From mithun.nayak patni.com Fri Nov 1 09:24:16 2002 From: mithun.nayak patni.com (Mithun N) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] Rounding of motion vectors for chroma? In-Reply-To: Message-ID: <001b01c2815a$59c33400$fe02a8c0@patni.com> hi christoph,, when quarter pel is active, the max accuracy u can seek is 0.25 ..right??? for eg..if the four motion vectors are 8.25, 6.25,3.25.and 5......for the luma blocks then the chrominance motion vector is calculated as mv_chr = sum/4 = 22. 75/4 = 5.6875 but since the chroma vector is calculated as the above given formula...the value can go upto 0.24/4 = 1/16 which is i/4 of quarter pel similarly if there are 3,2or 1 non transparent luma blocks then the motion vectors shall be 1/12,1/8 or 1/4 of quarter pel respectively Now these vectors which can go to the accuracy of 1/16 (0.0625) have to be rounded to the nearest halfpel resolution as given in the table 7.6 - 7.9.... since the result needs to be at a half pixel i hope this answers ur query... regds mithun -----Original Message----- From: technotes-admin@lists.m4if.org [mailto:technotes-admin@lists.m4if.org]On Behalf Of Christoph Lampert Sent: Wednesday, October 30, 2023 8:21 PM To: technotes@lists.m4if.org Subject: [M4IF Technotes] Rounding of motion vectors for chroma? Hi, I have a question about rounding motion vectors, especially for chroma if quarterpel is active. I'll use a simple example. Consider a macro block in inter4v mode with luminance motion vectors mv[0] to mv[3] in halfpel or quarterpel resolution. How do I calculate the motion vector for chroma from them? I know of course that it's mv[0]+mv[1]+mv[2]+mv[3] / (2*4) but how exactly is rounding supposed to be done? What I read from the standard about this is as unclear as possible. Thank you in advance, Christoph -- Christoph H. Lampert chl@math,uni-bonn,de | Diese Signature wurde maschi- Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell erstellt und bedarf Tel. (0228) 73-4708 Fax. +49 228 73-7916 | keiner Unterschrift. AZ 27B-6 _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes From noimark il.ibm.com Fri Nov 1 12:55:59 2002 From: noimark il.ibm.com (Yuval Noimark) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] AC prediction, follow document or follow reference software? In-Reply-To: <3B8749EA44FC2748956538D4A3096E42E08A98@RED-MSG-01.redmond.corp.microsoft.com> Message-ID: We've reported this problem during the last MPEG meeting in Shanghai. The decision was to use zero predictors in order to keep consistency with the reference software and the VM document. Yuval. -------------------------------- Yuval Noimark IBM Research Lab in Haifa "Gary Sullivan" Sent by: technotes-admin@lists.m4if.org 24/10/2023 17:16 To: "Lefan Zhong" , cc: Subject: RE: [M4IF Technotes] AC prediction, follow document or follow reference software? Lefan, That appears to be a bug in the MPEG-4 visual spec. I just looked too, and I couldn't find a clear statement that non-intra macroblocks are not used for AC coefficient prediction of Intra MBs, although I think it would be silly to use them. I don't recall whether that problem has been previously reported to MPEG or not. I will try to do so. Best Regards, Gary Sullivan +> -----Original Message----- +> From: Lefan Zhong [mailto:Lefan@mediaq.com] +> Sent: Wednesday, October 23, 2023 4:25 PM +> To: technotes@lists.m4if.org +> Subject: [M4IF Technotes] AC prediction, follow document or +> follow reference software? +> +> +> Hello folks, +> +> In MPEG-4 document 14496-2, subclause 7.4.3.3, adaptive ac +> coefficient prediction, it says: +> +> If the prediction block (block 'A' or block 'C') is outside +> of the boundary of VOP or video packet, then all the +> prediction coefficients of that block are assumed to be zero. +> +> +> Beware that statement doesn't tell us what to do when the +> prediction block is 'inter'. Should we use the prediction +> block coefficients to do prediction or set to zeros when the +> prediction block is 'inter' type? +> +> In the comformance test bitstream Hit007.m4v, 1st P-VOP, the +> 2nd macroblock (coordinate (16,0)), the macroblock is an +> INTRA block, and AC prediction flag is on, and the 1st +> Y-block needs to do AC prediction from the left block. But +> left block is an INTER block, the AC coefficients are: +> +> -1 0 0 -1 0 0 0 0 +> 1 -1 0 0 0 0 0 0 +> 0 -1 0 0 0 0 0 0 +> 0 0 0 0 0 0 0 0 +> 0 0 0 0 0 0 0 0 +> 0 0 0 0 0 0 0 0 +> 0 0 0 0 0 0 0 0 +> 0 0 0 0 0 0 0 0 +> +> Should we use +> 1024 +> 1 +> 0 +> 0 +> 0 +> 0 +> 0 +> 0 +> to do prediction or use +> 1024 +> 0 +> 0 +> 0 +> 0 +> 0 +> 0 +> 0 +> to do prediction? +> +> In the reference softwares (no matter momusys or microsoft, +> no matter version 1 or version 2), they all use zeros to do +> prediction in such cases. +> +> For scence of prediction, it is better to use zeros in such +> cases. But that is conflict to document. +> +> Thank you for your consideration. +> +> Lefan +> +> _______________________________________________ +> Technotes mailing list +> Technotes@lists.m4if.org +> http://lists.m4if.org/mailman/listinfo/technotes +> _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021101/b137f066/attachment.html From chl math.uni-bonn.de Fri Nov 1 13:56:16 2002 From: chl math.uni-bonn.de (Christoph Lampert) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] Rounding of motion vectors for chroma? In-Reply-To: <001b01c2815a$59c33400$fe02a8c0@patni.com> Message-ID: On Fri, 1 Nov 2002, Mithun N wrote: > hi christoph,, > > when quarter pel is active, the max accuracy u can > seek is > 0.25 ..right??? > [...] > Now these vectors which can go to the accuracy of > 1/16 (0.0625) have to be rounded to the nearest > halfpel resolution > as given in the table 7.6 - 7.9.... since the > result needs to be at a half pixel > i hope this answers ur query... Sorry, not really. Of course I know that vectors have to be rounded and that table 7.6 to 7.9 should be used for that. But the exact method _how_ this is supposed to be done is very unclear. If K=4 and the lumi motion vectors in quarterpel resolution, chroma (mv[0]+mv[1]+mv[2]+mv[3])/(2*K) would be 32th-pel(!) This is not allowed (and there's no table for that), so the standard talks about dividing quarterpel values by 2 before doing the prodecure. (If I am not mistaken, this section has been changed between different versions of drafts etc.) So is rounding supposed to be a two-step process in quarterpel-inter4v mode? First devide by 2. Then use table 7-6? What rounding is used for the first division by 2? Next, what about K=1? Then the result if mv[0]/2 is 8th-pel. Should the 8th-pel table 7-8 be used for rounding as the reference software does (if I remember correctly)? Why isn't it two-step again? And why is the whole process so complicated, why isn't there simply another table for 32th-pel to halfpel? Christoph > > -----Original Message----- > From: technotes-admin@lists.m4if.org > [mailto:technotes-admin@lists.m4if.org]On Behalf > Of Christoph Lampert > Sent: Wednesday, October 30, 2023 8:21 PM > To: technotes@lists.m4if.org > Subject: [M4IF Technotes] Rounding of motion > vectors for chroma? > > > Hi, > > I have a question about rounding motion vectors, > especially for > chroma if quarterpel is active. I'll use a simple > example. > Consider a macro block in inter4v mode with > luminance motion vectors > mv[0] to mv[3] in halfpel or quarterpel > resolution. > > How do I calculate the motion vector for chroma > from them? > I know of course that it's > mv[0]+mv[1]+mv[2]+mv[3] / (2*4) > but how exactly is rounding supposed to be done? > What I read from the > standard about this is as unclear as possible. > > Thank you in advance, > > Christoph > > -- > Christoph H. Lampert chl@math,uni-bonn,de | Diese > Signature wurde maschi- > Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell > erstellt und bedarf > Tel. (0228) 73-4708 Fax. +49 228 73-7916 | keiner > Unterschrift. AZ 27B-6 > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > From amit.klir emblaze.com Mon Nov 4 12:16:38 2002 From: amit.klir emblaze.com (Amit Klir) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] Bitstream syntax Message-ID: <980C7B879697C7448A54C9B065AAFAAC420062@embexcil01.emblaze.ent> Hi, Sometimes there is a need to add VOL header in the middle of the stream, before IVOP for example. Is there a must to add visual_object_sequence_start_code and visual_object_start_code for every insertion of video_object_layer_start_code ? Thanks, Amit ************************************************************************************************** The contents of this email and any attachments are confidential. It is intended for the named recipient(s) only. If you have received this email in error please notify the system manager or the sender immediately and do not disclose the contents to any one or make copies. ** eSafe scanned this email for viruses, vandals and malicious content ** ************************************************************************************************** From chl math.uni-bonn.de Mon Nov 4 12:00:04 2002 From: chl math.uni-bonn.de (Christoph Lampert) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] Parameters if VOL-Header Message-ID: Hi, if I have a simple MPEG-4 video stream, is it possible to change parameters like quarterpel motion compensation and sprite_enable from one VOL-Headers to the next? So e.g. have a VOLheader with sprite_enable='00', then a sequence of say 100 VOPs, then another VOL-Header with sprite_enable='10', again 100 VOPs using these new settings, etc.? Christoph Lampert -- Christoph H. Lampert chl@math,uni-bonn,de | Diese Signature wurde maschi- Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell erstellt und bedarf Tel. (0228) 73-4708 Fax. +49 228 73-7916 | keiner Unterschrift. AZ 27B-6 From d98rolb stud.hh.se Mon Nov 4 19:06:02 2002 From: d98rolb stud.hh.se (Roland Bengtsson) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] Documentation about MPEG-4 Message-ID: <5.1.1.6.0.20021104184312.00b43d08@studpop.hh.se> I'm doing a master thesis about MPEG-4 and searching for material for my report. I know there is a lot written about this subject but I still want to hear some recommendations from experts. I'm searching for papers and reports about: 1. How much CPU-power is required for different algorithms. I'm interrested in both encoding and decoding. Different profiles in MPEG-4, H.263, H.264 etc. I read about AVC that will do miracle and nearly half the bit rate for the same quality. Does this cost cpu-cycles? It's difficult to compare different platforms but I am concentrated on DSP's and embedded systems. 2. Reports about scalability, advantages and disadvantages about algorithms like FGS and other methods. 