1, mb_adaptive_frame_field_flag is equal to 0, field_pic_flag is 0 then coded picture is frame so the slice group map units are units of macroblock. Thanks, Regards, Atul Gary Sullivan wrote: To be more precise, see the following excerpts from the specification itself. Note that there is another case besides the MBAFF case in which the map units are not individual macroblocks (the last case listed below). 3.82 map unit to slice group map: A means of mapping slice group map units of a picture into slice groups. The map unit to slice group map consists of a list of numbers, one for each slice group map unit, specifying the slice group to which each coded slice group map unit belongs. 3.139 slice group map units: The units of the map unit to slice group map. 7.4.2.2 Slice group map units are specified as follows. - If frame_mbs_only_flag is equal to 0 and mb_adaptive_frame_field_flag is equal to 1 and the coded picture is a frame, the slice group map units are macroblock pair units. - Otherwise, if frame_mbs_only_flag is equal to 1 or a coded picture is a field, the slice group map units are units of macroblocks. - Otherwise (frame_mbs_only_flag is equal to 0 and mb_adaptive_frame_field_flag is equal to 0 and the coded picture is a frame), the slice group map units are units of two macroblocks that are vertically contiguous as in a frame macroblock pair of an MBAFF frame. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of +> Ye-Kui.Wang@nokia.com +> Sent: Monday, June 20, 2023 11:01 PM +> To: tomar_atul2000@yahoo.com; mp4-tech@lists.mpegif.org +> Subject: RE: [Mp4-tech] About map unit to slice group map. +> +> Hi Atul, +> +> "map unit" may either be a macroblock or a macroblock pair +> if MBAFF (macroblock adaptive frame field coding) is in use. +> "map unit to slice group map" tells the informaiton to which +> slice group each macroblock or macroblock pair belongs. +> +> BR, YK +> +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext atul tomar +> Sent: 16 June, 2005 17:15 +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] About map unit to slice group map. +> +> +> Hi All, +> +> Can anyone please explain me what standard mean by " map +> unit to slice group map". I read ITU standard many times to +> understand this but could not get it clear. +> Please expalin what does "map unit" represent. +> +> Thanks in advance. +> +> Reg, +> Atul +> +> +> +> _____ +> +> Discover Yahoo! +> Find restaurants, movies, travel & more fun for the weekend. +> Check +> +> weekend.html> it out! +> +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php --------------------------------- Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050622/394a363d/attachment.html From asic_soc 163.com Thu Jun 23 16:59:33 2005 From: asic_soc 163.com (Shen Sha) Date: Thu Jun 23 04:34:38 2005 Subject: [Mp4-tech] [video][mpeg2] Where can I get the MPEG2 TS stream? References: <200506221602.j5MG1AYJ002360@lists1.magma.ca> Message-ID: <001b01c577c9$7e189800$4d3210ac@TMT.COM> Hi all, Where can I get the MPEG2 TS stream? Thanks a lot. Shen Sha From gilberto ensilica.com Thu Jun 23 10:04:37 2005 From: gilberto ensilica.com (Gilberto) Date: Thu Jun 23 04:34:44 2005 Subject: [Mp4-tech] About intra prediction In-Reply-To: <20050622145749.31392.qmail@web52301.mail.yahoo.com> Message-ID: <20050623081615.8C6D84E54B@newint.iuma.ulpgc.es> Atul, For the first MB and/or subBlock you will use DC_128. All values equal to 128. DC has 4 modes DC_LEFT, DC_RIGHT, DC_128, DC_ALL Regards, Gilberto _____ From: atul tomar [mailto:tomar_atul2000@yahoo.com] Sent: 22 June 2023 15:58 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] About intra prediction Dear all, I have a question on intra prediction in h.264 standard. Would you please explain it to me. If the Intra4x4 subMB is the first subMB in the first MB(upper-left of frame) of an IDR-picture, and I want to decide predIntra4x4PredMode for this subMB. How to decide this beacuse it does not have any above and left SubMBs or any samples? If it is set to Mode2(DC prediction mode) even this mode require above and left samples to get Mean of these samples for prediction. Same case is with Intra16x16 prediction for first MB of an IDR or I picture. Thanks in advance. Regards, Atul _____ Do you Yahoo!? Make Yahoo! your home page -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/126b88d4/attachment.html From tma iis.fhg.de Thu Jun 23 13:43:50 2005 From: tma iis.fhg.de (Herbert Thoma) Date: Thu Jun 23 08:07:16 2005 Subject: [Mp4-tech] Doubt- Please write In-Reply-To: <20050623053402.21931.qmail@web8604.mail.in.yahoo.com> References: <20050623053402.21931.qmail@web8604.mail.in.yahoo.com> Message-ID: <42BA9266.3070307@iis.fhg.de> Amni Gupta wrote: > Hi; > I have a doubt. Please clarify. Does the simple profile of MPEg-4 > natural video support Overlapped motion compensation? Simple Profile does not support OBMC. OBMC is in the standard, but it is not used in any profile. > I also have the same doubt for the Horizontal and Vertical padding > sometimes done before motion compensation. Since in Simple profile the > shape is rectangular and the full frame is a VOP. Is padding > required before Motion compensation? What is the exact use of this padding? When coding shaped objects you need the padding process to fill up "holes" in the shape. In Simple Profile you may need padding for motion vectors that point outside the VOP. You can also use vector clipping in motion compensation for this. Regards, Herbert. > Please can any one clarify. > > Regards > Amni. > > ------------------------------------------------------------------------ > Too much spam in your inbox? Yahoo! Mail gives you the best spam > protection for FREE! > http://in.mail.yahoo.com > > > ------------------------------------------------------------------------ > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -- Herbert Thoma Group Manager Video Multimedia Realtime Systems Department Fraunhofer IIS Am Wolfsmantel 33, 91058 Erlangen, Germany Phone: +49-9131-776-323 Fax: +49-9131-776-399 email: tma@iis.fhg.de www: http://www.iis.fhg.de/ From ycchang smediatech.com Thu Jun 23 21:38:54 2005 From: ycchang smediatech.com (=?big5?B?sWmlw7JN?=) Date: Thu Jun 23 15:15:27 2005 Subject: [Mp4-tech] [H.264] inconsistent of intra prediction between standard and JM9.6 Message-ID: Dear all: H.264 standard 8.3.1.1 & (8-42) describes how to decide the prediction mode of a 4x4 block. The corresponding code in JM9.6 decoder is in macroblock.c\read_ipred_modes(). I will explain it by an example. According to the standard, if the block to the left of a 4x4 block is not available, the intra4x4PredModeModeA (variable leftIntraPredMode in JM) is set to 2. In our example, the block to the up is available, and intra4x4PredModeModeB (variable upIntraPredMode) is 0. According to the standard, the predIntra4x4PredMode (variable mostProbableIntraPredMode) is the minimal one of these two value, and therefore it should be 0. However, if this situation is decoded by JM, it set leftIntraPredMode = -1and upIntraPredMode = 0, and then decide the mostProbableIntraPredMode as: mostProbableIntraPredMode = (upIntraPredMode < 0 || leftIntraPredMode < 0) ? DC_PRED : upIntraPredMode < leftIntraPredMode ? upIntraPredMode : leftIntraPredMode; Because leftIntraPredMode < 0, so mostProbableIntraPredMode = DC_PRED = 2. Could anyone help me with this problem? Is there any information I missed? Best regards, Yung-Ching Chang From garysull windows.microsoft.com Thu Jun 23 11:23:18 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Thu Jun 23 15:21:49 2005 Subject: [Mp4-tech] Doubt- Please write Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F27875E@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Regarding the "vector clipping" idea, I think it is worth cautioning that such vector clipping would need to be performed on a sample-by-sample basis rather than on a block-wise basis, or else padding would be necessary. There are many cases that would involve some samples of a prediction block lying outside the VOP while others lie inside of it. Clipping the MV on a block-wise basis (without also performing padding) would produce incorrect results. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Herbert Thoma +> Sent: Thursday, June 23, 2023 3:44 AM +> To: Amni Gupta +> Cc: mp4-tech@lists.mpegif.org +> Subject: Re: [Mp4-tech] Doubt- Please write +> +> Amni Gupta wrote: +> > Hi; +> > I have a doubt. Please clarify. Does the simple profile +> of MPEg-4 +> > natural video support Overlapped motion compensation? +> +> Simple Profile does not support OBMC. +> OBMC is in the standard, but it is not used in any profile. +> +> > I also have the same doubt for the Horizontal and +> Vertical padding +> > sometimes done before motion compensation. Since in Simple +> profile the +> > shape is rectangular and the full frame is a VOP. Is padding +> > required before Motion compensation? What is the exact use +> of this padding? +> +> When coding shaped objects you need the padding process to fill up +> "holes" in the shape. +> +> In Simple Profile you may need padding for motion vectors that point +> outside the VOP. You can also use vector clipping in motion +> compensation +> for this. +> +> Regards, +> Herbert. +> +> > Please can any one clarify. +> > +> > Regards +> > Amni. +> > +> > +> ------------------------------------------------------------- +> ----------- +> > Too much spam in your inbox? Yahoo! Mail gives you the best spam +> > protection for FREE! +> > http://in.mail.yahoo.com +> > +> > +> > +> ------------------------------------------------------------- +> ----------- +> > +> > _______________________________________________ +> > NOTE: Please use clear subject lines for your posts. +> Include [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> > +> > Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> +> -- +> Herbert Thoma +> Group Manager Video +> Multimedia Realtime Systems Department +> Fraunhofer IIS +> Am Wolfsmantel 33, 91058 Erlangen, Germany +> Phone: +49-9131-776-323 +> Fax: +49-9131-776-399 +> email: tma@iis.fhg.de +> www: http://www.iis.fhg.de/ +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From lyzhou equator.com Thu Jun 23 11:50:48 2005 From: lyzhou equator.com (LiyongZhou) Date: Thu Jun 23 15:28:05 2005 Subject: [MP4-tech] intra8x8 mode In-Reply-To: <20050623054059.6298.qmail@web30205.mail.mud.yahoo.com> Message-ID: <009101c5781c$1728bb90$a610000a@win.equator.com> Robin, As you noticed, Transform8x8 is for FRExt (high profiles). However, this has nothing to do with intra8x8 prediction mode. If Transform8x8 flag is off, you use transform 4x4 no matter what the prediction mode is (can be 16x16, 16x8, 8x8, or 4x4 etc). --leon > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Robin Zoo > Sent: 2005?6?22? 22:41 > To: mp4-tech@lists.mpegif.org > Subject: [MP4-tech] intra8x8 mode > > > Dear experts, > > I noticed that transform_8x8_mode_flag "shall not be > present in picture parameter sets" for > Baseline/Main/Extended profiles, which means that 8x8 > transform decoding process is not in use. I'd like to > know whether intra8x8 prediction mode is used in the > profiles or not. If yes, is the 4x4 based inverse > transform being used accordingly to get the > reconstructed picture? > > thanks in advance! > > Robin > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php From lyzhou equator.com Thu Jun 23 12:06:32 2005 From: lyzhou equator.com (LiyongZhou) Date: Thu Jun 23 15:35:23 2005 Subject: [Mp4-tech] Inter prediction and de-blocking work in parallel?? In-Reply-To: <42BA1FBD.000082.17311@m0201.mail.163.com> Message-ID: <009201c5781e$49ddc5b0$a610000a@win.equator.com> DB and MC can not be done in 100% parallel. Even without FMO, there is a potential racing condition between the two: consider the case when a mv of a top MB points to the bottom MB of the previous being deblocked. So you have to run a Deblock-Progress-Checker in your MC. The good news is that this kind of "checker" is not that difficult: Deblock is performed in raster-scan order (even decoding is in FMO or ASO): one X and one Y check will be enough, plus the reference frame index. -- leon > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Jing Li > Sent: 2005?6?22? 