3. Implementations of MPEG-4 on a DSP like Texas Instruments have. Efficiency and costs are important. I'm hoping for a good discussion about this. Best regards Roland Bengtsson Halmstad University -------------- next part -------------- --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.410 / Virus Database: 231 - Release Date: 2023-10-31 From dwu jdl.ac.cn Tue Nov 5 20:07:51 2002 From: dwu jdl.ac.cn (Wu Di) Date: Wed Jul 30 14:09:08 2003 Subject: [M4IF Technotes] Searching for ASP VM Message-ID: <002e01c284c3$f738d020$222ae29f@sky> Hi, all, I am designing the video codec for Adanced Simpile Profile. Now I have the MPEG-4 Video Verification Model for FGS, which includs not only the base layer - advanced simple profile but also the enhancement layer. However, I want to get a standalone VM for advanced simple profile, which can clarify the algorithms and tools involved for ASP. Can anybody tell me how to get it? Thanks, david -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021105/027f843b/attachment.html From byan cse.cuhk.edu.hk Tue Nov 5 20:50:24 2002 From: byan cse.cuhk.edu.hk (Yan Bo) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] how to get the bit rate for the Mpeg-4 encoder. Message-ID: <200211051250.gA5CoMI28848@cse.cuhk.edu.hk> Hi£¬ I am now using microsoft Mpeg-4 visual codec. Now I have a problem. Can anyone tell me how to get the bit rate of the compressed bit streams? Thanks a lot. Yan Bo Dept. of Computer Science & Engineering The Chinese University of Hong Kong From rob.koenen m4if.org Tue Nov 5 15:19:35 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] Searching for ASP VM In-Reply-To: <002e01c284c3$f738d020$222ae29f@sky> Message-ID: <3C124172E7FDD511B510000347426D59011B3F79@exchange.epr.com> There is no separate VM for ASP. Just disable the FGS functionality, that's your best shot. Rob -----Original Message----- From: Wu Di [mailto:dwu@jdl.ac.cn] Sent: Tuesday, November 05, 2023 4:08 To: technotes@lists.m4if.org Cc: Discuss@lists.m4if.org Subject: [M4IF Technotes] Searching for ASP VM Hi, all, I am designing the video codec for Adanced Simpile Profile. Now I have the MPEG-4 Video Verification Model for FGS, which includs not only the base layer - advanced simple profile but also the enhancement layer. However, I want to get a standalone VM for advanced simple profile, which can clarify the algorithms and tools involved for ASP. Can anybody tell me how to get it? Thanks, david -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021105/492355ee/attachment.html From dwu jdl.ac.cn Wed Nov 6 11:46:05 2002 From: dwu jdl.ac.cn (Wu Di) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] Question for "video_object_layer_verid" Message-ID: <001401c28547$08c1dc50$222ae29f@sky> Hi, all, I am confused by the parameter - "video_object_layer_verid" in VideoObjectLayer of Advance Simple Profile bitstream. When I use the VM for FGS to compress ASP bitstream, the "video_object_layer_verid" is "0010". But in the standard ASP bitstream, this field is "0101". The question is whether this field can be chosen from "0001", "0010", "0100" and "0101", or definitely be "0101". thanks. david -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021106/92d46c43/attachment.html From pescatorimarco hotmail.com Wed Nov 6 15:20:31 2002 From: pescatorimarco hotmail.com (Marco Pescatori) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] FGS Message-ID: I have a question;what is the FGS and its applications?Is it similar to the other scalabilty type(Spatial/Temporal)? If somebody can help me please write to me. Thanks. Best Regards. Marco. _________________________________________________________________ Vuoi chattare e scambiare canzoni e foto con gli amici? http://messenger.msn.it From dwu jdl.ac.cn Wed Nov 6 22:35:56 2002 From: dwu jdl.ac.cn (Wu Di) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] A question Message-ID: <000a01c285a1$d1498380$222ae29f@sky> Hi, all, If there are no timestamps( PTS or DTS) in the MPEG-4 video stream, how to control the video decoder to work in real time and avoid the bit buffer neither overflow nor underflow. thanks, wudi. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021106/5bb8f34c/attachment.html From h8f gentek.com.cn Tue Nov 5 09:52:57 2002 From: h8f gentek.com.cn (=?GB2312?Q?=BB=C6=B0=CE=B7=E5?=) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] the problem happening in link Message-ID: <200211050158.gA51vWvh011037@mx3.magma.ca> Hi all The following happened when I rebuilt the Imp1(2001) Linking... Executiv.obj : error LNK2001: unresolved external symbol "public: virtual class Decoder * __thiscall InputSensorImp::CreateDecoder(class DecoderConfigDescriptor *,class QoS_Descriptor *,class Decoder *)" (?CreateDecoder@InputSensorImp@@UAEPAVDecoder @@PAVDecoderConfigDescriptor@@PAVQoS_Descriptor@@PAV2@@Z) Executiv.obj : error LNK2001: unresolved external symbol "public: virtual void __thiscall InputSensorImp::DeleteObject(void)" (?DeleteObject@InputSensorImp@@UAEXXZ) ./Debug/Im1Player.exe : fatal error LNK1120: 2 unresolved externals Error executing link.exe. Any answer is appreciated very much. Best regards Luoba Hwang Donghua University ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡h8f@gentek.com.cn ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡2002-11-05 From rob.koenen m4if.org Wed Nov 6 11:45:55 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] FGS In-Reply-To: Message-ID: <3C124172E7FDD511B510000347426D59011B3FA5@exchange.epr.com> FGS is Fine Grain Scalability. There are many applications, including delivery of pre-encoded content over networks where the bandwidth is unknown a-priori (e.g., IP) and/or can vary in real-time (e.g., mobile). Another application is statistical multiplexing of pre-eoncoded programs for broadcast applications. FGS adds layers of quality to a base layer, 11 in the case of MPEG-4, so you can do smooth transitions between different qualities and bitrates. In MPEG-4 Audio, it is even possible to add 1 kbit/s layers. More information is in the MPEG-4 Overview, http://www.m4if.org/resources/Overview.pdf Rob > -----Original Message----- > From: Marco Pescatori [mailto:pescatorimarco@hotmail.com] > Sent: Wednesday, November 06, 2023 6:21 > To: technotes@lists.m4if.org > Subject: [M4IF Technotes] FGS > > > > I have a question;what is the FGS and its applications?Is it > similar to the other scalabilty type(Spatial/Temporal)? If > somebody can help me please write to me. Thanks. > > > Best Regards. > > > Marco. > > > > > _________________________________________________________________ > Vuoi chattare e scambiare canzoni e foto con gli amici? > http://messenger.msn.it > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technot> es > From mithun.nayak patni.com Thu Nov 7 09:18:21 2002 From: mithun.nayak patni.com (Mithun N) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] Rounding of motion vectors for chroma? In-Reply-To: Message-ID: <002101c28610$84d140a0$fe02a8c0@patni.com> hi.. the motion vectors cant go to an accuracy of 1/32, the reason being that there can be max of 4 non transparent blocks in a macroblock, each having an accuracy of 1/4...which comes out to 1/16 and not 1/32. and as far as using the table is concerned the rounding is very clearly mentioned _how_ to be done. for eg.. if a macroblock has 4 non transparent blocks,and you are calculating the chroma vector for the same then add the 4 motion vectors(each with an accuracy of 1/4) Divide the sum by 4....the so obtained accuracy will be 1/16... Now check the table 7.6 which shows the rounding for 1/16 accuracy. If the fractional part of the motion vector is between 0/16 to 2/16 then the resulting fractional part becomes 0/2=0.0 If the fractional part of the motion vector is between 3/16 to 13/16 then the resulting fractional part becomes 1/2=0.5 If the fractional part of the motion vector is between 14/16 to 15/16 then the resulting fractional part becomes 2/2=1.0 similarly if u have 3 nontransparent blocks ...you will have amax accuracy of 1/4*1/3 =1/12..then check table 7.7 and perform similar steps as in table 7.6 Do ask if still this doesnt solve your query regards Mithun Nayak -----Original Message----- From: technotes-admin@lists.m4if.org [mailto:technotes-admin@lists.m4if.org]On Behalf Of Christoph Lampert Sent: Friday, November 01, 2023 6:26 PM To: technotes@lists.m4if.org Subject: RE: [M4IF Technotes] Rounding of motion vectors for chroma? Sorry, not really. Of course I know that vectors have to be rounded and that table 7.6 to 7.9 should be used for that. But the exact method _how_ this is supposed to be done is very unclear. If K=4 and the lumi motion vectors in quarterpel resolution, chroma (mv[0]+mv[1]+mv[2]+mv[3])/(2*K) would be 32th-pel(!) This is not allowed (and there's no table for that), so the standard talks about dividing quarterpel values by 2 before doing the prodecure. (If I am not mistaken, this section has been changed between different versions of drafts etc.) So is rounding supposed to be a two-step process in quarterpel-inter4v mode? First devide by 2. Then use table 7-6? What rounding is used for the first division by 2? Next, what about K=1? Then the result if mv[0]/2 is 8th-pel. Should the 8th-pel table 7-8 be used for rounding as the reference software does (if I remember correctly)? Why isn't it two-step again? And why is the whole process so complicated, why isn't there simply another table for 32th-pel to halfpel? Christoph > > -----Original Message----- > From: technotes-admin@lists.m4if.org > [mailto:technotes-admin@lists.m4if.org]On Behalf > Of Christoph Lampert > Sent: Wednesday, October 30, 2023 8:21 PM > To: technotes@lists.m4if.org > Subject: [M4IF Technotes] Rounding of motion > vectors for chroma? > > > Hi, > > I have a question about rounding motion vectors, > especially for > chroma if quarterpel is active. I'll use a simple > example. > Consider a macro block in inter4v mode with > luminance motion vectors > mv[0] to mv[3] in halfpel or quarterpel > resolution. > > How do I calculate the motion vector for chroma > from them? > I know of course that it's > mv[0]+mv[1]+mv[2]+mv[3] / (2*4) > but how exactly is rounding supposed to be done? > What I read from the > standard about this is as unclear as possible. > > Thank you in advance, > > Christoph > > -- > Christoph H. Lampert chl@math,uni-bonn,de | Diese > Signature wurde maschi- > Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell > erstellt und bedarf > Tel. (0228) 73-4708 Fax. +49 228 73-7916 | keiner > Unterschrift. AZ 27B-6 > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes From cloverleafland mail.hongkong.com Thu Nov 7 17:49:34 2002 From: cloverleafland mail.hongkong.com (cloverleafland@mail.hongkong.com) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] Re: motion estimation, please advise Message-ID: Dear all, From mithun.nayak patni.com Thu Nov 7 16:00:57 2002 From: mithun.nayak patni.com (Mithun N) Date: Wed Jul 30 14:09:09 2003 Subject: [M4IF Technotes] Re: motion estimation, please advise In-Reply-To: Message-ID: <00c301c28648$c2db1460$fe02a8c0@patni.com> hi i didnt understand your first question properly.. Could you clarify it a bit further?? As for the Motion estimation, there are papers available of the type of method to be used .. A combination of three step search and four step search is useful in case the motion isnt too much between 2 time instants. This method is generally called Majority Voting algorithm where the encoder decides between the 2 methods by using the surrounding macroblock motion vectors. This method is more useful if u plan to build a real time encoder. Else there are other methods such as Full Search.. which take a lot of processing time. regards Mithun Nayak -----Original Message----- From: technotes-admin@lists.m4if.org [mailto:technotes-admin@lists.m4if.org]On Behalf Of cloverleafland@mail.hongkong.com Sent: Thursday, November 07, 2023 3:20 PM To: technotes@lists.m4if.org Cc: discuss@lists.m4if.org Subject: [M4IF Technotes] Re: motion estimation, please advise Dear all, >From previous questions on motion estimation, it can be done on Y component in each macroblock or combining a fraction of Cr and Cb together with Y into a 16x16 blocks for estimation. I would like to ask is there any further information on which method is better based on quality, speed and other performance? any source or papers comparing the performance of these two methods that I would find? Thanks very much. Regards, Clover --------------------------------------------- ?w??????HongKong.com?l???t?? Thank you for using hongkong.com Email system _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes From chl math.uni-bonn.de Thu Nov 7 11:47:25 2002 From: chl math.uni-bonn.de (Christoph Lampert) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] Rounding of motion vectors for chroma? In-Reply-To: <002101c28610$84d140a0$fe02a8c0@patni.com> Message-ID: On Thu, 7 Nov 2002, Mithun N wrote: > hi.. > the motion vectors cant go to an accuracy of 1/32, > the reason being that there can be max of 4 non > transparent > blocks in a macroblock, each having an accuracy of > 1/4...which > comes out to 1/16 and not 1/32. Sorry to still bother you, but this is not the point. Since chroma is 1:2 _subsampled_, you have to divide chroma value not by K but by 2*K ! This _is_ clearly stated in the standard (right atop of the tables 7-6 to 7-9) and it's also completely logical. halfpel / (2*4) is 16th-pel, okay, no problem. But for quarterpel this is _not_ working, because quarterpel / (2*4) is 32th-pel ! I read the standard that this is not supposed to be done, but that _before_ that quarterpel-values should be divided by 2 to obtain halfpel and _then_ add up and divide by 2*K. However there is no operator given for this dividing by 2. So should there be no rounding? And "no rounding" means round to -infinity? Then if one component in qpel would get values 3/4, 3/4, 1/4, 1/4 the resulting vector would be (1+1+0+0)/2 /(4*2) ---Table_7.6---> 0/2 I simply find this very strange and non-symmetric and would like to get confirmation. Christoph > and as far as using the table is concerned the > rounding is very > clearly mentioned _how_ to be done. > for eg.. > if a macroblock has 4 non transparent blocks,and > you are calculating > the chroma vector for the same then add the 4 > motion vectors(each with an accuracy of 1/4) > Divide the sum by 4....the so obtained accuracy > will be 1/16... > Now check the table 7.6 which shows the rounding > for 1/16 accuracy. > If the fractional part of the motion vector is > between 0/16 to 2/16 then the resulting fractional > part becomes 0/2=0.0 > If the fractional part of the motion vector is > between 3/16 to 13/16 then the resulting > fractional part becomes 1/2=0.5 > If the fractional part of the motion vector is > between 14/16 to 15/16 then the resulting > fractional part becomes 2/2=1.0 > similarly if u have 3 nontransparent blocks ...you > will have amax accuracy of 1/4*1/3 =1/12..then > check table 7.7 and perform similar steps as in > table 7.6 > Do ask if still this doesnt solve your query > regards > Mithun Nayak > > -----Original Message----- > From: technotes-admin@lists.m4if.org > [mailto:technotes-admin@lists.m4if.org]On Behalf > Of Christoph Lampert > Sent: Friday, November 01, 2023 6:26 PM > To: technotes@lists.m4if.org > Subject: RE: [M4IF Technotes] Rounding of motion > vectors for chroma? > > > Sorry, not really. > > Of course I know that vectors have to be rounded > and that table 7.6 to > 7.9 should be used for that. But the exact method > _how_ this is supposed > to be done is very unclear. > > If K=4 and the lumi motion vectors in quarterpel > resolution, > chroma (mv[0]+mv[1]+mv[2]+mv[3])/(2*K) would be > 32th-pel(!) > > This is not allowed (and there's no table for > that), so the standard > talks about dividing quarterpel values by 2 before > doing the prodecure. > (If I am not mistaken, this section has been > changed between different > versions of drafts etc.) > So is rounding supposed to be a two-step process > in quarterpel-inter4v > mode? First devide by 2. Then use table 7-6? What > rounding is used for the > first division by 2? > > Next, what about K=1? Then the result if mv[0]/2 > is 8th-pel. > Should the 8th-pel table 7-8 be used for rounding > as the reference > software does (if I remember correctly)? Why isn't > it two-step again? > > And why is the whole process so complicated, why > isn't there simply > another table for 32th-pel to halfpel? > > Christoph > > > > > > > > -----Original Message----- > > From: technotes-admin@lists.m4if.org > > [mailto:technotes-admin@lists.m4if.org]On Behalf > > Of Christoph Lampert > > Sent: Wednesday, October 30, 2023 8:21 PM > > To: technotes@lists.m4if.org > > Subject: [M4IF Technotes] Rounding of motion > > vectors for chroma? > > > > > > Hi, > > > > I have a question about rounding motion vectors, > > especially for > > chroma if quarterpel is active. I'll use a > simple > > example. > > Consider a macro block in inter4v mode with > > luminance motion vectors > > mv[0] to mv[3] in halfpel or quarterpel > > resolution. > > > > How do I calculate the motion vector for chroma > > from them? > > I know of course that it's > > mv[0]+mv[1]+mv[2]+mv[3] / (2*4) > > but how exactly is rounding supposed to be done? > > What I read from the > > standard about this is as unclear as possible. > > > > Thank you in advance, > > > > Christoph > > > > -- > > Christoph H. Lampert chl@math,uni-bonn,de | > Diese > > Signature wurde maschi- > > Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell > > erstellt und bedarf > > Tel. (0228) 73-4708 Fax. +49 228 73-7916 | > keiner > > Unterschrift. AZ 27B-6 > > > > _______________________________________________ > > Technotes mailing list > > Technotes@lists.m4if.org > > http://lists.m4if.org/mailman/listinfo/technotes > > > > _______________________________________________ > > Technotes mailing list > > Technotes@lists.m4if.org > > http://lists.m4if.org/mailman/listinfo/technotes > > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > > From rob.koenen m4if.org Thu Nov 7 09:01:38 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] RE: [M4IF Discuss] Re: motion estimation, please advise In-Reply-To: Message-ID: <3C124172E7FDD511B510000347426D59011B3FC6@exchange.epr.com> Thanks Clover. Please keep these posts on the Technotes lists exclusively. Rob > -----Original Message----- > From: cloverleafland@mail.hongkong.com > [mailto:cloverleafland@mail.hongkong.com] > Sent: Thursday, November 07, 2023 1:50 > To: technotes@lists.m4if.org > Cc: discuss@lists.m4if.org > Subject: [M4IF Discuss] Re: motion estimation, please advise > > > Dear all, > > From previous questions on motion estimation, it can be done > on Y component in each macroblock or combining a fraction of > Cr and Cb together with Y into a 16x16 blocks for estimation. > I would like to ask is there any further information on which > method is better based on quality, speed and other > performance? any source or papers comparing the performance > of these two methods that I would find? > Thanks very much. > > Regards, > Clover > --------------------------------------------- > ?w??????HongKong.com?l???t?? > Thank you for using hongkong.com Email system > > _______________________________________________ > Discuss mailing list > Discuss@lists.m4if.org http://lists.m4if.org/mailman/listinfo/discuss > From cloverleafland mail.hongkong.com Fri Nov 8 11:13:24 2002 From: cloverleafland mail.hongkong.com (cloverleafland@mail.hongkong.com) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] Re: motion estimation, please advise Message-ID: Hi.. Thanks for your reply.... actually I would like to use PMVFAST method... but I don't know would there be help if using a combination of Y, Cr, Cb elements during the estimation in PMVFAST method. Besides, I am confused that, isn't that the same between different methods?as the criteria for finding a motion vector is by the minimum SAD value? regards, Clover > hi > i didnt understand your first question properly.. > Could you clarify it a bit further?? > As for the Motion estimation, there are papers > available of the type of method to be used .. > A combination of three step search and four step > search is useful in case the motion isnt too much > between 2 time instants. > This method is generally called Majority Voting > algorithm where the encoder decides between > the 2 methods by using the surrounding macroblock > motion vectors. > This method is more useful if u plan to build a > real time > encoder. > Else there are other methods such as Full Search.. > which take a lot of processing time. > regards > Mithun Nayak > > > -----Original Message----- > From: technotes-admin@lists.m4if.org > [mailto:technotes-admin@lists.m4if.org]On Behalf > Of > cloverleafland@mail.hongkong.com > Sent: Thursday, November 07, 2023 3:20 PM > To: technotes@lists.m4if.org > Cc: discuss@lists.m4if.org > Subject: [M4IF Technotes] Re: motion estimation, > please advise > > > Dear all, > > >From previous questions on motion estimation, it > can be done on Y component in each macroblock or > combining a fraction of Cr and Cb together with Y > into a 16x16 blocks for estimation. I would like > to ask is there any further information on which > method is better based on quality, speed and other > performance? any source or papers comparing the > performance of these two methods that I would > find? > Thanks very much. > > Regards, > Clover > --------------------------------------------- > Åwªï¨Ï¥ÎHongKong.com¶l¥ó¨t²Î > Thank you for using hongkong.com Email system > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > ---------------------------------------------- Åwªï¨Ï¥ÎHongKong.com¶l¥ó¨t²Î Thank you for using hongkong.com Email system From xxh soft.thtf.com.cn Fri Nov 8 11:34:55 2002 From: xxh soft.thtf.com.cn (Xue Xuhong) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] About Advanced Simple Profile tools Message-ID: <001d01c286d7$ce648520$32d06fa6@XUEXUHONG> Hi, I am doing some work about Advanced Simple Profile(ASP), I know three tools are used in ASP,they are: 1)Interlace 2)GMC 3)1/4-pel MC I wonder if there are bitstream files that only used Interlace tool, only GMC tool, or only 1/4-pel MC tool? That is, should the three tools must be used together to implement ASP, they can not be used separately? Thank you. Regards, Xue Xuhong From mithun.nayak patni.com Fri Nov 8 09:44:57 2002 From: mithun.nayak patni.com (Mithun N) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] Re: motion estimation, please advise In-Reply-To: Message-ID: <001a01c286dd$665f2100$fe02a8c0@patni.com> Hi.. Motion estimation isnt performed on Cr and Cb components. As such I have no idea about the PMVFAST method. Guess its based on the diamond search method...with a large diamond and small diamond pattern. Ya but since all the methods are based on correlation between the 2 frames, they essentially need to find at what point in the reference frame, the correlation is maximum.(to achieve as small VLC for the error block as possible) The only difference between the methods is in the points at which the search is done in the reference frame. regards Mithun Nayak -----Original Message----- From: cloverleafland@mail.hongkong.com [mailto:cloverleafland@mail.hongkong.com] Sent: Friday, November 08, 2023 8:43 AM To: ; Cc: discuss@lists.m4if.org Subject: Re: RE: [M4IF Technotes] Re: motion estimation, please advise Hi.. Thanks for your reply.... actually I would like to use PMVFAST method... but I don't know would there be help if using a combination of Y, Cr, Cb elements during the estimation in PMVFAST method. Besides, I am confused that, isn't that the same between different methods?as the criteria for finding a motion vector is by the minimum SAD value? regards, Clover > hi > i didnt understand your first question properly.. > Could you clarify it a bit further?? > As for the Motion estimation, there are papers > available of the type of method to be used .. > A combination of three step search and four step > search is useful in case the motion isnt too much > between 2 time instants. > This method is generally called Majority Voting > algorithm where the encoder decides between > the 2 methods by using the surrounding macroblock > motion vectors. > This method is more useful if u plan to build a > real time > encoder. > Else there are other methods such as Full Search.. > which take a lot of processing time. > regards > Mithun Nayak > > > -----Original Message----- > From: technotes-admin@lists.m4if.org > [mailto:technotes-admin@lists.m4if.org]On Behalf > Of > cloverleafland@mail.hongkong.com > Sent: Thursday, November 07, 2023 3:20 PM > To: technotes@lists.m4if.org > Cc: discuss@lists.m4if.org > Subject: [M4IF Technotes] Re: motion estimation, > please advise > > > Dear all, > > >From previous questions on motion estimation, it > can be done on Y component in each macroblock or > combining a fraction of Cr and Cb together with Y > into a 16x16 blocks for estimation. I would like > to ask is there any further information on which > method is better based on quality, speed and other > performance? any source or papers comparing the > performance of these two methods that I would > find? > Thanks very much. > > Regards, > Clover > --------------------------------------------- > ?w??????HongKong.com?l???t?? > Thank you for using hongkong.com Email system > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > ---------------------------------------------- ?w??????HongKong.com?l???t?? Thank you for using hongkong.com Email system From rob.koenen m4if.org Thu Nov 7 21:48:24 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] About Advanced Simple Profile tools In-Reply-To: <001d01c286d7$ce648520$32d06fa6@XUEXUHONG> Message-ID: <3C124172E7FDD511B510000347426D59011B3FFE@exchange.epr.com> In an ENcoder, you can do what you want as long as you do not use tools that aren't in the profile (and you observe the level restrictions). If you use only a part of the too set, you can still output a compliant bistream (although your quality may be significant behind that of your competition) I na DEcoder, you MUST implement all tools in the profile to be compliant. Of course there are more tools than just the three you mention, such as B frames (and P frames, etc. etc.) Note that Interlace is not required in the lower levels (1,2,3 and 3a) For more information, see http://www.m4if.org/resources.php towards the bottom. Best, Rob > -----Original Message----- > From: Xue Xuhong [mailto:xxh@soft.thtf.com.cn] > Sent: Thursday, November 07, 2023 19:35 > To: technotes@lists.m4if.org > Subject: [M4IF Technotes] About Advanced Simple Profile tools > > > Hi, > > I am doing some work about Advanced Simple Profile(ASP), I > know three tools are used in ASP,they are: 1)Interlace 2)GMC > 3)1/4-pel MC > > I wonder if there are bitstream files that only used > Interlace tool, only GMC tool, or only 1/4-pel MC tool? > > That is, should the three tools must be used together to > implement ASP, they can not be used separately? > > > > Thank you. > > Regards, > Xue Xuhong > > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > From xxh soft.thtf.com.cn Fri Nov 8 14:17:04 2002 From: xxh soft.thtf.com.cn (Xue Xuhong) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] About Advanced Simple Profile tools References: <3C124172E7FDD511B510000347426D59011B3FFE@exchange.epr.com> Message-ID: <003201c286ee$757deac0$32d06fa6@XUEXUHONG> Hi, Rob: I am developing an ASP decoder, and what I based is Core Profiledecoder. So that only three tools need to be implement. They are: 1)Interlace 2)GMC 3)1/4-pel MC To implement the ASP decoder, I should implement all tools in the profile to be compliant, but when I need to test the ASP decoder ,I use the following bitstream files: 1) bitstream file only used Interlace( based on Core profile ) 2)bitstream file only used GMC ( based on Core profile ) 3)bitstream file only used 1/4-pel MC( based on Core profile ) 4)bitstream file used both Interlace and GMC( based on Core profile ) 5)bitstream file used both Interlace and 1/4-pel MC( based on Core profile ) 6)bitstream file used both GMC and 1/4-pel MC( based on Core profile ) 7)bitstream file used Interlace,GMC and 1/4-pel MC( based on Core profile ) Is it OK? Thank you. Regards, Xue Xuhong ----- Original Message ----- From: "Rob Koenen" To: "'Xue Xuhong'" ; Sent: Friday, November 08, 2023 1:48 PM Subject: RE: [M4IF Technotes] About Advanced Simple Profile tools > In an ENcoder, you can do what you want as long as you do not use > tools that aren't in the profile (and you observe the level > restrictions). > If you use only a part of the too set, you can still output a compliant > bistream (although your quality may be significant behind that of your > competition) > > I na DEcoder, you MUST implement all tools in the profile to be > compliant. > > Of course there are more tools than just the three you mention, such > as B frames (and P frames, etc. etc.) > > Note that Interlace is not required in the lower levels (1,2,3 and 3a) > > For more information, see http://www.m4if.org/resources.php towards > the bottom. > > Best, > Rob > > > -----Original Message----- > > From: Xue Xuhong [mailto:xxh@soft.thtf.com.cn] > > Sent: Thursday, November 07, 2023 19:35 > > To: technotes@lists.m4if.org > > Subject: [M4IF Technotes] About Advanced Simple Profile tools > > > > > > Hi, > > > > I am doing some work about Advanced Simple Profile(ASP), I > > know three tools are used in ASP,they are: 1)Interlace 2)GMC > > 3)1/4-pel MC > > > > I wonder if there are bitstream files that only used > > Interlace tool, only GMC tool, or only 1/4-pel MC tool? > > > > That is, should the three tools must be used together to > > implement ASP, they can not be used separately? > > > > > > > > Thank you. > > > > Regards, > > Xue Xuhong > > > > > > > > _______________________________________________ > > Technotes mailing list > > Technotes@lists.m4if.org > > http://lists.m4if.org/mailman/listinfo/technotes > > From rob.koenen m4if.org Thu Nov 7 22:19:26 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] About Advanced Simple Profile tools In-Reply-To: <003201c286ee$757deac0$32d06fa6@XUEXUHONG> Message-ID: <3C124172E7FDD511B510000347426D59011B4005@exchange.epr.com> Don't ask me to devise a testing strategy for you :-) There are conformance bitstreams for ASP but I am unsure about their general availability. Perhaps someone else on the list knows if they are already available or when they will be if not now. In the future, M4IF will be able to help you with bistreams, but that is of little help for you now. Good luck, Rob > -----Original Message----- > From: Xue Xuhong [mailto:xxh@soft.thtf.com.cn] > Sent: Thursday, November 07, 2023 22:17 > To: rob.koenen@m4if.org; technotes@lists.m4if.org > Subject: Re: [M4IF Technotes] About Advanced Simple Profile tools > > > Hi, Rob: > > I am developing an ASP decoder, and what I based is Core > Profiledecoder. So that only three tools need to be > implement. They are: 1)Interlace 2)GMC 3)1/4-pel MC > > To implement the ASP decoder, I should implement all tools in > the profile to be compliant, but when I need to test the ASP > decoder ,I use the following bitstream files: > 1) bitstream file only used Interlace( based on Core profile > ) 2)bitstream file only used GMC ( based on Core profile ) > 3)bitstream file only used 1/4-pel MC( based on Core profile > ) 4)bitstream file used both Interlace and GMC( based on > Core profile ) 5)bitstream file used both Interlace and > 1/4-pel MC( based on Core profile ) 6)bitstream file used > both GMC and 1/4-pel MC( based on Core profile ) 7)bitstream > file used Interlace,GMC and 1/4-pel MC( based on Core profile ) > > Is it OK? > > Thank you. > > Regards, > Xue Xuhong > > > > > > > ----- Original Message ----- > From: "Rob Koenen" > To: "'Xue Xuhong'" ; > Sent: Friday, November 08, 2023 1:48 PM > Subject: RE: [M4IF Technotes] About Advanced Simple Profile tools > > > > In an ENcoder, you can do what you want as long as you do not use > > tools that aren't in the profile (and you observe the level > > restrictions). If you use only a part of the too set, you can still > > output a compliant bistream (although your quality may be > significant > > behind that of your > > competition) > > > > I na DEcoder, you MUST implement all tools in the profile to be > > compliant. > > > > Of course there are more tools than just the three you > mention, such > > as B frames (and P frames, etc. etc.) > > > > Note that Interlace is not required in the lower levels > (1,2,3 and 3a) > > > > For more information, see http://www.m4if.org/resources.php towards > > the bottom. > > > > Best, > > Rob > > > > > -----Original Message----- > > > From: Xue Xuhong [mailto:xxh@soft.thtf.com.cn] > > > Sent: Thursday, November 07, 2023 19:35 > > > To: technotes@lists.m4if.org > > > Subject: [M4IF Technotes] About Advanced Simple Profile tools > > > > > > > > > Hi, > > > > > > I am doing some work about Advanced Simple Profile(ASP), I know > > > three tools are used in ASP,they are: 1)Interlace 2)GMC > 3)1/4-pel MC > > > > > > I wonder if there are bitstream files that only used > Interlace tool, > > > only GMC tool, or only 1/4-pel MC tool? > > > > > > That is, should the three tools must be used together to > implement > > > ASP, they can not be used separately? > > > > > > > > > > > > Thank you. > > > > > > Regards, > > > Xue Xuhong > > > > > > > > > > > > _______________________________________________ > > > Technotes mailing list > > > Technotes@lists.m4if.org > > > http://lists.m4if.org/mailman/listinfo/technotes > > > > From cloverleafland mail.hongkong.com Sat Nov 9 18:50:35 2002 From: cloverleafland mail.hongkong.com (cloverleafland@mail.hongkong.com) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] Re: Qs on Motion estimation again Message-ID: Hi all, From cloverleafland mail.hongkong.com Sun Nov 10 17:53:16 2002 From: cloverleafland mail.hongkong.com (cloverleafland@mail.hongkong.com) Date: Wed Jul 30 14:09:10 2003 Subject: [M4IF Technotes] Re: about vlc Message-ID: Dear all, I would like to ask what's the use of "0000001" the 7 bits eacape code in VLC table for Intra and Inter encoding? Regards, Clover ---------------------------------------------- ¤¤¤å¦rhtml\305w\252\357\250\317\245\316 hongkong.com \271q\244l\266l\245\363\250t\262\316 Thank you for using hongkong.com Email system From cloverleafland mail.hongkong.com Mon Nov 11 19:25:15 2002 From: cloverleafland mail.hongkong.com (cloverleafland@mail.hongkong.com) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] Re: about vlc Message-ID: hello, Thanks for your reply, Further Qs on vlc.. From the Table B-18 in the standard, when LEVEL = -2048 and 0, there's no code representing these two situation, it's stated with "forbidden" only. So may I know how to treat with these values? Besides, from the standard, there are three modes to encode such situation when short_video_header = 0, May I knwo are there any criteria on which modes to be used for encoding? or we would simply freely choose one for encoding. Thanks... Regards, Clover > > Hai Cloverleafland, > > Many possible combinations of runs and levels have no variable length code > to represent them. > In order to encode these statistically rare combinations an Escape Coding > method is used. > > Please refer the Standard: Chapter: 7.3 Texture decoding > in particular : 7.3.1.3 Escape code > > Regards, > Vinayagam.M > ----------------------------------------------- > M.Vinayagam > Senior Software Engineer > L&T Infotech Limited > KIADB Industrial Estate, > Hootagalli, > Mysore - 571 186 > Tel: 91-821-40561 Ext:2613 > E-mail: vinayagam_m@yahoo.com > --------------------------------------------- Åwªï¨Ï¥ÎHongKong.com¶l¥ó¨t²Î Thank you for using hongkong.com Email system From moses kampsax.dtu.dk Mon Nov 11 19:40:22 2002 From: moses kampsax.dtu.dk (moses wold) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] I'm looking error resilience documents, which is used in the MoMuSys and MPEG4IP Message-ID: hi all. I'd appreciate it, if you could help me. I need the document for the Error resilience, which is used in the MoMuSys and MPEG4IP. I found the document of MPEG4IP from TOSHIBA web site, but it was in japan. any link. i'm looking forward to get your help. Best regards Mosese <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> YOUR BRAIN IS HUNGRY Wireless Networks Technology, News and Opinion. <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> From garysull microsoft.com Mon Nov 11 11:25:43 2002 From: garysull microsoft.com (Gary Sullivan) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] Re: about vlc Message-ID: <3B8749EA44FC2748956538D4A3096E42E08BEC@RED-MSG-01.redmond.corp.microsoft.com> +> -----Original Message----- +> From: cloverleafland@mail.hongkong.com +> [mailto:cloverleafland@mail.hongkong.com] +> Sent: Monday, November 11, 2023 3:25 AM +> To: technotes@lists.m4if.org +> Subject: Re: [M4IF Technotes] Re: about vlc +> +> +> hello, +> +> Thanks for your reply, +> Further Qs on vlc.. From the Table B-18 in the standard, +> when LEVEL = -2048 and 0, there's no code representing these +> two situation, it's stated with "forbidden" only. So may I +> know how to treat with these values? Sending 0 doesn't make much sense - just increase the RUN if a LEVEL value is zero. And if you get the urge to send -2048, just send -2047 instead. +> Besides, from the standard, there