19:35 > To: mp4-tech@lists.mpegif.org > Cc: garysull@windows.microsoft.com > Subject: [Mp4-tech] Inter prediction and de-blocking work in > parallel?? > > > Hello, Experts: > > I have a question on making inter prediction and de-blocking > work in parallel in baselie profile FMO. > > When inter prediction processes an MB in the current frame, > it refers to pixels in previous frame, and these pixels have > to be de-blocked already. Since de-blocking processes the MBs > in the last frame, if the current MB refers to the pixels 100 > MB ahead of the MB being de-blocked, then inter prediction > cannot start until de-blocking catches up. This will cause > the system performance wait for a long time. > > Is this designed intentionly in baseline profile FMO? Is > there a way to make inter prediction and de-blocking work in parallel? > > Thanks. > > Jing > From garysull windows.microsoft.com Thu Jun 23 13:44:49 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Fri Jun 24 04:30:15 2005 Subject: [Mp4-tech] [H.264] inconsistent of intra prediction betweenstandard and JM9.6 Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F278A67@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> To Yung-Ching Chang (et al), I think you may have missed the concept of "dcPredModePredictedFlag" (which used to be called something else - hopefully you're looking at a recent version of the standard text). Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ??? +> Sent: Thursday, June 23, 2023 5:39 AM +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] [H.264] inconsistent of intra prediction +> betweenstandard and JM9.6 +> +> Dear all: +> +> H.264 standard 8.3.1.1 & (8-42) describes how to decide the +> prediction mode of a 4x4 block. +> The corresponding code in JM9.6 decoder is in +> macroblock.c\read_ipred_modes(). +> +> I will explain it by an example. According to the standard, +> if the block to the left of a 4x4 block is not available, +> the intra4x4PredModeModeA (variable leftIntraPredMode in JM) +> is set to 2. In our example, the block to the +> up is available, and intra4x4PredModeModeB (variable +> upIntraPredMode) is 0. According to the standard, +> the predIntra4x4PredMode (variable +> mostProbableIntraPredMode) is the minimal one of these two value, +> and therefore it should be 0. +> +> However, if this situation is decoded by JM, it set +> leftIntraPredMode = -1and upIntraPredMode = 0, and then +> decide the mostProbableIntraPredMode as: +> +> mostProbableIntraPredMode = (upIntraPredMode < 0 || +> leftIntraPredMode < 0) ? DC_PRED : upIntraPredMode < +> leftIntraPredMode ? upIntraPredMode : leftIntraPredMode; +> +> Because leftIntraPredMode < 0, so mostProbableIntraPredMode +> = DC_PRED = 2. +> +> Could anyone help me with this problem? Is there any +> information I missed? +> +> Best regards, +> +> Yung-Ching Chang +> +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From garysull windows.microsoft.com Thu Jun 23 13:59:55 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Fri Jun 24 04:30:21 2005 Subject: [MP4-tech] intra8x8 mode Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F278ACF@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Leon, Robin, et al, Actually, that's not the whole story. The Intra_8x8 prediction process can only be invoked if transform_size_8x8_flag is 1. But transform_size_8x8_flag is not present and has an inferred value of 0 when transform_8x8_mode_flag is equal to 0, and transform_8x8_mode_flag has an inferred value of 0 when it is not present. As Robin noted, transform_8x8_mode_flag is not present in the Baseline, Main, and Extended profiles. To summarize, 8x8 intra prediction is never invoked unless the 8x8 transform type is invoked for the macroblock. So 8x8 intra prediction is never used with a 4x4 residual transform, and it is never used with any non-FRExt profiles. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of LiyongZhou +> Sent: Thursday, June 23, 2023 10:51 AM +> To: 'Robin Zoo'; mp4-tech@lists.mpegif.org +> Subject: RE: [MP4-tech] intra8x8 mode +> +> Robin, +> +> As you noticed, Transform8x8 is for FRExt (high profiles). +> However, this +> has nothing to do with intra8x8 prediction mode. If +> Transform8x8 flag is +> off, you use transform 4x4 no matter what the prediction +> mode is (can be +> 16x16, 16x8, 8x8, or 4x4 etc). +> --leon +> +> > -----Original Message----- +> > From: mp4-tech-bounces@lists.mpegif.org +> > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Robin Zoo +> > Sent: 2005?6?22? 22:41 +> > To: mp4-tech@lists.mpegif.org +> > Subject: [MP4-tech] intra8x8 mode +> > +> > +> > Dear experts, +> > +> > I noticed that transform_8x8_mode_flag "shall not be +> > present in picture parameter sets" for +> > Baseline/Main/Extended profiles, which means that 8x8 +> > transform decoding process is not in use. I'd like to +> > know whether intra8x8 prediction mode is used in the +> > profiles or not. If yes, is the 4x4 based inverse +> > transform being used accordingly to get the +> > reconstructed picture? +> > +> > thanks in advance! +> > +> > Robin +> > +> > +> > __________________________________________________ +> > Do You Yahoo!? +> > Tired of spam? Yahoo! Mail has the best spam protection around +> > http://mail.yahoo.com +> > _______________________________________________ +> > NOTE: Please use clear subject lines for your posts. Include +> > [audio, [video], [systems], [general] or another apppropriate +> > identifier to indicate the type of question you have. +> > +> > Note: Conduct on the mailing list is subject to the Antitrust +> > guidelines found at +> > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti +> trust.php +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant itrust.php +> From alexismt comcast.net Thu Jun 23 21:51:34 2005 From: alexismt comcast.net (Alexis Michael Tourapis) Date: Fri Jun 24 04:30:28 2005 Subject: [MP4-tech] intra8x8 mode Message-ID: <062320052051.7451.42BB20D5000D226B00001D1B22070206539B039C07980A040E@comcast.net> Dear Leon, Actually to clarify, intra_8x8 is only allowed when transform_size_8x8_flag is equal to 1. Please check table 7-11. Since this is not available in non FRExt profiles (i.e. transform_8x8_mode_flag is inferred as 0) then this mode is also implicitly not allowed in these profiles. Best regards, Alexis -------------- Original message -------------- > Robin, > > As you noticed, Transform8x8 is for FRExt (high profiles). However, this > has nothing to do with intra8x8 prediction mode. If Transform8x8 flag is > off, you use transform 4x4 no matter what the prediction mode is (can be > 16x16, 16x8, 8x8, or 4x4 etc). > --leon > > > -----Original Message----- > > From: mp4-tech-bounces@lists.mpegif.org > > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Robin Zoo > > Sent: 2005$BG/(B6$B7n(B22$BF|(B 22:41 > > To: mp4-tech@lists.mpegif.org > > Subject: [MP4-tech] intra8x8 mode > > > > > > Dear experts, > > > > I noticed that transform_8x8_mode_flag "shall not be > > present in picture parameter sets" for > > Baseline/Main/Extended profiles, which means that 8x8 > > transform decoding process is not in use. I'd like to > > know whether intra8x8 prediction mode is used in the > > profiles or not. If yes, is the 4x4 based inverse > > transform being used accordingly to get the > > reconstructed picture? > > > > thanks in advance! > > > > Robin > > > > > > __________________________________________________ > > Do You Yahoo!? > > Tired of spam? Yahoo! Mail has the best spam protection around > > http://mail.yahoo.com > > _______________________________________________ > > NOTE: Please use clear subject lines for your posts. Include > > [audio, [video], [systems], [general] or another apppropriate > > identifier to indicate the type of question you have. > > > > Note: Conduct on the mailing list is subject to the Antitrust > > guidelines found at > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti > trust.php > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], > [systems], [general] or another apppropriate identifier to indicate the type of > question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found > at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050623/26c7f4c6/attachment-0001.html From yi-jen.chiu intel.com Thu Jun 23 14:59:48 2005 From: yi-jen.chiu intel.com (Chiu, Yi-jen) Date: Fri Jun 24 04:30:35 2005 Subject: [Mp4-tech] [H.264] inconsistent of intra prediction betweenstandard and JM9.6 Message-ID: <2F226B6DEEDB8F4299E206C7DA5313D0041E1651@scsmsx403.amr.corp.intel.com> Hi, In your example, if the block to the left of a 4x4 block is Not available, dcPredModePredictedFlag will be equal to 1. With dcPredModePredictedFlag=1, both intraMxMPredModeA and intraMxMPredModeB will be set equal to 2 accoring to the standard. Thus, the predIntra4x4PredMode will be equal to 2. Yi-Jen -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ycchang@smediatech.com Sent: Thursday, June 23, 2023 5:39 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] inconsistent of intra prediction betweenstandard and JM9.6 Dear all: H.264 standard 8.3.1.1 & (8-42) describes how to decide the prediction mode of a 4x4 block. The corresponding code in JM9.6 decoder is in macroblock.c\read_ipred_modes(). I will explain it by an example. According to the standard, if the block to the left of a 4x4 block is not available, the intra4x4PredModeModeA (variable leftIntraPredMode in JM) is set to 2. In our example, the block to the up is available, and intra4x4PredModeModeB (variable upIntraPredMode) is 0. According to the standard, the predIntra4x4PredMode (variable mostProbableIntraPredMode) is the minimal one of these two value, and therefore it should be 0. However, if this situation is decoded by JM, it set leftIntraPredMode = -1and upIntraPredMode = 0, and then decide the mostProbableIntraPredMode as: mostProbableIntraPredMode = (upIntraPredMode < 0 || leftIntraPredMode < 0) ? DC_PRED : upIntraPredMode < leftIntraPredMode ? upIntraPredMode : leftIntraPredMode; Because leftIntraPredMode < 0, so mostProbableIntraPredMode = DC_PRED = 2. Could anyone help me with this problem? Is there any information I missed? Best regards, Yung-Ching Chang _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From nam817 ms25.hinet.net Fri Jun 24 11:52:09 2005 From: nam817 ms25.hinet.net (ChiChang_Kuo) Date: Fri Jun 24 04:30:42 2005 Subject: [Mp4-tech] [H.264] inconsistent of intra prediction betweenstandard and JM9.6 References: Message-ID: <42BB7559.FF59CA0D@ms25.hinet.net> Hi, After I read the standard document "ITU-T H.264 (03/2005)" section 8.3.1.1, It describes that: =============================================================================================== predIntra4x4PredMode = Min( intraMxMPredModeA, intraMxMPredModeB ) //!!!HERE!!! if( prev_intra4x4_pred_mode_flag[ luma4x4BlkIdx ] ) Intra4x4PredMode[ luma4x4BlkIdx ] = predIntra4x4PredMode else (8-42) if( rem_intra4x4_pred_mode[ luma4x4BlkIdx ] < predIntra4x4PredMode ) Intra4x4PredMode[ luma4x4BlkIdx ] = rem_intra4x4_pred_mode[ luma4x4BlkIdx ] else Intra4x4PredMode[ luma4x4BlkIdx ] = rem_intra4x4_pred_mode[ luma4x4BlkIdx ] + 1 =============================================================================================== where 'intraMxMPredModeA' and 'intraMxMPredModeB' are set to 2 (DC_PRED) if 'dcPredModePredictedFlag' is equal to 1. And the document says that: =============================================================================================== The variable 'dcPredModePredictedFlag' is derived as follows. - If any of the following conditions are true, dcPredModePredictedFlag is set equal to 1 - the macroblock with address mbAddrA is not available - the macroblock with address mbAddrB is not available - the macroblock with address mbAddrA is available and coded in Inter prediction mode and constrained_intra_pred_flag is equal to 1 - the macroblock with address mbAddrB is available and coded in Inter prediction mode and constrained_intra_pred_flag is equal to 1 - Otherwise, dcPredModePredictedFlag is set equal to 0. =============================================================================================== So, as my point of view, the behavior of JM S/W matching that in the standard document. Hope this is helpful. Regards, Chi-Chang Kuo ąiĨÃ˛M wrote: > Dear all: > H.264 standard 8.3.1.1 & (8-42) describes how to decide the prediction mode of a 4x4 block. > The corresponding code in JM9.6 decoder is in macroblock.c\read_ipred_modes(). > I will explain it by an example. According to the standard, if the block to the left of a 4x4 block is not available, > the intra4x4PredModeModeA (variable leftIntraPredMode in JM) is set to 2. In our example, the block to the > up is available, and intra4x4PredModeModeB (variable upIntraPredMode) is 0. According to the standard, > the predIntra4x4PredMode (variable mostProbableIntraPredMode) is the minimal one of these two value, > and therefore it should be 0. > However, if this situation is decoded by JM, it set leftIntraPredMode = -1and upIntraPredMode = 0, and then > decide the mostProbableIntraPredMode as: > mostProbableIntraPredMode = (upIntraPredMode < 0 || leftIntraPredMode < 0) ? DC_PRED : upIntraPredMode < leftIntraPredMode ? upIntraPredMode : leftIntraPredMode; > Because leftIntraPredMode < 0, so mostProbableIntraPredMode = DC_PRED = 2. > Could anyone help me with this problem? Is there any information I missed? > Best regards, > Yung-Ching Chang -- Chi-Chang Kuo (ŗĸ¨äŠ÷) Digital Signal Processing Tech. Dept., V300, (ŧÆĻė°T¸šŗB˛z§ŪŗNŗĄV300) Computer & Communications Research Laboratories, (šq¸ŖģPŗq°T¤uˇ~Ŧã¨sŠŌ) Industrial Technology Research Institute (¤uˇ~§ŪŗNŦã¨s°|) ADDR: PO Box 7-297, Hsinchu 300, Taiwan (300 ˇsĻËļlŦF 7-297 ¸šĢHŊc) TEL: +886-3-5914784, FAX: +886-3-5829731 From ycchang smediatech.com Fri Jun 24 12:08:09 2005 From: ycchang smediatech.com (=?big5?B?sWmlw7JN?=) Date: Fri Jun 24 04:30:48 2005 Subject: [Mp4-tech] [H.264] inconsistent of intra prediction between standard and JM9.6 Message-ID: Dear all: Thanks for the kindly help of Sullivan, Chiu and Kuo. I found I miss-understand the standard. I do read the description about dcOnlyPredictionFlag, but I thought that there are dcOnlyPredictionFlag(A) and dcOnlyPredictionFlag(B) effect only the corresponding intra4x4PredModeA or intra4x4PredModeB. But the truth is that only one dcOnlyPredictionFlag, and in our example, both intra4x4PredModeA and intra4x4PredModeB are effected by this dcOnlyPredictionFlag to become Intra_4x4_DC. So the result predIntra4x4PredMode should be Intra_4x4_DC. Thanks for your help once more. Yung-Ching Chang -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050624/0323a389/attachment.html From sathyavratha cisctechnology.com Fri Jun 24 11:23:53 2005 From: sathyavratha cisctechnology.com (sathyavratha) Date: Fri Jun 24 04:30:55 2005 Subject: [MP4-tech] padding issue Message-ID: <000e01c57878$ba982930$a101a8c0@CISCDESKTOP029> hi experts, we are developing mpeg-4 simple profile visual decoder and encoder IP's my doubt is whether vector padding is necessary before MV decoding for simple profile to get motion compensated output. please let me know the steps for developing motion compensation. thanx in advance. sathyavratha -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050624/0a8bca52/attachment.html From jingli888ca 163.com Fri Jun 24 15:56:07 2005 From: jingli888ca 163.com (=?gb2312?B?SmluZyBMaQ==?=) Date: Fri Jun 24 04:31:02 2005 Subject: [Mp4-tech] Parallel Processing of Y, U and V?? Message-ID: <42BBAE87.00003F.07559@m178.163.com> Hello, Experts: I wonder if it is a recommended way to parallely process the Y, U and V components of an MB in IDCT and inter prediction? What are cons and pros of doing so?? Is parallel processing more favourable over sequential processing of Y, U and V? Thanks. Jing -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050624/9d19de80/attachment.html From ksuehring web.de Fri Jun 24 11:22:12 2005 From: ksuehring web.de (Karsten Suehring) Date: Fri Jun 24 04:56:03 2005 Subject: [Mp4-tech] [H.264] inconsistent of intra prediction between standard and JM9.6 In-Reply-To: References: Message-ID: <42BBC2B4.1010209@web.de> Dear Yung-Ching Chang, I think you might be looking at an old version of the standard text where the decription was not very clear. >From current text: dcPredModePredictedFlag is set equal to 1 if ĄV the macroblock with address mbAddrA is not available ĄV the macroblock with address mbAddrB is not available [...] If dcPredModePredictedFlag is equal to 1 or the macroblock with address mbAddrN is not coded in Intra_4x4 or Intra_8x8 macroblock prediction mode, intraMxMPredModeN is set equal to 2 (Intra_4x4_DC prediction mode). Note that dcPredModePredictedFlag is derived only once, it does not differ for left or upper neighbour. So in your case dcPredModePredictedFlag is equal to one, which results in setting both intraMxMPredModeA and intraMxMPredModeB equal to 2. This is the same behaviour that your observed in the JM software. Best regards, Karsten ąiĨÃ˛M wrote: > Dear all: > > H.264 standard 8.3.1.1 & (8-42) describes how to decide the prediction mode of a 4x4 block. > The corresponding code in JM9.6 decoder is in macroblock.c\read_ipred_modes(). > > I will explain it by an example. According to the standard, if the block to the left of a 4x4 block is not available, > the intra4x4PredModeModeA (variable leftIntraPredMode in JM) is set to 2. In our example, the block to the > up is available, and intra4x4PredModeModeB (variable upIntraPredMode) is 0. According to the standard, > the predIntra4x4PredMode (variable mostProbableIntraPredMode) is the minimal one of these two value, > and therefore it should be 0. > > However, if this situation is decoded by JM, it set leftIntraPredMode = -1and upIntraPredMode = 0, and then > decide the mostProbableIntraPredMode as: > > mostProbableIntraPredMode = (upIntraPredMode < 0 || leftIntraPredMode < 0) ? DC_PRED : upIntraPredMode < leftIntraPredMode ? upIntraPredMode : leftIntraPredMode; > > Because leftIntraPredMode < 0, so mostProbableIntraPredMode = DC_PRED = 2. > > Could anyone help me with this problem? Is there any information I missed? > > Best regards, > > Yung-Ching Chang > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From ksuehring web.de Fri Jun 24 11:31:25 2005 From: ksuehring web.de (Karsten Suehring) Date: Fri Jun 24 04:56:10 2005 Subject: [Mp4-tech] About map unit to slice group map. In-Reply-To: <20050623065454.54442.qmail@web52305.mail.yahoo.com> References: <20050623065454.54442.qmail@web52305.mail.yahoo.com> Message-ID: <42BBC4DD.1020506@web.de> Atul, please see my comments inline. > I have doubt in third last option, where (frame_mbs_only_flag is equal > to 0 and mb_adaptive_frame_field_flag is equal to 0 and the coded > picture is a frame), the slice group map units are units of two > macroblocks that are vertically contiguous as in a frame macroblock pair > of an MBAFF frame. > My question is here, " How can we say that coded picture is frame?", if > frame_mbs_only_flag is equal to 0 which says that macroblock units are > taken from picture which is complete frame not filed. The second option applies to field pictures: "if frame_mbs_only_flag is equal to 1 or a coded picture is a field" The third option will only be applied if prevoius conditions didn't match. > Second please correct me if I am wrong, In baseline profile we set > frame_mbs_only_flag equal to 1, mb_adaptive_frame_field_flag is equal to > 0, field_pic_flag is 0 then coded picture is frame so the slice group > map units are units of macroblock. This is correct. Karsten > Thanks, > > Regards, > Atul > > */Gary Sullivan /* wrote: > > > To be more precise, see the following excerpts from the specification > itself. Note that there is another case besides the MBAFF case in which > the map units are not individual macroblocks (the last case listed > below). > > 3.82 map unit to slice group map: A means of mapping slice group map > units of a picture into slice groups. The map unit to slice group map > consists of a list of numbers, one for each slice group map unit, > specifying the slice group to which each coded slice group map unit > belongs. > > 3.139 slice group map units: The units of the map unit to slice group > map. > > 7.4.2.2 > > Slice group map units are specified as follows. > - If frame_mbs_only_flag is equal to 0 and > mb_adaptive_frame_field_flag is equal to 1 and the coded picture is a > frame, the slice group map units are macroblock pair units. > - Otherwise, if frame_mbs_only_flag is equal to 1 or a coded > picture is a field, the slice group map units are units of macroblocks. > - Otherwise (frame_mbs_only_flag is equal to 0 and > mb_adaptive_frame_field_flag is equal to 0 and the coded picture is a > frame), the slice group map units are units of two macroblocks that are > vertically contiguous as in a frame macroblock pair of an MBAFF frame. > > Best Regards, > > Gary Sullivan > > > +> -----Original Message----- > +> From: mp4-tech-bounces@lists.mpegif.org > +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > +> Ye-Kui.Wang@nokia.com > +> Sent: Monday, June 20, 2023 11:01 PM > +> To: tomar_atul2000@yahoo.com; mp4-tech@lists.mpegif.org > +> Subject: RE: [Mp4-tech] About map unit to slice group map. > +> > +> Hi Atul, > +> > +> "map unit" may either be a macroblock or a macroblock pair > +> if MBAFF (macroblock adaptive frame field coding) is in use. > +> "map unit to slice group map" tells the informaiton to which > +> slice group each macroblock or macroblock pair belongs. > +> > +> BR, YK > +> > +> -----Original Message----- > +> From: mp4-tech-bounces@lists.mpegif.org > +> [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of ext atul tomar > +> Sent: 16 June, 2005 17:15 > +> To: mp4-tech@lists.mpegif.org > +> Subject: [Mp4-tech] About map unit to slice group map. > +> > +> > +> Hi All, > +> > +> Can anyone please explain me what standard mean by " map > +> unit to slice group map". I read ITU standard many times to > +> understand this but could not get it clear. > +> Please expalin what does "map unit" represent. > +> > +> Thanks in advance. > +> > +> Reg, > +> Atul > +> > +> > +> > +> _____ > +> > +> Discover Yahoo! > +> Find restaurants, movies, travel & more fun for the weekend. > +> Check > +> +> weekend.html> it out! > +> > +> > +> _______________________________________________ > +> NOTE: Please use clear subject lines for your