are three modes to encode +> such situation when short_video_header = 0, May I knwo are +> there any criteria on which modes to be used for encoding? +> or we would simply freely choose one for encoding. I think you can use whichever one you want. In fact I think you can use different escape coding methods for different coefficients in the same IDCT block if you want to. One good strategy may be to use the escape coding method that uses the fewest bits for each coefficient. Best Regards, -Gary S. +> Thanks... +> +> Regards, +> Clover +> +> > +> > Hai Cloverleafland, +> > +> > Many possible combinations of runs and levels have no +> variable length code +> > to represent them. +> > In order to encode these statistically rare combinations +> an Escape Coding +> > method is used. +> > +> > Please refer the Standard: Chapter: 7.3 Texture decoding +> > in particular : 7.3.1.3 Escape code +> > +> > Regards, +> > Vinayagam.M +> > ----------------------------------------------- +> > M.Vinayagam +> > Senior Software Engineer +> > L&T Infotech Limited +> > KIADB Industrial Estate, +> > Hootagalli, +> > Mysore - 571 186 +> > Tel: 91-821-40561 Ext:2613 +> > E-mail: vinayagam_m@yahoo.com +> > +> --------------------------------------------- +> ?w??????HongKong.com?l???t?? +> Thank you for using hongkong.com Email system +> +> _______________________________________________ +> Technotes mailing list +> Technotes@lists.m4if.org +> http://lists.m4if.org/mailman/listinfo/technotes +> From ben interframemedia.com Mon Nov 11 13:06:50 2002 From: ben interframemedia.com (Ben Waggoner) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] Call for MPEG-4 Encoders - Robert X. Cringely project Message-ID: Everybody, Okay, I've got a Big Deal to announce. Robert X. Cringely, computer industry journalist and broadcaster extraordinaire, is launching a new internet talk show, NerdTV, using MPEG-4 technology. Anyone familiar with Bob's past PBS series like "The Triumph of the Nerds" and his seminal book "Accidental Empires" knows this is a big deal. Here are some past columns from him describing this project, in chronological order: http://www.pbs.org/cringely/pulpit/pulpit20020321.html http://www.pbs.org/cringely/pulpit/pulpit20020328.html http://www.pbs.org/cringely/pulpit/pulpit20020912.html http://www.pbs.org/cringely/pulpit/pulpit20020704.html I've been consulting with him on the media delivery technologies part of the project, and we've decided that it's progressive download MPEG-4 all the way for the video part. Users will even be welcome to download and edit the camera sources, and edit together their own versions! All going well, this will be the first really big public use of MPEG-4, and while we picked the technology on its own merits, it's my hope that this can help jump start the industry some in the video-playing-on-computers market. The next phase of the project is to pick the right encoder, hence this email. We're doing an open call for encoders for evaluation for this project. Since I've got relationships with a number of vendors, I'm going to do all the testing, but Bob's going to pick the final results he likes best. I've talked to some encoder vendors directly, but wanted to make a public call to any companies who would like to submit their software. As an added bonus, I'll be doing an article for DV about the process of designing the media technologies for this project, including how we went about picking an encoder. So, here is the basic constraints we're looking for in the file, and I'm happy to look at any encoder that can meet these specs: Compatible with playback with ISMA Profile 1, QuickTime 6.0.2, Envivio plug-in, and MPEG4IP. Progressive download compatible file (MooV atom at the front of the file) Total data rate of 100-120 Kbps (2-pass VBR would be welcome). We'll be tweaking with the right resolution and data rate, but compression efficiency provided by the tool will be a major factor in our decision. Simple Visual video AAC-LC audio (should be able to do 22.050 mono at 20 Kbps) Source will be progressive PAL DV. It'll probably be in QuickTime format, but a transcode to .AVI might be possible. Anyone interested in providing an encoder for evaluation, drop me a private note at ben@benwaggoner.com Ben Waggoner Compressed Video Consulting, Training, and Encoding My Book: http://www.benwaggoner.com/books.htm Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html Compression Books: http://www.benwaggoner.com/bookshelf.htm From ben interframemedia.com Mon Nov 11 13:14:20 2002 From: ben interframemedia.com (Ben Waggoner) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] Call for MPEG-4 Encoders - Robert X. Cringely project Message-ID: Everybody, Okay, I've got a Big Deal to announce. Robert X. Cringely, computer industry journalist and broadcaster extraordinaire, is launching a new internet talk show, NerdTV, using MPEG-4 technology. Anyone familiar with Bob's past PBS series like "The Triumph of the Nerds" and his seminal book "Accidental Empires" knows this is a big deal. Here are some past columns from him describing this project, in chronological order: http://www.pbs.org/cringely/pulpit/pulpit20020321.html http://www.pbs.org/cringely/pulpit/pulpit20020328.html http://www.pbs.org/cringely/pulpit/pulpit20020912.html http://www.pbs.org/cringely/pulpit/pulpit20020704.html I've been consulting with him on the media delivery technologies part of the project, and we've decided that it's progressive download MPEG-4 all the way for the video part. Users will even be welcome to download and edit the camera sources, and edit together their own versions! All going well, this will be the first really big public use of MPEG-4, and while we picked the technology on its own merits, it's my hope that this can help jump start the industry some in the video-playing-on-computers market. The next phase of the project is to pick the right encoder, hence this email. We're doing an open call for encoders for evaluation for this project. Since I've got relationships with a number of vendors, I'm going to do all the testing, but Bob's going to pick the final results he likes best. I've talked to some encoder vendors directly, but wanted to make a public call to any companies who would like to submit their software. As an added bonus, I'll be doing an article for DV about the process of designing the media technologies for this project, including how we went about picking an encoder. So, here is the basic constraints we're looking for in the file, and I'm happy to look at any encoder that can meet these specs: Compatible with playback with ISMA Profile 1, QuickTime 6.0.2, Envivio plug-in, and MPEG4IP. Progressive download compatible file (MooV atom at the front of the file) Total data rate of 100-120 Kbps (2-pass VBR would be welcome). We'll be tweaking with the right resolution and data rate, but compression efficiency provided by the tool will be a major factor in our decision. Simple Visual video AAC-LC audio (should be able to do 22.050 mono at 20 Kbps) Source will be progressive PAL DV. It'll probably be in QuickTime format, but a transcode to .AVI might be possible. Anyone interested in providing an encoder for evaluation, drop me a private note at ben@benwaggoner.com Ben Waggoner Compressed Video Consulting, Training, and Encoding My Book: http://www.benwaggoner.com/books.htm Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html Compression Books: http://www.benwaggoner.com/bookshelf.htm From rob.koenen m4if.org Mon Nov 11 14:30:31 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] Call for MPEG-4 Encoders - Robert X. Cringely project In-Reply-To: Message-ID: <3C124172E7FDD511B510000347426D59011B4047@exchange.epr.com> Exciting project, Ben. Let me know if you can use M4IF's support. We'll advertise this on the website and in the weekly news digest. Through the Technotes and Discuss lists, you'll have reached the great majority of vendors; the ones that aren't there should get the news thought the news list. What is the target date for evaluation? Keep us posted, Rob ps: Bit of a shame you are keeping with Simple Visual Profile, because there is much promising ASP stuff out there. > -----Original Message----- > From: Ben Waggoner [mailto:ben@interframemedia.com] > Sent: Monday, November 11, 2023 13:07 > To: Streaming Media MPEG-4; technotes@lists.m4if.org; > discuss@lists.m4if.org > Subject: [M4IF Technotes] Call for MPEG-4 Encoders - Robert > X. Cringely project > > > Everybody, > > Okay, I've got a Big Deal to announce. Robert X. > Cringely, computer industry journalist and broadcaster > extraordinaire, is launching a new internet talk show, > NerdTV, using MPEG-4 technology. > > Anyone familiar with Bob's past PBS series like "The > Triumph of the Nerds" and his seminal book "Accidental > Empires" knows this is a big deal. Here are some past columns > from him describing this project, in chronological order: > > http://www.pbs.org/cringely/pulpit/pulpit20020321.html > http://www.pbs.org/cringely/pulpit/pulpit20020328.html > http://www.pbs.org/cringely/pulpit/pulpit20020912.html > http://www.pbs.org/cringely/pulpit/pulpit20020704.html > > I've been consulting with him on the media delivery > technologies part of the project, and we've decided that it's > progressive download MPEG-4 all the way for the video part. > Users will even be welcome to download and edit the camera > sources, and edit together their own versions! > > All going well, this will be the first really big public > use of MPEG-4, and while we picked the technology on its own > merits, it's my hope that this can help jump start the > industry some in the video-playing-on-computers market. > > The next phase of the project is to pick the right > encoder, hence this email. We're doing an open call for > encoders for evaluation for this project. Since I've got > relationships with a number of vendors, I'm going to do all > the testing, but Bob's going to pick the final results he > likes best. I've talked to some encoder vendors directly, > but wanted to make a public call to any companies who would > like to submit their software. As an added bonus, I'll be > doing an article for DV about the process of designing the > media technologies for this project, including how we went > about picking an encoder. > > So, here is the basic constraints we're looking for in > the file, and I'm happy to look at any encoder that can meet > these specs: > > Compatible with playback with ISMA Profile 1, QuickTime > 6.0.2, Envivio plug-in, and MPEG4IP. > > Progressive download compatible file (MooV atom at the front > of the file) > > Total data rate of 100-120 Kbps (2-pass VBR would be > welcome). We'll be tweaking with the right resolution and > data rate, but compression efficiency provided by the tool > will be a major factor in our decision. > > Simple Visual video > > AAC-LC audio (should be able to do 22.050 mono at 20 Kbps) > > Source will be progressive PAL DV. It'll probably be in > QuickTime format, but a transcode to .AVI might be possible. > > > Anyone interested in providing an encoder for evaluation, > drop me a private note at ben@benwaggoner.com > > > Ben Waggoner > Compressed Video Consulting, Training, and Encoding > > My Book: http://www.benwaggoner.com/books.htm > Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html > Compression Books: http://www.benwaggoner.com/bookshelf.htm > > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technot> es > From ben interframemedia.com Mon Nov 11 14:48:58 2002 From: ben interframemedia.com (Ben Waggoner) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] Call for MPEG-4 Encoders - Robert X. Cringely project In-Reply-To: <3C124172E7FDD511B510000347426D59011B4047@exchange.epr.com> Message-ID: on 11/11/02 14:30, Rob Koenen at rob.koenen@m4if.org wrote: > Exciting project, Ben. > > Let me know if you can use M4IF's support. We'll advertise this > on the website and in the weekly news digest. > > Through the Technotes and Discuss lists, you'll have reached the > great majority of vendors; the ones that aren't