posts. Include > +> [audio, [video], [systems], [general] or another > +> apppropriate identifier to indicate the type of question you have. > +> > +> Note: Conduct on the mailing list is subject to the > +> Antitrust guidelines found at > +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant > +> itrust.php > +> > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to > indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > Yahoo! Sports > Rekindle the Rivalries. Sign up for Fantasy Football > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From ksuehring web.de Fri Jun 24 11:38:55 2005 From: ksuehring web.de (Karsten Suehring) Date: Fri Jun 24 04:56:18 2005 Subject: [MP4-tech] intra8x8 mode In-Reply-To: <20050623054059.6298.qmail@web30205.mail.mud.yahoo.com> References: <20050623054059.6298.qmail@web30205.mail.mud.yahoo.com> Message-ID: <42BBC69F.6060202@web.de> The Intra_8x8 MbPartPredMode mode is signaled by mb_type equal to I_NxN and transform_size_8x8_flag equal to 1 (see table 7-11) The semantics section says for transform_size_8x8_flag: "When transform_size_8x8_flag is not present in the bitstream, it shall be inferred to be equal to 0." So Intra_8x8 MbPartPredMode is not possible in baseline/main/extended. Best regards, Karsten Robin Zoo wrote: > Dear experts, > > I noticed that transform_8x8_mode_flag "shall not be > present in picture parameter sets" for > Baseline/Main/Extended profiles, which means that 8x8 > transform decoding process is not in use. I'd like to > know whether intra8x8 prediction mode is used in the > profiles or not. If yes, is the 4x4 based inverse > transform being used accordingly to get the > reconstructed picture? > > thanks in advance! > > Robin > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From JelyZhu innosis.com.cn Fri Jun 24 19:10:06 2005 From: JelyZhu innosis.com.cn (JelyZhu@innosis.com.cn) Date: Fri Jun 24 09:53:48 2005 Subject: [Mp4-tech] about horizontal motion vector component range inbaseline profile! Message-ID: Hi Gary: Thanks for your reply! I have another question about the redundant slices feature of BaseLine profile in standard and JM code. Where does the redundant slice fit into the overall picture? How do peopel use this feature? What are the related concepts? Thanks. Jely From garysull windows.microsoft.com Fri Jun 24 13:04:56 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Sun Jun 26 18:28:28 2005 Subject: [MP4-tech] padding issue Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F2E5645@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Sathyavratha (et al), The usual practice is to "pad" (i.e., copy) samples to extrapolate a border of 16 [perhaps 15 would be enough?] extra samples (8 for chroma) around the edge of the picture. Then, when decoding, the decoder clips the value of the block motion vector prior to extracting the reference samples to ensure that no sample in the reference block lie outside of the padded region. I don't recall for certain whether this decoder motion vector clipping is officially needed or not (there may be an encoder restriction on motion vector values to ensure that it is not officially needed, although a robust decoder design should probably do it anyway). The motion vector used for the prediction of subsequent motion vectors is the one BEFORE any such block-wise motion vector value clipping. You need to be especially careful about the right border and the bottom border for pictures that have a size that is not divisible by 16 in width or height. This was the subject of some recent corrigendum clarification activity. I don't recall what was the final decision about that (i.e., whether the padding should begin at the edge of the cropping rectangle or at the edge of the area that is within the decoded region but outside of the cropping rectangle). Since this aspect was not as clear in prior versions of the standard, it may be possible that different decoder implementers interpreted the standard in different ways. So encoders should be designed with caution about the way they use motion vectors over picture boundaries for picture sizes that are not divisible by 16 in height or width. It might be best to avoid using such motion vectors in an encoder with such picture sizes, in order to ensure interoperability. Hopefully the latest set of conformance bitstreams checks this aspect of decoder behavior, but I'm not sure whether it does or not. For full details, refer to the standard itself, and make sure you have the very latest version (including corrigenda). Do not assume that other standards (e.g., H.263) handle all the details in exactly the same way. Best Regards, Gary Sullivan ________________________________ From: sathyavratha [mailto:sathyavratha@cisctechnology.com] Sent: Thursday, June 23, 2023 9:54 PM To: Gary Sullivan Cc: mp4-tech@lists.mpegif.org Subject: [MP4-tech] padding issue hi experts, we are developing mpeg-4 simple profile visual decoder and encoder IP's my doubt is whether vector padding is necessary before MV decoding for simple profile to get motion compensated output. please let me know the steps for developing motion compensation. thanx in advance. sathyavratha -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050624/d622800e/attachment.html From garysull windows.microsoft.com Fri Jun 24 13:25:09 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Sun Jun 26 18:34:42 2005 Subject: [Mp4-tech] about horizontal motion vector component range inbaseline profile! Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F2E56A2@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Jely (et al), A redundant slice would ordinarily only be used if some part of the picture was lost or corrupted. The decoder is not required to use them. Simple decoders can just ignore them if that's what the decoder designer wants to do. In fact the decoder is required not to use them unless there is a problem with the bitstream. I'm not sure I understand the question "Where does the redundant slice fit into the overall picture?" A redundant slice has a macroblock address, so it fits into the picture in the location identified by its macroblock address. The idea is that if some part of the picture is lost, then the data in the redundant slice (if available) can be used as a substitute in the problem area. Best Regards, Gary Sullivan +> -----Original Message----- +> From: JelyZhu@innosis.com.cn [mailto:JelyZhu@innosis.com.cn] +> Sent: Friday, June 24, 2023 3:10 AM +> To: Gary Sullivan +> Cc: mp4-tech@lists.mpegif.org +> Subject: RE: [Mp4-tech] about horizontal motion vector +> component range inbaseline profile! +> +> +> Hi Gary: +> +> Thanks for your reply! +> +> I have another question about the redundant slices feature +> of BaseLine +> profile in standard and JM code. Where does the redundant +> slice fit into +> the overall picture? How do peopel use this feature? What +> are the related +> concepts? +> +> Thanks. +> +> +> Jely +> +> +> From fghxu hotmail.com Fri Jun 24 18:57:59 2005 From: fghxu hotmail.com (Frank Xu) Date: Sun Jun 26 18:34:48 2005 Subject: [Mp4-tech] How to encode layered MPEG4 with temporal and SNR scalability Message-ID: Hi all, I am able to encode layered MPEG4 with spatial scalability using c036089_ISO_IEC_14496-5_2001_Amd_1_2002_Reference_Software by change its sample configure files. But I don't know how to changed those files to encoded the layered MPEG4 with Temporal and/or SNR scalability. I have spent long time in this. Any one can help me on that (e.g. give me a copy of sample control files)? Thanks in advance! Frank -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050624/6ca434f2/attachment.html From dmitriy graphics.cs.msu.ru Mon Jun 27 01:30:39 2005 From: dmitriy graphics.cs.msu.ru (Dmitriy Vatolin) Date: Sun Jun 26 18:34:55 2005 Subject: [Mp4-tech] LIST OF MPEG-4 AVC/H.264 VIDEO CODECS Message-ID: <1181245950.20050627003039@graphics.cs.msu.ru> Hello all! Now we prepare information for second h.264 codecs comparison test (first available from: http://compression.ru/video/codec_comparison/mpeg-4_avc_h264_en.html ), that will be compared with more advanced metrics (download program from http://compression.ru/video/quality_measure/video_measurement_tool_en.html ). So we collect information about current available codecs and there stage. If you have ANY information about this codecs or another codecs please replay this message! Updated codecs list available from http://www.compression-links.info/MPEG-4_AVC_H264 ============================================================================ LIST OF MPEG-4 AVC/H.264 VIDEO CODECS ============================================================================ SOFTWARE CODECS: ---------------- 1. Ateme H.264/MPEG-4 AVC Codec (Ateme) No download on original site: http://www.ateme.com/products/h264.php See Nero Recode from ftp://ftp.nero.com/ --- 2. VSS H.264 Video Codec (Vanguard Software Solutions, Inc.) Trial available, VfW. http://www.vsofts.com/h264/codecs.html --- 3. Mainconcept H.264 (MainConcept AG) Trial available http://www.mainconcept.com/h264_encoder.shtml See download page and SDK page Mainconcept declare partnership with Elecard in H.264 in March 2005. --- 4. Elecard OneClick Compressor & Elecard AVC/H.264 Decoder Package (Elecard) 21-day free trial. Have H.264 SDK http://www.elecard.com/products/oneclick.shtml --- 5. Moonlight H.264 Video Codec (Moonlight Cordless LTD.) 21-day evaluation, DirectX filter. http://www.moonlight.co.il/products...er/oneclickavc/ This codec was developed with Elecard team. Now Elecard states they have fully another codec, but they are looks quite similar with Moonlight. --- 6. QuickTime 7 H.264 (Apple Computer, Inc.) Integrated. Who test it? Any remarks welcome! http://www.apple.com/quicktime/technologies/h264/ --- 7. FastVDO H.264 (FastVDO LLC) Demo has decode time limit of 5 min. http://www.fastvdo.com/H.264.html --- 8. LEAD H.264 Video codec (LEAD Technologies) Only demo, but you can buy codec online (100$). Who test it? http://www.leadcodecs.com/Codecs/LEAD-H264.htm --- 9. SVM H.264 Decoder Kit & MKi DVD Converter (Pegasus Information Technology Inc.) H.264 encoder and decoder with non standart format. http://www.h-264.com/downloads.htm --- 10. AVC Alliance free demo player (AVC Alliance, written by Philips Electronics) Poor but free. http://www.avc-alliance.nl/main/downloads.htm --- 11. Mpegable AVC Codec (dicas digital image coding GmbH) No download, VfW http://www.mpegable.com/show/mpegableavc.html Old version from free-codecs --- 12. Expert H.264 (PixelTools Corporation) Download on request http://www.pixeltools.com/experth264.html --- 13. Softstream H.264/MPEG-4 (Media Excel) Audio & Video encoders ane decoders. Evaluation copy on request http://www.mediaexcel.com/products.htm --- 14. Fraunhofer IIS H.264 Codec (Fraunhofer IIS) No download. http://www.iis.fraunhofer.de/amm/download/mpeg4/ --- 15. ViBE MPEG-4 AVC (Thomson) No download http://www.thomsongrassvalley.com/p.../encoder_mpeg4/ --- 16. UBLive-264-C64 (UB Video Incorporated) Demo available on request. http://www.ubvideo.com/mainmenu.html --- 17. Sorenson Squeeze 4 Compression Suite (Sorenson) No download. Converter. http://www.sorensonmedia.com/soluti...od/mpg4_win.php --- 18. Sonic?s HD-Series AVC encoder (Sonic Solutions) No download. Only press release now. http://www.sonic.com/about/press/ne...4/hdseries.aspx --- 19. Enchansed sklmp4 (Pascal Massimino) No download. Announce of 264 development. http://skal.planet-d.net/coding/mpeg4codec.html ---------------------------------------------------------- OPEN SOURCE: --------------------- 20. x264 (free, supported by VideoLAN) OPEN SOURCE. Good publically available high profile encoder. Developers declare "early development stage", but encode results are one of the best. http://www.videolan.org/x264.html http://trac.videolan.org/x264/browser/trunk/ http://x264.nl/ --- 21. JM 9.6 reference H.264 encodec (free, supported by Joint Video Team ISO+ITU-T) OPEN SOURCE. Can be used only for stream testing. http://bs.hhi.de/~suehring/tml/ --- 22. Hdot264 Pre-Alpha (Open Source Project Hdot264) OPEN SOURCE. http://sourceforge.net/projects/hdot264/ --- 23. Minimal H.264 decoder (Einige Fakten) OPEN SOURCE. Basic H.264 decoder. http://www-user.tu-chemnitz.de/~mfie/stuff/SA/ --- See also Software Developments Kits for software codec usage: 1. Elecard AVC&MPEG2 Encoder SDK (Elecard) MPEG2 and H.264 Video Encoder and Decoders other components within the Microsoft? DirectShow? technology. http://www.elecard.com/products/encoder_sdk.shtml http://www.elecard.com/products/decoder_sdk.shtml 2. MainConcept H.264 SDK (MainConcept AG) Windows, Apple, Linux and TI DSP supported. http://developer.mainconcept.com/h264-sdk.html 3. LEADTOOLS Multimedia Developer SDKs (LEAD Technologies, Inc) include support for AVI, WMV, WMA, ASF, MPEG, MPEG-1, MPEG-2, MPG, WAV, OGG, MIDI, QuickTime, SND, AIF files, video and audio processing filters (rotate, resize, deinterlace, etc). http://www.leadtools.com/SDK/Multim...