there should > get the news thought the news list. Excellent, thanks. > What is the target date for evaluation? I'll be starting it this week. We'll probably try to get it wrapped up in a few weeks, but don't have a hard deadline. Interested vendors should try to at least contact me this week, though. > Keep us posted, > Rob > > ps: Bit of a shame you are keeping with Simple Visual Profile, > because there is much promising ASP stuff out there. I fully agree! Alas, we need QuickTime 6 compatibility, which rules out ASP support at this point. As QuickTime's support improves, I'd like to look at migrating to ASP or H.264 when feasible. Ben Waggoner Compressed Video Consulting, Training, and Encoding My Book: http://www.benwaggoner.com/books.htm Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html Compression Books: http://www.benwaggoner.com/bookshelf.htm From ashok broadware.com Mon Nov 11 16:00:29 2002 From: ashok broadware.com (Ashok Yerneni) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] MPEG4-streaming format... In-Reply-To: Message-ID: <006b01c289de$83198cf0$38320a0a@earth> I am looking for a description of the stream syntax and having a hard time finding a good source. I bought the MPEG-4 Jump-Start and the MPEG-4 book by Fernando Pereira, but neither of them details the stream syntax! My main interest is audio+video streaming. Can someone point to a good description of the stream syntax? Thanks much, Ashok === Ashok Yerneni BroadWare Technologies ashok@broadware.com 408-342-2630 20863 Stevens Creek Blvd, #100 Cupertino, CA, 95014 From rob.koenen m4if.org Mon Nov 11 16:02:39 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] RE: MPEG4-streaming format... In-Reply-To: <006b01c289de$83198cf0$38320a0a@earth> Message-ID: <3C124172E7FDD511B510000347426D59011B4051@exchange.epr.com> The stream syntax? Hmmm... books explain, but do not copy the standard. A book cannot begin to copy all that stuff (even it it were allowed to). The bitstream syntax is specified in the MPEG-4 Standard. You can buy it from ISO, see references on M4IF's resources page (www.m4if.org/resources) The transport Syntax is in IETF specifications. Perhaps you would benefit from getting the ISMA (www.isma.tv) spec too, which details how to use IETF specs and MPEG-4 to do interoperable streaming. Hope it helps, Rob > -----Original Message----- > From: Ashok Yerneni [mailto:ashok@broadware.com] > Sent: Monday, November 11, 2023 16:00 > To: 'Ben Waggoner'; rob.koenen@m4if.org; 'Streaming Media > MPEG-4'; technotes@lists.m4if.org > Subject: MPEG4-streaming format... > > > I am looking for a description of the stream syntax and > having a hard time finding a good source. I bought the MPEG-4 > Jump-Start and the MPEG-4 book by Fernando Pereira, but > neither of them details the stream syntax! > > My main interest is audio+video streaming. Can someone point > to a good description of the stream syntax? > > Thanks much, > > Ashok > > === > Ashok Yerneni > BroadWare Technologies > ashok@broadware.com > 408-342-2630 > > 20863 Stevens Creek Blvd, #100 > Cupertino, CA, 95014 > > > From lcheng62 yahoo.com Mon Nov 11 21:37:45 2002 From: lcheng62 yahoo.com (Liang Cheng) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] B-frame in simple profile Message-ID: <20021112053745.18095.qmail@web21108.mail.yahoo.com> Hi, Can anyone of you tell me why there is no B-frame in mpeg4 simple profile? How long is the mp4 suggested maximum refresh period. H.263 takes about 132 frames before the next I-frame. Thanks, Liang __________________________________________________ Do you Yahoo!? U2 on LAUNCH - Exclusive greatest hits videos http://launch.yahoo.com/u2 From chl math.uni-bonn.de Tue Nov 12 10:35:29 2002 From: chl math.uni-bonn.de (Christoph Lampert) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] B-frame in simple profile In-Reply-To: <20021112053745.18095.qmail@web21108.mail.yahoo.com> Message-ID: On Mon, 11 Nov 2002, Liang Cheng wrote: > Hi, > Can anyone of you tell me why there is no B-frame in > mpeg4 simple profile? How long is the mp4 suggested > maximum refresh period. H.263 takes about 132 frames > before the next I-frame. No _macroblock_ is allowed to be coded using a predicted mode (inter,inter4v,gmc,...) more than 131 times in a row. Christoph -- Christoph H. Lampert chl@math,uni-bonn,de | Diese Signature wurde maschi- Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell erstellt und bedarf Tel. (0228) 73-4708 Fax. +49 228 73-7916 | keiner Unterschrift. AZ 27B-6 From tma iis.fhg.de Tue Nov 12 10:56:32 2002 From: tma iis.fhg.de (Herbert Thoma) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] B-frame in simple profile References: <20021112053745.18095.qmail@web21108.mail.yahoo.com> Message-ID: <3DD0D050.BD9F13F9@iis.fhg.de> Liang Cheng schrieb: > > Hi, > Can anyone of you tell me why there is no B-frame in > mpeg4 simple profile? May be, because B-frames are not simple enough? :-) > How long is the mp4 suggested > maximum refresh period. I am not aware of a suggested maximum refresh period in MPEG-4 Video. > H.263 takes about 132 frames > before the next I-frame. H.263 does not require an I-frame every 132 frames. H.263 reqires that every macroblock is coded as INTRA macroblock at least every 132 frames. You don't need to send any I-frame after the first frame. Best Regards, Herbert. > Thanks, > Liang > > __________________________________________________ > Do you Yahoo!? > U2 on LAUNCH - Exclusive greatest hits videos > http://launch.yahoo.com/u2 > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes -- Herbert Thoma FhG-IIS A, Studio Department 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 mithun.nayak patni.com Tue Nov 12 15:25:07 2002 From: mithun.nayak patni.com (Mithun N) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] total_data_number Message-ID: <001601c28a31$953982c0$fe02a8c0@patni.com> Hi, Annex L in 14496-2 standard uses total_data_number parameter for Rate Control. The standard however, does not explain the semantics of the term. Does anyone know what does it indicate? Thanks & Regards, Mithun Nayak From fp lx.it.pt Tue Nov 12 10:02:26 2002 From: fp lx.it.pt (Fernando Pereira) Date: Wed Jul 30 14:09:11 2003 Subject: [M4IF Technotes] MPEG4-streaming format... References: <006b01c289de$83198cf0$38320a0a@earth> Message-ID: <3DD0D1B2.50405@lx.it.pt> Dear Ashok Ashok Yerneni wrote: >I am looking for a description of the stream syntax and having a hard >time finding a good source. I bought the MPEG-4 Jump-Start and the >MPEG-4 book by Fernando Pereira, but neither of them details the stream >syntax! > >My main interest is audio+video streaming. Can someone point to a good >description of the stream syntax? > > I am sorry to disappoint you but there a single solution for you: THE STANDARD ITSELF. I believe no book will ever include the full bitstream syntax becaus it is long and boring ;-) REgards Fernando Pereira -- Fernando Manuel Bernardo Pereira, Ph.D., Professor Instituto Superior T?cnico - Instituto de Telecomunica??es Av. Rovisco Pais, 1049-001 Lisboa, PORTUGAL Phone: + 351 21 8418460 Fax: + 351 21 8418472 E-mail: Fernando.Pereira@lx.it.pt WWW: http://www.img.lx.it.pt/~fp/ From moses kampsax.dtu.dk Tue Nov 12 12:24:58 2002 From: moses kampsax.dtu.dk (moses wold) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] error resiilience algorithms document in the MPEG4 reference Message-ID: hi all I'm looking for the error resilience algorithms documents, which are used in the MoMuSys, MPEG4IP and microsoft. I need the algorithms. I'm looking forward to get yourhelp. Best regards Mosese <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> YOUR BRAIN IS HUNGRY Wireless Networks Technology, News and Opinion. <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> From lmwang hf.tvia.com Tue Nov 12 19:44:11 2002 From: lmwang hf.tvia.com (Limin Wang) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] =?gb2312?B?tPC4tDogW000SUYgVGVjaG5vdGVzXSBlcnJvciByZXNpaWxpZW5jZQ==?= =?gb2312?B?IGFsZ29yaXRobXMgZG9jdW1lbnQgaW4gdGhlIE1QRUc0IHJlZmVyZW5jZQ==?= In-Reply-To: Message-ID: Dear Mosese, You can refer to SC29/WG11/N3908(VM18.0) document. Section 3.6 is talking about Error Resilience. Hopes that helps, Best, Limin -----ÓʼþÔ­¼þ----- ·¢¼þÈË: technotes-admin@lists.m4if.org [mailto:technotes-admin@lists.m4if.org] ´ú±í moses wold ·¢ËÍʱ¼ä: Tuesday, November 12, 2023 7:25 PM ÊÕ¼þÈË: Technotes@lists.m4if.org Ö÷Ìâ: [M4IF Technotes] error resiilience algorithms document in the MPEG4 reference hi all I'm looking for the error resilience algorithms documents, which are used in the MoMuSys, MPEG4IP and microsoft. I need the algorithms. I'm looking forward to get yourhelp. Best regards Mosese <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> YOUR BRAIN IS HUNGRY Wireless Networks Technology, News and Opinion. <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes From rob.koenen m4if.org Tue Nov 12 10:06:40 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] error resiilience algorithms document in the MPEG4 reference In-Reply-To: Message-ID: <3C124172E7FDD511B510000347426D59011B405F@exchange.epr.com> There are no standard algorithms to deal with errors in MPEG-4 There is, however, a lot of literature on how to conceal and cope with errors. The MPEG-4 standard contains some tools that can help you cope with errors, but does not prescribe how to do that. The same applies to the (MoMuSys, Microsoft) reference software. I suggest that you do a literature search. There are plenty of references out there, including articles on how to make MPEG-4 error-resilient. Best, Rob > -----Original Message----- > From: moses wold [mailto:moses@kampsax.dtu.dk] > Sent: Tuesday, November 12, 2023 3:25 > To: Technotes@lists.m4if.org > Subject: [M4IF Technotes] error resiilience algorithms > document in the MPEG4 reference > > > hi all > > > I'm looking for the error resilience algorithms documents, > which are used in the MoMuSys, MPEG4IP and microsoft. > > I need the algorithms. > > I'm looking forward to get yourhelp. > > Best regards > Mosese > <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> > YOUR BRAIN IS HUNGRY > Wireless Networks Technology, News and Opinion. > <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technot> es > From sunx pollux.usc.edu Tue Nov 12 14:26:25 2002 From: sunx pollux.usc.edu (sunx) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] FGST Message-ID: Hi All, I have a question regarding FGST. I checked the Microsoft reference software for FGS. I think it has the complete implementation for FGS, but I doubt if FGST is completely implemented. I am working on 4711 and 3905. Any suggestions will be highly appreciated! BTW, I think the implementation of 4711 also has some problems. I am using Windows XP Home Edition and VC 6.0 + SP 5. Xiaoming From dwu jdl.ac.cn Wed Nov 13 08:37:46 2002 From: dwu jdl.ac.cn (David) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] Slice Resynchronization Message-ID: <002a01c28aac$e2bd25d0$222ae29f@sky> Hi, all, Slice Resynchronization is one of visual tools for error resilience. How to realize the function in MPEG-4 encoder? What corresponds to it the visual bitstream? Thanks, david -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021113/ee91c606/attachment.html From tma iis.fhg.de Wed Nov 13 11:49:12 2002 From: tma iis.fhg.de (Herbert Thoma) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] Slice Resynchronization References: <002a01c28aac$e2bd25d0$222ae29f@sky> Message-ID: <3DD22E28.7A5B05EA@iis.fhg.de> Hi, > David schrieb: > > Hi, all, > > Slice Resynchronization is one of visual tools for error resilience. How to realize the function in MPEG-4 encoder? What corresponds to it > the visual bitstream? that's video_packet_header, syntax on page 46 of ISO/IEC 14496-2:2001 Regards, Herbert. > Thanks, > > david -- Herbert Thoma FhG-IIS A, Studio Department 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 gertzen irt.de Wed Nov 13 13:22:30 2002 From: gertzen irt.de (Gertzen) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] SL packet header syntax for MPEG-4:1 in MPEG-2 PES Message-ID: Hi, does anybody know, what elements of the SL packet header syntax should be used, when a visual elementary stream of a MPEG-4 scene ougth be to integrate in a MPEG-2 packetized elementary stream (PES)? Thanks for hints Carsten From rob.koenen m4if.org Wed Nov 13 22:43:41 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] RE: [M4IF Discuss] IS MPEg4 simple profile doesn't contain shape coding In-Reply-To: <20021113060703.82764.qmail@web11608.mail.yahoo.com> Message-ID: <3C124172E7FDD511B510000347426D59011B40BE@exchange.epr.com> > I have a question MPEG4 simple profile does not > contain any Shape coding block? Indeed Simple Profile (and Advanced Simple, for that matter) does not contain shape coding. Rob From Manu.Batura patni.com Thu Nov 14 15:44:24 2002 From: Manu.Batura patni.com (Manu V Batura) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] Technical paper? Message-ID: <000201c28bc6$9c2bb210$5402a8c0@pcp33246> Hi, Is an public domain copy available for the following paper or its latest version (preferably): Video Group, "Text of ISO/IEC 14496-2 MPEG4 video VM--version 8.0", ISO/IEC JTC1/SC29/WG11 Coding of Moving Pictures and Associated Audio MPEG 97/W1796, Stockholm, Sweden, 1997. Thanks & Regards, Manu From Gus.Reid motorola.com Thu Nov 14 14:13:12 2002 From: Gus.Reid motorola.com (Reid Angus-BAR04Z) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] SSP in Momusys Message-ID: Hi, I'm having a lot of trouble getting the MoMuSys-1.0-001220_sony encoder to generate Simple Scalable Profile bitstreams. Are there any sample ctl/cfg files available that will work? -Gus From raghu vidiator.com Thu Nov 14 17:21:11 2002 From: raghu vidiator.com (Raaghu Nagaraj) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] RE: H26L for PC platforms Message-ID: <30A6344AEBFC85429C30FF6C9C501CD10EE49E@bays02.mediator.com> Hello all, Are there any license-able H.26L implementations for the PC (x86) platform that you are aware of? Does M4IF maintain any such information on their website? Best regards, Raghu From rob.koenen m4if.org Thu Nov 14 21:18:49 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] RE: H26L for PC platforms In-Reply-To: <30A6344AEBFC85429C30FF6C9C501CD10EE49E@bays02.mediator.com> Message-ID: <3C124172E7FDD511B510000347426D59011B40F7@exchange.epr.com> People that have stuff to license can now get in touch with you :-) We do not have such information on our website (yet). We will start doing interop tests for MPEG-4 Advanced Video Coding / H.264 early next year, though. You are welcome to join those tests. Best, Rob > -----Original Message----- > From: Raaghu Nagaraj [mailto:raghu@vidiator.com] > Sent: Thursday, November 14, 2023 17:21 > To: technotes@lists.m4if.org > Subject: [M4IF Technotes] RE: H26L for PC platforms > > > > Hello all, > > Are there any license-able H.26L implementations for the PC > (x86) platform that you are aware of? > > Does M4IF maintain any such information on their website? > > Best regards, > Raghu > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technot> es > From garysull microsoft.com Thu Nov 14 21:56:32 2002 From: garysull microsoft.com (Gary Sullivan) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] RE: H26L for PC platforms Message-ID: <3B8749EA44FC2748956538D4A3096E42E08C97@RED-MSG-01.redmond.corp.microsoft.com> UB Video, HHI, and Videolocus seem to be doing the most in terms of real-time implementation demos of H.264/AVC so far. Others are also working on imlementations. But the standard is not finished yet and no bitstream exchanges have happened yet. So there's no way anyone could claim a complete conforming implementation yet. Best Regards, -Gary S. +> -----Original Message----- +> From: Raaghu Nagaraj [mailto:raghu@vidiator.com] +> Sent: Thursday, November 14, 2023 5:21 PM +> To: technotes@lists.m4if.org +> Subject: [M4IF Technotes] RE: H26L for PC platforms +> +> +> +> Hello all, +> +> Are there any license-able H.26L implementations for the PC +> (x86) platform that you are aware of? +> +> Does M4IF maintain any such information on their website? +> +> Best regards, +> Raghu +> _______________________________________________ +> Technotes mailing list +> Technotes@lists.m4if.org +> http://lists.m4if.org/mailman/listinfo/technotes +> From hwang jdl.ac.cn Fri Nov 15 19:53:06 2002 From: hwang jdl.ac.cn (WangHui) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] about end of the RVLC Message-ID: <013b01c28c9d$8f5ee430$242ae29f@WangHui> hi, all. about the RVLC in MPEG-4, when decode backward, how to recognize the end of the RVLC codeword? For the resync_marker is bytealigned, so if the RVLC is transferred by byte or word, there may be some stuffing bits between the end of RVLC and resync_marker, in this case, how to identify the end of the RVLC segment? best regards WangHui -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021115/7ac06816/attachment.html From christian.weigel epost.de Fri Nov 15 18:11:06 2002 From: christian.weigel epost.de (Christian Weigel) Date: Wed Jul 30 14:09:12 2003 Subject: [M4IF Technotes] MediaControl node problem Message-ID: <000201c28cc9$fc1ded10$0500a8c0@christianpc> Hello! I have a problem using the MediaControl node in a scene. I tried the following: - I extracted a mediaStream (video) and the corresponding *.nhnt *.info from a video only stream with mp4tool from ENST (the video was a sample from mpegable: MPEG-4 video 360x288, 600 kBit/s) - I included it in my scene via the MovieTexture node and created a new mp4 file - again with mp4tool - I loaded the file with the latest envivoTV It works and I am able to watch the video BUT I can't control it's playback. I added a MediaControl node to the scene with mediaStartTime -1 (the default) If I deactivate the mediaStream by setting the stopTime field of the MovieTexture to NOW it stops. But if I reactivate it (by setting the startTime field to NOW) it plays not from the time it was stopped (like it is said in the standard when the MediaControlNode.startMediaTime is -1) but so as if it has never been stopped (it jumps forward) The strange thing is, that there are scenes where exactly the same works perfectly (ENST conformanace Bitstreams: Advanced2D_Level1\Envivio\mediacontrol) So I guess the mediaStartTime and mediaStopTime fields of the MediaControl node are ignored. It is said in the standard that this should be the case if the mediaTime of the media stream is undefinded. Maybe this is my problem? I found out, that after extraction and remuxing of the stream its properties differ a little,e.g. mp4info says with the original stream: video MPEG-4 18.8800 secs, 0kbps, 352x240 @ 12.55 fps and after mp4tool demux/remux video MPEG-4 18.0000 secs, 561kbps, 0x0 @ 13.17 fps Maybe some time information is lost during this process. Does someone has an idea or had same problems? Regards Christian Weigel From garysull microsoft.com Fri Nov 15 12:40:11 2002 From: garysull microsoft.com (Gary Sullivan) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] about end of the RVLC Message-ID: <3B8749EA44FC2748956538D4A3096E42E08CA7@RED-MSG-01.redmond.corp.microsoft.com> Skipped content of type multipart/alternative From jean-claude.dufourd enst.fr Fri Nov 15 22:17:26 2002 From: jean-claude.dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] MediaControl node problem References: <000201c28cc9$fc1ded10$0500a8c0@christianpc> Message-ID: <3DD56466.8020303@enst.fr> Christian Weigel wrote: >The strange thing is, that there are scenes where exactly the same works >perfectly (ENST conformance Bitstreams: >Advanced2D_Level1\Envivio\mediacontrol) > Precision: the files are *downloadable* from ENST web site, but this file comes, as the directory name says, from Envivio, the creator of EnvivioTV. >If I deactivate the mediaStream by setting the stopTime field of the >MovieTexture to NOW it stops. But if I reactivate it (by setting the >startTime field to NOW) it plays not from the time it was stopped (like >it is said in the standard when the MediaControlNode.startMediaTime is >-1) but so as if it has never been stopped (it jumps forward) > This is as it should be. The standard says exactly this. When a MediaControl is present and active, then the MovieTexture fields are just doing mute and demute. I have had problems with the EnvivioTV MediaControl implementation, but it comes from their interpretation of the spec, that you need to mute the video for a change in MediaControl fields to have an effect. >I found out, that after extraction and >remuxing of the stream its properties differ a little,e.g. mp4info says >with the original stream: > mp4tool had bugs: rounding of the duration, losing the width and height BUT it computes the bitrate exactly! If the video plays without MediaControl, then there is no consequence. Best regards JC From rob.koenen m4if.org Fri Nov 15 18:26:46 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] Technical paper? In-Reply-To: <000201c28bc6$9c2bb210$5402a8c0@pcp33246> Message-ID: <3C124172E7FDD511B510000347426D59011B4111@exchange.epr.com> This is now a part of the Reference Software, which can be downloaded from ISO's website. See www.m4if.org/resources.php Rob > -----Original Message----- > From: Manu V Batura [mailto:Manu.Batura@patni.com] > Sent: Thursday, November 14, 2023 2:14 > To: technotes@lists.m4if.org > Subject: [M4IF Technotes] Technical paper? > > > Hi, > > Is an public domain copy available for the following paper or > its latest version (preferably): > > Video Group, "Text of ISO/IEC 14496-2 MPEG4 video VM--version > 8.0", ISO/IEC JTC1/SC29/WG11 Coding of Moving Pictures and > Associated Audio MPEG 97/W1796, Stockholm, Sweden, 1997. > > Thanks & Regards, > Manu > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technot> es > From hwang jdl.ac.cn Sat Nov 16 12:40:23 2002 From: hwang jdl.ac.cn (WangHui) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] about end of the RVLC References: <3B8749EA44FC2748956538D4A3096E42E08CA7@RED-MSG-01.redmond.corp.microsoft.com> Message-ID: <006e01c28d2a$46e4fc70$242ae29f@WangHui> Message Thank you for your email. For the stuffing bits, I think it is to make the resync_marker bytealigned, and avoid emulation of resync_marker. So, it takes the form just reverse to that of resync_marker. But for the RVLC, the end of the RVLC is not written directly, just like the EOB in MPEG-2. You must decode the RVLC sequence to get the "LAST". I think the method you described will not work, for example, given the RVLC codeword "0,0111,1111,1101s" is just bytealigned, if the 's' bit is '1', according to your method the Last three bits '011' will be stuffing bits, which will result in wrong RVLC codeword error. ----- Original Message ----- From: Gary Sullivan To: WangHui ; technotes@lists.m4if.org Sent: Saturday, November 16, 2023 4:40 AM Subject: RE: [M4IF Technotes] about end of the RVLC The stuffing bits always consist of one bit equal to zero followed by 7 bits or less equal to 1. So I think you can just look back from the end while removing everything up to and including the last bit equal to zero -- the rest will be valid data. Best Regards, -Gary S. -----Original Message----- From: WangHui [mailto:hwang@jdl.ac.cn] Sent: Friday, November 15, 2023 3:53 AM To: technotes@lists.m4if.org Subject: [M4IF Technotes] about end of the RVLC hi, all. about the RVLC in MPEG-4, when decode backward, how to recognize the end of the RVLC codeword? For the resync_marker is bytealigned, so if the RVLC is transferred by byte or word, there may be some stuffing bits between the end of RVLC and resync_marker, in this case, how to identify the end of the RVLC segment? best regards WangHui -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021116/df8507e4/attachment.html From garysull microsoft.com Sat Nov 16 07:11:21 2002 From: garysull microsoft.com (Gary Sullivan) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] about end of the RVLC Message-ID: <3B8749EA44FC2748956538D4A3096E42E08CC6@RED-MSG-01.redmond.corp.microsoft.com> Skipped content of type multipart/alternative From