-Products-n.htm 4. VSS H.264 SDK (Vanguard Software Solutions, Inc.) Based on there codec. Windows and Linux support. http://www.vsofts.com/h264/sdk.html 5. Pegasus H.264 SDK Development Kits (Pegasus) Looks like simple codec, but prefilters declared. http://www.h-264.com/Products.htm ---------------------------------------------------------- HARDWARE IMPLEMENTED: --------------------- 24. Sony PlayStation Portable (Sony Computer Entertainment Inc.) Product: Support H.264/MPEG-4 AVC Main Profile Level3 decoding. http://www.us.playstation.com/consoles.aspx?id=4 --- 25. VSX 300 Videoconferencing Terminal with H.264 support (Polycom Inc) Product: Biggest HW videoconferencing company, prepare many solutions with 264 support http://www.polycom.com/products_ser...35-6197,00.html --- 26. Tandberg EN5930 (Tandberg Television) Product for videoconferencing. Who test it? I have bad review. http://www.tandbergtv.com/productview.asp?n=65 --- 27. 4Caster MPEG-4 / H.264 Encoder (Envivio, Inc.) Product: Hardware box encoder http://www.envivio.com/products/4caster.html http://www.envivio.com/products/4caster_se.html --- 28. Tut Systems MPEG-4 Advanced Video Codec (Tut Systems, Inc.) Product: Hardware transcoder box. http://www.tutsys.com/pdflibrary/pd..._AVC_2_9_05.pdf --- 29. Hardware MGW 5100 (Optibase, Inc.) Product: Hardware box encoder http://www.optibase.com/Content.aspx?id=38 --- 30. Harmonic DiviCom? MV 100 Encoder (Harmonic Inc. ) Product: Hardware box encoder http://www.harmonicinc.com/view_csd...fm?classID=1010 --- 31. Broadcom chips (Broadcom Corp. former Sandvideo) Chip with 264 HD decoding http://www.broadcom.com/products/Ca...r-Chips/BCM7411 --- 32. SA New Advanced Compression AVC Encoder (Scientific Atlanta) Product: Transmission system (for satellite TV) http://www.saeurope.com/products/customerhome.htm --- 33. H.264 VLE4000 (LSI Logic Corp. former: VideoLocus) Chips with 264 support http://www.lsilogic.com/products/vi...ng/vle4000.html --- 34. Modulus Video SDTV, HDTV Encoder (Modulus Video, Inc.) Chips. Declarations only now. License H.264 from LSI Logic. http://modulusvideo.com/main.php?Page=12 --- 35. BSP 15 (System-on-chip) (Equator Technologies, Inc.) Chips. No exact codec info. http://www.equator.com/solutions/digitalmedia.html --- 36. BT Exact H.264 Suite (BT Exact) temporarily not available http://www.btexact.com/technologies...files?doc=42895 --- 37. Surveillance Real Time H.264 Encoders (DSP Research, Inc.) Several chips implementation. http://www.dspr.com/www/products/overview_video.htm http://www.dspr.com/www/support/dow...eo_download.htm http://www.wwcoms.com/products/index.htm --- 38. VP3-PMC / Hardware TwinPEG Pro (Vitec Multimedia) Chips. TMS320DM642TMDSP & etc http://www.vitecmm.com/product.php?id=41 --- We plan prepare next MSU H.264 codecs comparison test soon will be available from http://compression.ru/video/codec_comparison/mpeg-4_avc_h264_en.html If you know any more information about H.264 codecs, please reply this message or add information to Compression-links.info H.264 page: http://www.compression-links.info/MPEG-4_AVC_H264 -- Best regards, Dmitriy Vatolin MSU Graphics&Media Lab From AngelZhang innosis.com.cn Mon Jun 27 10:52:51 2005 From: AngelZhang innosis.com.cn (AngelZhang@innosis.com.cn) Date: Mon Jun 27 03:21:02 2005 Subject: [Mp4-tech] H.264: about random test sequence! Message-ID: hi experts, I want to list some test sequences for baseline decoder function verification. and i know one means is to change parameters of encoder.cfg to make diffrent h264 files. but i don't know whether there are still other good means? besides, i don't know how to realize random test case. please give me some hints! thanks in advance! angel From JelyZhu innosis.com.cn Mon Jun 27 13:02:48 2005 From: JelyZhu innosis.com.cn (JelyZhu@innosis.com.cn) Date: Mon Jun 27 03:27:35 2005 Subject: [Mp4-tech] about horizontal motion vector component range inbaseline profile! Message-ID: Hi Gary: Thank you for your reply. I still have some more questions. It seems there is no operation for redundant slices such as discarding the coded slice in JM96, so I am not sure if I really need to care about it in Baseline Profile decoder. In encoder configure file, there is a sentence like this: UseRedundantSlice = 0; # 0 : not used 1: one redundant slice used for each slice ( other modes not supported yet ) My interpretation is that if I set this syntax element to 1 in encoding process, then I will receive the redundant slices after the primary coded picture in decoding process, and if these slices of this picture have some errors or losses , these redundant slices will be used to overwrite the corresponding primary slices, and if not, they will be discarded. In fact, there is not any redundant slices ( redundant_pic_cnt is always equal to 0 ) in bit stream. Even though redundant_pic_cnt is set to 1 intentionally in parsing process, the decoder still doesnĄ¯t care about this. Thanks in advance. Jely From sandipkr dgbmicro.com Mon Jun 27 12:01:45 2005 From: sandipkr dgbmicro.com (Sandip Ray) Date: Mon Jun 27 03:31:13 2005 Subject: [Mp4-tech] Parallel Processing of Y, U and V?? Message-ID: Hi Jing, It is not clear from your email that which video coding standard are you talking about. In general, most of the video coding algorithms give emphasis on luma(Y) processing as it is more important for human visual system. If you are interested to have better understanding of the importance Y, U and V, I can say you should go through the publications available in http://www.ece.rochester.edu/~gsharma/ In general, I can say Y contains more visual information than U and V. In most of the existing video coding standard(e.g. MPEG-4, H.264 etc.), Y processing is more compared to U and V. Both of U and V processing are usually same. If you are talking about MPEG-4, then yes, you can do parallel processsing of IDCT and inter prediction. The pros and cons depend on the platform where you want to run the codec. As far as the standard is concerned, you have to be careful about maintaining enough precision for the residual data(output of IDCT). You can look into the MPEG-4 standard for recommended precision. I don't think there can be any other issue. Regards, Sandip -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050627/1dc5becd/attachment.html From Ye-Kui.Wang nokia.com Mon Jun 27 12:01:00 2005 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Mon Jun 27 04:09:28 2005 Subject: [Mp4-tech] redundant slices Message-ID: <1C1F3D15859526459B4DD0A7A9B2268BA38B78@trebe101.NOE.Nokia.com> Hi Jely, Let me try to answer your questions below. What Gary said is basically the intended use of redundant slices, though you may also use it in other ways. Yes the feature has not been adequately implemented in the reference software. However, that is not really relevant to whether you need to care about redundant slices in your baseline decoder, which heavily depends on your application requirements. BR, YK > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] > Sent: 27 June, 2005 07:03 > To: Gary Sullivan > Cc: mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] about horizontal motion vector component > rangeinbaseline profile! > > > > > > Hi Gary: > > Thank you for your reply. > > I still have some more questions. It seems there is no > operation for > redundant slices such as discarding the coded slice in > JM96, so I am not > sure if I really need to care about it in Baseline Profile > decoder. In > encoder configure file, there is a sentence like this: > > UseRedundantSlice = 0; # 0 : not used 1: one > redundant slice used > for each slice ( other modes not supported yet ) > > My interpretation is that if I set this syntax element to > 1 in encoding > process, then I will receive the redundant slices after the primary > coded picture in decoding process, and if these slices of > this picture > have some errors or losses , these redundant slices will be used to > overwrite the corresponding primary slices, and if not, > they will be > discarded. > > In fact, there is not any redundant slices ( > redundant_pic_cnt is always > equal to 0 ) in bit stream. Even though redundant_pic_cnt > is set to 1 > intentionally in parsing process, the decoder still doesnĄ¯ > t care about > this. > > Thanks in advance. > > Jely > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From sagar sarayusoftech.com Mon Jun 27 14:05:15 2005 From: sagar sarayusoftech.com (sagar) Date: Mon Jun 27 04:09:36 2005 Subject: [Mp4-tech] mpeg4 porting issues In-Reply-To: Message-ID: Hi Experts, I am working on MPEG4 on Blackfin BF533 processor. Can anybody suggest me some way to prepare a test batch file or some means to test for number of other test vectors. Regards, Sagar -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of AngelZhang@innosis.com.cn Sent: Monday, June 27, 2023 7:23 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] H.264: about random test sequence! hi experts, I want to list some test sequences for baseline decoder function verification. and i know one means is to change parameters of encoder.cfg to make diffrent h264 files. but i don't know whether there are still other good means? besides, i don't know how to realize random test case. please give me some hints! thanks in advance! angel _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From satish_babudv yahoo.com Mon Jun 27 04:46:15 2005 From: satish_babudv yahoo.com (satish babu) Date: Mon Jun 27 08:17:14 2005 Subject: [Mp4-tech] H.264 video baseline decoder supports USB video class driver or not? Message-ID: <20050627104615.24253.qmail@web50306.mail.yahoo.com> hi experts, can any one tell whether H.264 baseline decoder supports USB video class driver waiting 4 ur reply Regards, satish --- mp4-tech-bounces@lists.mpegif.org wrote: > Send Mp4-tech mailing list submissions to > mp4-tech@lists.mpegif.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.mpegif.org/mailman/listinfo/mp4-tech > or, via email, send a message with subject or body 'help' to > mp4-tech-request@lists.mpegif.org > > You can reach the person managing the list at > mp4-tech-owner@lists.mpegif.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Mp4-tech digest..." > > > Today's Topics: > > 1. RE: padding issue (Gary Sullivan) > 2. RE: about horizontal motion vector component range inbaseline > profile! (Gary Sullivan) > 3. How to encode layered MPEG4 with temporal and SNR scalability > (Frank Xu) > 4. LIST OF MPEG-4 AVC/H.264 VIDEO CODECS (Dmitriy Vatolin) > 5. H.264: about random test sequence! (AngelZhang@innosis.com.cn) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Fri, 24 Jun 2023 12:04:56 -0700 > From: "Gary Sullivan" > Subject: RE: [MP4-tech] padding issue > To: "sathyavratha" > Cc: Jens-Rainer Ohm , tchiang@cc.nctu.edu.tw, > mp4-tech@lists.mpegif.org > Message-ID: > <91D7F2CEE3425A4A9D11311D09FCE2460F2E5645@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> > > Content-Type: text/plain; charset="us-ascii" > > Sathyavratha (et al), > > The usual practice is to "pad" (i.e., copy) samples to extrapolate a > border of 16 [perhaps 15 would be enough?] extra samples (8 for chroma) > around the edge of the picture. > Then, when decoding, the decoder clips the value of the block motion > vector prior to extracting the reference samples to ensure that no > sample in the reference block lie outside of the padded region. > I don't recall for certain whether this decoder motion vector clipping > is officially needed or not (there may be an encoder restriction on > motion vector values to ensure that it is not officially needed, > although a robust decoder design should probably do it anyway). > > The motion vector used for the prediction of subsequent motion vectors > is the one BEFORE any such block-wise motion vector value clipping. > > You need to be especially careful about the right border and the bottom > border for pictures that have a size that is not divisible by 16 in > width or height. This was the subject of some recent corrigendum > clarification activity. I don't recall what was the final decision > about that (i.e., whether the padding should begin at the edge of the > cropping rectangle or at the edge of the area that is within the decoded > region but outside of the cropping rectangle). Since this aspect was > not as clear in prior versions of the standard, it may be possible that > different decoder implementers interpreted the standard in different > ways. > > So encoders should be designed with caution about the way they use > motion vectors over picture boundaries for picture sizes that are not > divisible by 16 in height or width. It might be best to avoid using > such motion vectors in an encoder with such picture sizes, in order to > ensure interoperability. > > Hopefully the latest set of conformance bitstreams checks this aspect of > decoder behavior, but I'm not sure whether it does or not. > > For full details, refer to the standard itself, and make sure you have > the very latest version (including corrigenda). Do not assume that > other standards (e.g., H.263) handle all the details in exactly the same > way. === Message Truncated === __________________________________ Do you Yahoo!? Read only the mail you want - Yahoo! Mail SpamGuard. http://promotions.yahoo.com/new_mail From v_v_bhat yahoo.com Mon Jun 27 04:55:27 2005 From: v_v_bhat yahoo.com (vinayak bhat) Date: Mon Jun 27 08:24:33 2005 Subject: [Mp4-tech] Regarding H.263+ encoder(Annex K) Message-ID: <20050627105527.59635.qmail@web33005.mail.mud.yahoo.com> Hi all In case Annex K slice structure mode what is maximum packet size we can give ? For H.263 Encoder. thanks& regards vinayak --------------------------------- Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050627/e628b653/attachment.html From satish_babudv yahoo.com Mon Jun 27 05:06:21 2005 From: satish_babudv yahoo.com (satish babu) Date: Mon Jun 27 08:30:25 2005 Subject: [Mp4-tech] H.264 video baseline decoder supports USB video class driver or not? Message-ID: <20050627110621.60151.qmail@web50310.mail.yahoo.com> hi experts, can any one tell me whether H.264 video baseline decoder supports USB video class driver or not? Regards, satish. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From indrajit.chakrabarty gmail.com Mon Jun 27 12:19:56 2005 From: indrajit.chakrabarty gmail.com (Indrajit Chakrabarty) Date: Tue Jun 28 08:30:23 2005 Subject: [Mp4-tech] H.264 video baseline decoder supports USB video class driver or not? In-Reply-To: <20050627110621.60151.qmail@web50310.mail.yahoo.com> References: <20050627110621.60151.qmail@web50310.mail.yahoo.com> Message-ID: <6feb73a4050627111929e4b741@mail.gmail.com> What do you mean by a "USB video class driver"? Be more specific... On 6/27/05, satish babu wrote: > > hi experts, > > can any one tell me whether H.264 video baseline > decoder supports USB video class driver or not? > > > Regards, > satish. > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050627/f37e49f0/attachment.html From garysull windows.microsoft.com Mon Jun 27 13:21:57 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Tue Jun 28 08:30:31 2005 Subject: [Mp4-tech] about horizontal motion vector component range inbaseline profile! Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F3746C1@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Jely (et al), As I said in my previous message, decoders can ignore redundant slices. I cannot fully understand what you are saying about the reference software decoder and encoder. If the reference software decoder does not operate properly in response to redundant slices in a bitstream, that is a bug and we would like to know about it. Please note that the definition of "operate properly" may be as simple as ignoring any redundant slices found in the bitstream. Decoders are also allowed to respond to redundant slices in a more sophisticated way, but the only requirement is that they produce the correct output when decoding a conforming bitstream (and redundant slices are mostly about what decoders might want to do with non-conforming bitstreams). Whether the encoder reference software produces redundant slices or not is irrelevant. Decoders must me made to operate properly with all conforming encoders, not just the reference software encoder. The lack of support for a feature in the encoder reference software does not mean the feature is not in the standard - the standard is primarily about decoders rather than encoders. Unless an encoder produces a bitstream that does not conform to the constraints specified in the standard, it is considered to be operating properly as far as the standard is concerned. Please see what I said before, which I have quoted below. I feel like both of us are repeating ourselves, so I fear that there may not be a full understanding of each other's messages. Best Regards, Gary Sullivan +> -----Original Message----- +> From: JelyZhu@innosis.com.cn [mailto:JelyZhu@innosis.com.cn] +> Sent: Sunday, June 26, 2023 9:03 PM +> To: Gary Sullivan +> Cc: mp4-tech@lists.mpegif.org +> Subject: RE: [Mp4-tech] about horizontal motion vector +> component range inbaseline profile! +> +> +> +> +> Hi Gary: +> +> Thank you for your reply. +> +> I still have some more questions. It seems there is no +> operation for +> redundant slices such as discarding the coded slice in +> JM96, so I am not +> sure if I really need to care about it in Baseline +> Profile decoder. In +> encoder configure file, there is a sentence like this: +> +> UseRedundantSlice = 0; # 0 : not used 1: one +> redundant slice used +> for each slice ( other modes not supported yet ) +> +> My interpretation is that if I set this syntax element to +> 1 in encoding +> process, then I will receive the redundant slices after +> the primary +> coded picture in decoding process, and if these slices of +> this picture +> have some errors or losses , these redundant slices will +> be used to +> overwrite the corresponding primary slices, and if not, +> they will be +> discarded. +> +> In fact, there is not any redundant slices ( +> redundant_pic_cnt is always +> equal to 0 ) in bit stream. Even though redundant_pic_cnt +> is set to 1 +> intentionally in parsing process, the decoder still +> doesn't care about +> this. +> +> Thanks in advance. +> +> Jely +> +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gary Sullivan +> Sent: Friday, June 24, 2023 12:25 PM +> To: JelyZhu@innosis.com.cn +> Cc: mp4-tech@lists.mpegif.org +> Subject: RE: [Mp4-tech] about horizontal motion vector +> component rangeinbaseline profile! +> +> Jely (et al), +> +> A redundant slice would ordinarily only be used if some part of the +> picture was lost or corrupted. The decoder is not required +> to use them. +> Simple decoders can just ignore them if that's what the +> decoder designer +> wants to do. +> +> In fact the decoder is required not to use them unless there is a +> problem with the bitstream. +> +> I'm not sure I understand the question "Where does the +> redundant slice +> fit into the overall picture?" A redundant slice has a macroblock +> address, so it fits into the picture in the location +> identified by its +> macroblock address. +> +> The idea is that if some part of the picture is lost, then +> the data in +> the redundant slice (if available) can be used as a substitute in the +> problem area. +> +> Best Regards, +> +> Gary Sullivan +> +> +> -----Original Message----- +> +> From: JelyZhu@innosis.com.cn [mailto:JelyZhu@innosis.com.cn] +> +> Sent: Friday, June 24, 2023 3:10 AM +> +> To: Gary Sullivan +> +> Cc: mp4-tech@lists.mpegif.org +> +> Subject: RE: [Mp4-tech] about horizontal motion vector +> +> component range inbaseline profile! +> +> +> +> +> +> Hi Gary: +> +> +> +> Thanks for your reply! +> +> +> +> I have another question about the redundant slices feature +> +> of BaseLine +> +> profile in standard and JM code. Where does the redundant +> +> slice fit into +> +> the overall picture? How do peopel use this feature? What +> +> are the related +> +> concepts? +> +> +> +> Thanks. +> +> +> +> +> +> Jely +> +> +> +> +> +> +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant itrust.php +> From garysull windows.microsoft.com Mon Jun 27 14:10:48 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Tue Jun 28 08:30:38 2005 Subject: [Mp4-tech] Regarding H.263+ encoder(Annex K) Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F3747A5@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Vinayak (et al), I don't think the standard specifies a maximum video packet size. (Same answer regardless whether the question is about H.263 or MPEG-4 part 2 or part 10.) If you're talking about how to use some particular software encoder package, then probably we can't give you the answer. Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of vinayak bhat Sent: Monday, June 27, 2023 3:55 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Regarding H.263+ encoder(Annex K) Hi all In case Annex K slice structure mode what is maximum packet size we can give ? For H.263 Encoder. thanks& regards vinayak ________________________________ Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050627/b2a1a9a0/attachment.html From AngelZhang innosis.com.cn Tue Jun 28 10:46:33 2005 From: AngelZhang innosis.com.cn (AngelZhang@innosis.com.cn) Date: Tue Jun 28 08:30:44 2005 Subject: [Mp4-tech] H.264: about the caculation of mvCLX Message-ID: hi experts, In the part 8.4.1.4 of H264 standard,the description of mvCLX's derivation is written as following: "if the current macroblock is a frame macroblock,the horizontal and vertical components of the chroma motion vector mvCLX are derived by multiplying the corresponding components of luma motion vector mvLX by 2, through mapping one-quarter sample mvLX units to one-eighth sample mvCLX units" but the formular : mvCLX[0]=mvLX[0] mvCLX[1]=mvLX[1] havn't implement what the standard has said to multiply 2. i am confused about it very much.please give me your idea. thanks angel From g.lakshmankumar gmail.com Tue Jun 28 11:17:43 2005 From: g.lakshmankumar gmail.com (Lakshman Kumar) Date: Tue Jun 28 08:30:51 2005 Subject: [Mp4-tech] SNR calculation Message-ID: <138aa52f05062721475046e54f@mail.gmail.com> Hi all Can i know how SNR ratio is calculated for YUV component in h.264 baseline profile thx in advance -- lakshman kumar -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050628/9744b71e/attachment.html From iamsangu gmail.com Tue Jun 28 17:45:48 2005 From: iamsangu gmail.com (lisa sangs) Date: Tue Jun 28 08:30:58 2005 Subject: [Mp4-tech] MPEG 2.5 Message-ID: <42C13164.9000603@gmail.com> hi I want to know from where can i download *MPEG 2.5 documents* for its implementation.please give me the links. I need *testvectors* for MPEG 2.5 also. thanks in advance, lisa sangs From A.Thomson indigovision.com Tue Jun 28 16:29:56 2005 From: A.Thomson indigovision.com (Alan Thomson) Date: Tue Jun 28 14:14:30 2005 Subject: [Mp4-tech] H.264: about the caculation of mvCLX Message-ID: <67336BF4D5D23C4AA30EC72DFCFCA6DF012BFE0E@peebles.indigovision.com> Hi - The equations are correct; the text is misleading. Delete "multiplying", insert "dividing" and it makes more sense. For example a luma component of 3 (quarter-luma-samples) yields a chroma component of 3 (eighth-chroma-samples). BR, Alan > -----Original Message----- > From: AngelZhang@innosis.com.cn [mailto:AngelZhang@innosis.com.cn] > Sent: 28 June 2023 02:47 > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] H.264: about the caculation of mvCLX > > > hi experts, > In the part 8.4.1.4 of H264 standard,the description of mvCLX's > derivation is written as following: > "if the current macroblock is a frame macroblock,the horizontal and > vertical components of the chroma motion vector mvCLX are derived by > multiplying the corresponding components of luma motion > vector mvLX by 2, > through mapping one-quarter sample mvLX units to one-eighth > sample mvCLX > units" > but the formular : > mvCLX[0]=mvLX[0] > mvCLX[1]=mvLX[1] > havn't implement what the standard has said to multiply 2. > > i am confused about it very much.please give me your idea. > > thanks > angel > > > > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php ---------------------------------------------------------------------------- -------- scanned for viruses and spam by indigovision If you believe this email is spam, please forward to spam@emailfiltering.com From tomar_atul2000 yahoo.com Tue Jun 28 09:05:25 2005 From: tomar_atul2000 yahoo.com (atul tomar) Date: Tue Jun 28 14:21:13 2005 Subject: [Mp4-tech] about InverseRasterScan.. Message-ID: <20050628150525.429.qmail@web52310.mail.yahoo.com> Hi All, In ITU standard, 6.4 index describes about Inverse scanning processes and derivation processes for neighbors. I could not understand InverseRasterScan( , , , , 0) and InverseRasterScan( , , , , 1) functions to determine x, y locations of upper left sample for a block. Also I could not get any explanation about these two functions. If would appreciate if someone could give me its details. Thanks in advance. Regards, Atul --------------------------------- Yahoo! Mail Stay connected, organized, and protected. Take the tour -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050628/1630c60b/attachment.html From garysull windows.microsoft.com Tue Jun 28 12:33:18 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Tue Jun 28 15:52:45 2005 Subject: [Mp4-tech] H.264: about the caculation of mvCLX Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F375304@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> You are looking at an old version of the standard. Get a new copy. You can get it for free from the ITU by following the directions on the MPEG-IF web site. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Alan Thomson +> Sent: Tuesday, June 28, 2023 7:30 AM +> To: AngelZhang@innosis.com.cn; mp4-tech@lists.mpegif.org +> Subject: RE: [Mp4-tech] H.264: about the caculation of mvCLX +> +> Hi - +> +> The equations are correct; the text is misleading. +> Delete "multiplying", insert "dividing" and it makes more sense. +> For example a luma component of 3 (quarter-luma-samples) +> yields a chroma component of 3 (eighth-chroma-samples). +> +> BR, +> Alan +> +> > -----Original Message----- +> > From: AngelZhang@innosis.com.cn [mailto:AngelZhang@innosis.com.cn] +> > Sent: 28 June 2023 02:47 +> > To: mp4-tech@lists.mpegif.org +> > Subject: [Mp4-tech] H.264: about the caculation of mvCLX +> > +> > +> > hi experts, +> > In the part 8.4.1.4 of H264 standard,the description of mvCLX's +> > derivation is written as following: +> > "if the current macroblock is a frame macroblock,the +> horizontal and +> > vertical components of the chroma motion vector mvCLX are +> derived by +> > multiplying the corresponding components of luma motion +> > vector mvLX by 2, +> > through mapping one-quarter sample mvLX units to one-eighth +> > sample mvCLX +> > units" +> > but the formular : +> > mvCLX[0]=mvLX[0] +> > mvCLX[1]=mvLX[1] +> > havn't implement what the standard has said to multiply 2. +> > +> > i am confused about it very much.please give me your idea. +> > +> > thanks +> > angel +> > +> > +> > +> > +> > +> > _______________________________________________ +> > NOTE: Please use clear subject lines for your posts. Include +> > [audio, [video], [systems], [general] or another apppropriate +> > identifier to indicate the type of question you have. +> > +> > Note: Conduct on the mailing list is subject to the Antitrust +> > guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant itrust.php +> +> ------------------------------------------------------------- +> --------------- +> -------- +> scanned for viruses and spam by indigovision +> If you believe this email is spam, please forward to +> spam@emailfiltering.com +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant itrust.php +> From dharjanto gmail.com Tue Jun 28 12:56:45 2005 From: dharjanto gmail.com (dharjanto) Date: Tue Jun 28 15:59:14 2005 Subject: [Mp4-tech] Capture MP4/3GP video snapshot Message-ID: <403345ee0506281156b7c4b0a@mail.gmail.com> Is there any available source code (or .dll) to extract MP4/3GP video frame (just 1st frame) and save it to JPG format? I need to make a thumbnail view out of MP4/3GP video files. Thanks, d?n From loloo 21cn.com Wed Jun 29 10:13:02 2005 From: loloo 21cn.com (loloo) Date: Wed Jun 29 06:08:02 2005 Subject: [Mp4-tech] Re: H.264: about the caculation of mvCLX Message-ID: <003401c57c47$b635ab60$251a60ac@EAT> As the standard said--"through mapping one-quarter sample mvLX units to one-eighth sample mvCLX units",and also we know that the chroma MV is half of the Luma MV in length, so: mvCLX'[0]=mvLX[0]/2 in 1/4 pixel unit of luma mvCLX[0]=mvCLX'[0]*2 convert to 1/8 pixel unit in chroma so we have: mvCLX[0]=mvCLX'[0]*2=mvLX[0]/2*2=mvLX[0]; and same is mvCLX[1] RuiLu >>hi experts, >> In the part 8.4.1.4 of H264 standard,the description of mvCLX's >>derivation is written as following: >> "if the current macroblock is a frame macroblock,the horizontal and >>vertical components of the chroma motion vector mvCLX are derived by >>multiplying the corresponding components of luma motion vector mvLX by 2, >>through mapping one-quarter sample mvLX units to one-eighth sample mvCLX >>units" >> but the formular : >> mvCLX[0]=mvLX[0] >> mvCLX[1]=mvLX[1] >> havn't implement what the standard has said to multiply 2. >> i am confused about it very much.please give me your idea. >>thanks >>angel From ranzhao arcsoft.com.cn Wed Jun 29 10:33:44 2005 From: ranzhao arcsoft.com.cn (=?gb2312?B?U0gtWmhhbyBSYW5b1dTIu10=?=) Date: Wed Jun 29 06:14:58 2005 Subject: [Mp4-tech] H.264: about the caculation of mvCLX Message-ID: <20F0DCA3A6EA8949B96BC2C04BFE92CC0F6614@hz-email01.apac.arcsoft.corp> Hi, angel, If you review this paragraph more carefully, you can take note of the difference between mvLX and mvCLX. The units of former is 1/4 pix and the latter is 1/8, so the formula imply the double. Best Regards, Zhao Ran In the part 8.4.1.4 of H264 standard,the description of mvCLX's derivation is written as following: "if the current macroblock is a frame macroblock,the horizontal and vertical components of the chroma motion vector mvCLX are derived by multiplying the corresponding components of luma motion vector mvLX by 2, through mapping one-quarter sample mvLX units to one-eighth sample mvCLX units" but the formular : mvCLX[0]=mvLX[0] mvCLX[1]=mvLX[1] havn't implement what the standard has said to multiply 2. i am confused about it very much.please give me your idea. thanks angel -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050629/d649e50e/attachment.html From JelyZhu innosis.com.cn Wed Jun 29 15:27:48 2005 From: JelyZhu innosis.com.cn (JelyZhu@innosis.com.cn) Date: Wed Jun 29 06:21:52 2005 Subject: [Mp4-tech] about error concealment in baseline decoder Message-ID: Hi, Gary: I donĄ¯t know if the following interpretation is correct. It is unnecessary to care about the redundant slice in slice layer of baseline decoder which includes MB header and MB data parsing. Even if some error appears in the parsing process and the whole slice cannot be parsed correctly any more, we can use p_skip type MB to substitute all MB of this slice to rebuild it. You have said if there is a problem with the bitstream, redundant slice will be used to substitute the problem area. I donĄ¯t know if this kind of problem is NAL layer problem, that is, the slice has error from the start of the NAL. I really confused about this. I donĄ¯t know how to deal with the parsing error or slice lost in decoder. Best wishes, Jely From g.lakshmankumar gmail.com Wed Jun 29 18:08:34 2005 From: g.lakshmankumar gmail.com (Lakshman Kumar) Date: Wed Jun 29 09:55:48 2005 Subject: [Mp4-tech] MPS and LPS ,SNR calculation In-Reply-To: <1C1F3D15859526459B4DD0A7A9B2268BA38B78@trebe101.NOE.Nokia.com> References: <1C1F3D15859526459B4DD0A7A9B2268BA38B78@trebe101.NOE.Nokia.com> Message-ID: <138aa52f05062904387f9281aa@mail.gmail.com> Dear all I could able to understand the fixing of range parameters during decoding , in the function biari_decode_symbol() of H.264 decoder.how it is done for MPS and LPS also how SNR is calculated for YUV Components plz do help me. with reg lakshman -- lakshman kumar -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050629/a077b87c/attachment.html From may_ank77 yahoo.com Wed Jun 29 10:22:20 2005 From: may_ank77 yahoo.com (mayank agarwal) Date: Thu Jun 30 14:29:14 2005 Subject: [Mp4-tech] test vector for MPEG-4 Message-ID: <20050629162221.78534.qmail@web52807.mail.yahoo.com> Hi all, >From where can i get the test vectors for testing MPEG-4.Secondly how should i be able to test the MPEG=4 encoder and decoder using those test vectors.For testing if MPEG-4 APIs are sufficient or if the full code is needed.Further how o determine in which functions i should put test vector as input where i should see the output. what all is the information needed if i want to test the MPEG-4 using test vectors and compare the results with those standardly available. Please give the above information as soon as possible. Regards, Mayank ____________________________________________________ Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football http://football.fantasysports.yahoo.com From garysull windows.microsoft.com Wed Jun 29 19:18:35 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Thu Jun 30 17:14:26 2005 Subject: [Mp4-tech] RE: about error concealment in baseline decoder Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F43FE89@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Reply interspersed below. Best Regards, Gary Sullivan +> -----Original Message----- +> From: JelyZhu@innosis.com.cn [mailto:JelyZhu@innosis.com.cn] +> Sent: Tuesday, June 28, 2023 11:28 PM +> To: Gary Sullivan +> Cc: mp4-tech@lists.mpegif.org +> Subject: about error concealment in baseline decoder +> +> +> Hi, Gary: +> +> I don't know if the following interpretation is correct. +> +> It is unnecessary to care about the redundant slice in +> slice layer of +> baseline decoder which includes MB header and MB data +> parsing. Even if +> some error appears in the parsing process and the whole +> slice cannot be +> parsed correctly any more, we can use p_skip type MB to +> substitute all +> MB of this slice to rebuild it. Yes, that's fine. Any redundant slices that are present can be ignored by the decoder. +> +> You have said if there is a problem with the bitstream, +> redundant slice +> will be used to substitute the problem area. I don't know +> if this kind +> of problem is NAL layer problem, that is, the slice has +> error from the +> start of the NAL. The redundant slice CAN be used if there is a problem with ordinary slices the bitstream (e.g., if some of them are lost during transmission). It is not necessary that it WILL be used. I'm not sure I understand the rest of your above remark. +> +> I really confused about this. I don't know how to deal +> with the parsing +> error or slice lost in decoder. +> +> Best wishes, +> Jely +> From dyipan gmail.com Thu Jun 30 12:17:17 2005 From: dyipan gmail.com (yipan) Date: Thu Jun 30 17:21:30 2005 Subject: [Mp4-tech] Strange AAC Packet header, could sb. gives me a hint? Thanks. Message-ID: <30c5945a05062920174081666d@mail.gmail.com> Hi everybody, Recently I am trying to decode an .mp4 file. I have already successfully make the video stream playingsmoothly, but when I am dealing with the AAC audio stream reside inthe MP4 file, I find something really strange which makes me confused. I find the 'mp4a' tag in the 'stsd' box, from this information Ican know the audio stream in my testing mp4 file is encoded in AAC. Then I turn to AAC for some detail information. I got to knowthere exists two commonly used header in AAC stream, ADTS and ADIF.The former is existed at the beginning of each AAC audio frame withfirst 12 bits set; The latter is only existed at the beginning of thewhole stream with the identification "0x41444946". With my experience in AVI, I suggest the AAC audio stream withADTS format in mp4 file would be the better choice. In such case, Imay find '0xFFF', namely the sync word in each audio packet. But whenI view the testing mp4 file in binary, I find out each stream startswith '0x21', and I can't find the sync word in each packet. Following is the first three audio frames in my testing .mp4 file: Frame 1:(6 bytes) 21 10 03 20 64 1C Frame 2:(146 bytes) 21 1A 53 FD ?? Frame 3:(161 bytes) 21 1A 54 B5 ?? Can anybody give me a hint? Thanks! Best Regards, yipan From guohuaxu gmail.com Thu Jun 30 02:46:13 2005 From: guohuaxu gmail.com (Frank Xu) Date: Thu Jun 30 17:26:58 2005 Subject: [Mp4-tech] which Mpeg-4 reference software are you guys currently using that can do scalable encode/decode? In-Reply-To: References: Message-ID: <3ebb504105062922461e9433e9@mail.gmail.com> Hi all, I am new to the MPEG4 world, I need to do the SNP ( or temporal) scalable encoding/decoding MPEG4. I found MoMusys is the one that can do this, However. all questions about MoMuSys in this forum appear 2~3 years ago. So I am assuming there is a newer free avaliable reference sotware? Could any one point it out to me where to find it. Also, does the new software capable of doing SNP or temporal scalable encoding/decoding? I have spent weeks in searching and testing difference versions of MoMuSys and MS version, I can't make them do the SNP scalable encoding/decoding. I am so frustrated. Please help me on this Best regards Frank From JelyZhu innosis.com.cn Thu Jun 30 17:49:34 2005 From: JelyZhu innosis.com.cn (JelyZhu@innosis.com.cn) Date: Thu Jun 30 17:30:55 2005 Subject: [Mp4-tech] some questions about H.264 standard! Message-ID: Hi, Gary: Thank you very much for your kindly reply! I have some other questions about H.264. 1. As we all know, there are three kinds of samples in interpolation: quarter-pel samples, half-pel samples and integer-pel samples. They have different calculation complexity. My interpretation is that the quarter-pel interpolation will be used in the detailed area and on the contrary, the integer-pel interpolation is used in homogeneous areas which have more probability to appear than the detailed area. So can I get the statistic of the number of three kinds of samples, respectively, to evaluate the performance of the interpolation process and if this evaluation is reasonable or not? 2. The sps and pps can be activated again once IDR access unit appears (I donĄ¯t know my interpretation is correct or not). My question is if itĄ¯s essential to store 32 sps and 256 pps in decoding process because sps id has the range of 0 to 31 and pps id has the range of 0 to 255, inclusive. 3. I cannot understand the use of the syntax elements about POC which determine the display order. I found the JM reference encoder increments POC by 2 for every complete frame. When is the decode order unconformable with the display order? Will frame_num set to zero once IDR access unit appears? Is log2_max_frame_num_minus4 the limit of frame_num? The value of log2_max_frame_num_minus4 is in the range of 0 to 12. Is this means that the max encoded frame number should not exceed (2<<(12+4))? 4. If one design which only supports baseline Profile, will the change of the H.264 standard from the version of 2003.5 to the version of 2005.3 really impact the design? I had browsed the new standard and didnĄ¯t find any change of baseline profile feature. Did I miss anything? I really appreciate your patience to answer my questions. Best wishes, Jely From gbmallikarjunarao yahoo.com Thu Jun 30 13:18:31 2005 From: gbmallikarjunarao yahoo.com (mallikarjun rao) Date: Thu Jun 30 17:31:03 2005 Subject: [Mp4-tech] MP3-joint stereo Message-ID: <20050630111831.1553.qmail@web52013.mail.yahoo.com> Hi , Can any body helpme in the following . 1. In one Mp3 file, both stereo and joint stereo modes, are possible or not? i.e. some frames encoded as stereo and some frames as joint stereo. 2. is it same for both MPEG-1, MPEG-2 layerIII? Thanks Mallikarjun --------------------------------- How much free photo storage do you get? Store your holiday snaps for FREE with Yahoo! Photos. Get Yahoo! Photos -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20050630/0226b571/attachment.html From garysull windows.microsoft.com Thu Jun 30 13:32:26 2005 From: garysull windows.microsoft.com (Gary Sullivan) Date: Thu Jun 30 17:31:09 2005 Subject: [Mp4-tech] RE: some questions about H.264 standard! Message-ID: <91D7F2CEE3425A4A9D11311D09FCE2460F440630@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Jely et al, Reply interspersed. Best Regards, Gary Sullivan +> -----Original Message----- +> From: JelyZhu@innosis.com.cn [mailto:JelyZhu@innosis.com.cn] +> Sent: Thursday, June 30, 2023 1:50 AM +> To: Gary Sullivan +> Cc: mp4-tech@lists.mpegif.org +> Subject: some questions about H.264 standard! +> +> +> Hi, Gary: +> +> Thank you very much for your kindly reply! +> +> I have some other questions about H.264. +> +> 1. As we all know, there are three kinds of samples in +> interpolation: +> quarter-pel samples, half-pel samples and integer-pel +> samples. They have +> different calculation complexity. My interpretation is that the +> quarter-pel interpolation will be used in the detailed +> area and on the +> contrary, the integer-pel interpolation is used in +> homogeneous areas +> which have more probability to appear than the detailed +> area. So can I +> get the statistic of the number of three kinds of +> samples, respectively, +> to evaluate the performance of the interpolation process +> and if this +> evaluation is reasonable or not? I think you may be confusing the concept of when a certain degree of motion vector accuracy would be expected to be beneficial with the concept of when certain values of motion vectors will be used by an encoder. In the H.264/MPEG-4 AVC standard, luma motion vectors are always represented with 1/4-sample precision. I think your interest is primarily on decoding. Your decoder should be able to handle whatever the encoder chooses to use. Statistically, I would expect zero-valued motion vectors to be the most common, since some parts of the video scene are often sitting still. When the motion vector value is not zero, I would expect all fractional positions (i.e., fractional remainders equal to 0/4, 1/4, 2/4, and 3/4) to occur at roughly the same frequency relative to each other (at least for high-resolution video images). When the camera is moving with respect to the imaged content, of course, approximately none of the motion vectors will be zero-valued, so it is important to make a decoder than can handle a full diversity of motion. Ultimately, no assumptions should be made about the relative frequency of different types of motion when building a decoder. Decoders should be able to handle worst-case encoder behavior. +> +> 2. The sps and pps can be activated again once IDR access +> unit appears +> (I don't know my interpretation is correct or not). My +> question is if +> it's essential to store 32 sps and 256 pps in decoding +> process because +> sps id has the range of 0 to 31 and pps id has the range +> of 0 to 255, +> inclusive. Yes, you need to be able to store that many SPS and PPS structures in order to make a decoder that conforms to the standard. +> +> 3. I cannot understand the use of the syntax elements +> about POC which +> determine the display order. I found the JM reference +> encoder increments +> POC by 2 for every complete frame. When is the decode order +> unconformable with the display order? Will frame_num set +> to zero once +> IDR access unit appears? Is log2_max_frame_num_minus4 the limit of +> frame_num? The value of log2_max_frame_num_minus4 is in +> the range of 0 +> to 12. Is this means that the max encoded frame number +> should not exceed +> (2<<(12+4))? Pay no attention to what the reference encoder does. If you are building a decoder, you must do what the spec says and you must not make any assumptions about encoder behavior (except as expressed in mandatory bitstream requirements in the standard). I think the frame_num and the POC of an IDR picture will always be equal to 0. However, there may be pictures that arrive after the IDR picture that have POC values less than 0. +> +> 4. If one design which only supports baseline Profile, +> will the change +> of the H.264 standard from the version of 2003.5 to the version of +> 2005.3 really impact the design? I had browsed the new +> standard and +> didn't find any change of baseline profile feature. Did I +> miss anything? There were a number of small errrors that were fixed. Some details might be wrong if you base your implementation on the first version. Personally, I think it's very important to make sure you're using the latest version of the standard. +> +> I really appreciate your patience to answer my questions. +> +> +> Best wishes, +> Jely +>