christian.weigel epost.de Sat Nov 16 17:00:33 2002 From: christian.weigel epost.de (Christian Weigel) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] MediaControl node problem In-Reply-To: <3DD56466.8020303@enst.fr> Message-ID: <000401c28d89$4b74d8a0$0500a8c0@christianpc> Jean-Claude Duford wrote: > Christian Weigel wrote: > >If I deactivate the mediaStream by setting the stopTime field of the > >MovieTexture to NOW it stops. But if I reactivate it (by setting the > >startTime field to NOW) it plays not from the time it was > >stopped (like it is said in the standard when the > >MediaControlNode.startMediaTime is > >-1) but so as if it has never been stopped (it jumps forward) > > > This is as it should be. The standard says exactly this. When a > MediaControl is present and active, then the MovieTexture fields are > just doing mute and demute. Thank you for the answer. I assume that mute and demute is the same as deactivation and activation of the MovieTexture. If so I am a little confused because in the standard in section 9.4.2.75.2 is written: "When the controlled stream becomes inactive and then active again, then if mediaStartTime is -1 the stream starts playing from the point where it was last stopped. Otherwise the first controlled stream object in the url field restarts playing from mediaStartTime." So if the MediaControl's is mediaStartTime field is set to -1 (which it is by default) it should be possible to PAUSE the video by muting (deactivate) it and to CONTINUE it at the point in media Time where it was paused by demuting it. Or do I unterstand something the wrong way. I looked in the envivio sample how they have done it and they did it exactly like this and it works. So I still don't know why my scene is not working. Maybe someone from envivio can help? I put my sample scene on a web sever together with the scene description. Here's the url: http://www.tu-ilmenau.de/~cweigel/mpeg4/media_control.zip The working envivio sample is downloadable from ENST: http://www.comelec.enst.fr/~dufourd/mpeg-4/adv2Dl1.zip (inside the zip use path: \Advanced2D_Level1\Envivio\mediacontrol) Regards Christian Weigel From jean-claude.dufourd enst.fr Sat Nov 16 18:18:06 2002 From: jean-claude.dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] MediaControl node problem References: <000401c28d89$4b74d8a0$0500a8c0@christianpc> Message-ID: <3DD67DCE.6050004@enst.fr> Christian Weigel wrote: >I assume that mute and demute is the same as deactivation and activation >of the MovieTexture. > Yes > If so I am a little confused because in the >standard in section 9.4.2.75.2 is written: > >"When the controlled stream becomes inactive and then active again, then >if mediaStartTime is -1 the stream starts playing from the point where >it was last stopped. Otherwise the first controlled stream object in the >url field restarts playing from mediaStartTime." > What I remember from the discussions around MediaControl is, when MediaControl is present, the MovieTexture field manipulation has NO consequence on the stream at all. This means the formulation above is unfortunate. I would welcome confirmation of my statement by people more knowledgeable about MediaControl (Zvi, Yuval, Carsten, Jean, Cyril...) >So if the MediaControl's is mediaStartTime field is set to -1 (which it >is by default) it should be possible to PAUSE the video by muting >(deactivate) it and to CONTINUE it at the point in media Time where it >was paused by demuting it. Or do I unterstand something the wrong way. > >I looked in the envivio sample how they have done it and they did it >exactly like this and it works. > No, they did not do it exactly like that. Here is the relevant script node from this file (which I know almost by heart now from trying to understand the way Envivio implemented MediaControl): DEF N40 Script { url "javascript:function setTime(value, timestamp) { if (value) { video1.stopTime = timestamp; control.mediaStartTime = ( what.translation.x + 40 ) / 8.8; } else video1.startTime = timestamp; } [snip] " eventIn SFVec2f setVolume eventIn SFBool setTime eventOut SFFloat transpar1 field SFNode video1 USE N29 field SFNode what USE N17 field SFNode control USE N33 } N29 is the MovieTexture N33 is the MediaControl They are doing: - on isActive=true (mouse button down), N29.stopTime=now and N33.mediaStartTime=something - on isActive=false (mouse button up), N29.startTime=now (which is necessarily another "now" than above) If you do not do it this way, the MediaControl ignores your changes, which is wrong in my interpretation of the standard, but this is moot. I could not make my scenes work, so I'll stop here. I cannot make constructive suggestions. I tried emulating the above with Conditionals, to no avail. We are working on implementing MediaControl in Osmose: it is supposed to work as of last Thrusday. I have yet to test it. Best regards JC From pantaloonmc hotmail.com Fri Nov 15 16:03:30 2002 From: pantaloonmc hotmail.com (Padraig Murphy) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] SL packets Message-ID: Hello, I am trying to write a program that puts an elementary stream through the sync layer (i.e. into an SL-packetized stream). I want to use the simplest method possible, as the main reason I am doing it is so that I can use FlexMux. If I use the null SL packet header, am I still able to use Flexmux?? and what is the best way of writing code for this?? thanks, Padraig _________________________________________________________________ Help STOP SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail From guraaf yahoo.co.in Mon Nov 18 15:33:00 2002 From: guraaf yahoo.co.in (=?iso-8859-1?q?Gaurav=20Aggarwal?=) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] deblocking/deringing usage model Message-ID: <20021118153300.28056.qmail@web8202.mail.in.yahoo.com> hi all, would appreciate if someone could talk about deblocking/deringing filters. are decoders using them for SP or ASP? does anybody implement them in silicon as yet? what is their promised benefits? at what kind of bitrates would one really need these filters? our gut feeling is that at broadcast quality video (over satellite/cable) at 5Mbps and higher, we could do without them. does that sound reasonable? any views or pointers to technical papers would be appreciated. thanks, gaurav ________________________________________________________________________ Missed your favourite TV serial last night? Try the new, Yahoo! TV. visit http://in.tv.yahoo.com From ben interframemedia.com Mon Nov 18 10:40:50 2002 From: ben interframemedia.com (Ben Waggoner) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] Practical differences between MPEG-4 and H.263 quantization Message-ID: Folks, Say, any thoughts on the practical differences between H.263 and MPEG-4 quantization? In which areas would you expect one or the other to have an advantage in output quality? Ben Waggoner Compressed Video Consulting, Training, and Encoding My Book: http://www.benwaggoner.com/books.htm Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html Compression Books: http://www.benwaggoner.com/bookshelf.htm From Cyril.Concolato enst.fr Tue Nov 19 14:10:26 2002 From: Cyril.Concolato enst.fr (Cyril Concolato) Date: Wed Jul 30 14:09:13 2003 Subject: [M4IF Technotes] MediaControl node problem References: <000401c28d89$4b74d8a0$0500a8c0@christianpc> <3DD67DCE.6050004@enst.fr> Message-ID: <3DDA3842.EAF095DB@enst.fr> Hi all, > >"When the controlled stream becomes inactive and then active again, then > >if mediaStartTime is -1 the stream starts playing from the point where > >it was last stopped. Otherwise the first controlled stream object in the > >url field restarts playing from mediaStartTime." > What I remember from the discussions around MediaControl is, when > MediaControl is present, the MovieTexture field manipulation has NO > consequence on the stream at all. This means the formulation above is > unfortunate. I would welcome confirmation of my statement by people > more knowledgeable about MediaControl (Zvi, Yuval, Carsten, Jean, Cyril...) I think this interpretation is wrong because it explicitely says in the spec that one way to pause a stream is to set stopTime to Now and mediaStartTime to -1. One interpretation of the spec is that if you don't route Now to stopTime then you cannot change mediaStartTime because the media is active. > >So if the MediaControl's is mediaStartTime field is set to -1 (which it > >is by default) it should be possible to PAUSE the video by muting > >(deactivate) it and to CONTINUE it at the point in media Time where it > >was paused by demuting it. Or do I unterstand something the wrong way. I think that Christian's example is right and the fact that it does not work is probably a bug in ETV. Try to ROUTE -1 to the mediaStartTime field of the MediaControl node, instead of using the static default value, it should work. BR, Cyril Concolato -- Ecole Nationale Sup?rieure des T?l?communications, Paris Dept. Comelec 46, rue Barrault 75013 Paris Tel: +33145817991 Fax: +33145804036 From gertzen irt.de Wed Nov 20 12:50:07 2002 From: gertzen irt.de (Gertzen) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] SL packet header syntax for MPEG-4:1 in MPEG-2 PES Message-ID: Thanks Dominique, I suppose that all Elements needed by a MPEG-4 SL-Header can be taken over by an extended MPEG-2 PES header. This is similar to the RTP header when MPEG-4 is streamed over IP. Conclusion is, that the SL-Header exist, out of the view from MPEG-4 Systems. But its bit-length in the PES-Payload is zero. So an individual MPEG-4 elementary stream, packed as described in ISO/IEC 13818-1:2000, can be part of a MPEG-4 scene without physical modifications. The organisation is part of a DMIF instance to be develop for MPEG-2 TS. Am I right? Is anybody working on a implementation of a DMIF instance for MPEG-2 TS? I already modify a software-based DVB-Multiplexer and Demultiplexer to carry individual MPEG-4 elementary streams over a DVB kompliant network. But currently I can't demonstrate a live stream based on DVB datastreaming. Regards Carsten Am 20.11.2023 09:37:34, schrieb "CURET Dominique FTRD/DMI/REN" : >dear Carsten, > >in a PES you can start with : > PES_priority > PTS_DTS flags & Timestamps (all needed when you have B frames) > ESCR flag & field > ES_rate flag & field > >I think that an SL header can have all the corresponding fields (CTS=PTS) > >the copyright, original_or_copy, data_alignment_indicator can be set or not. > >About PES_scrambling_control > DSM_trick_modes > additional_copy_info > PES_CRC > PES_extension > PES_header_data_length > this needs more reflexion from my side. > >Dominique Curet >FT R&D Rennes, Laboratoire DMI/PIA >Tel : (33) 2 99 12 40 05 >Fax : (33) 2 99 12 40 98 >E-mail : dominique.curet@rd.francetelecom.fr >Mail : FT R & D , DMI/PIA >4, rue du clos Courtel BP 91226 >35512 Cesson-Sévigné Cedex, FRANCE > > >-----Message d'origine----- >De : AVARO Olivier FTRD/DIH/REN >Envoyé : mercredi 20 novembre 2002 08:57 >À : CURET Dominique FTRD/DMI/REN >Objet : TR: [M4IF Technotes] SL packet header syntax for MPEG-4:1 in >MPEG-2 PES > > > >Same ... >-----Message d'origine----- >De : Gertzen [mailto:gertzen@irt.de] >Envoyé : mercredi 13 novembre 2002 13:23 >À : Technotes@lists.m4if.org >Objet : [M4IF Technotes] SL packet header syntax for MPEG-4:1 in MPEG-2 >PES > > >Hi, >does anybody know, what elements of the SL packet header syntax should be used, when >a visual elementary stream of a MPEG-4 scene ougth be to integrate in a MPEG-2 >packetized elementary stream (PES)? > >Thanks for hints >Carsten > > >_______________________________________________ >Technotes mailing list >Technotes@lists.m4if.org >http://lists.m4if.org/mailman/listinfo/technotes > > -------- Ende der weitergeleiteten Nachricht -------- From christian.weigel epost.de Wed Nov 20 14:25:56 2002 From: christian.weigel epost.de (Christian Weigel) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] MediaControl node problem In-Reply-To: <3DDA3842.EAF095DB@enst.fr> Message-ID: <000a01c29098$5ec08820$0500a8c0@christianpc> Hello! Cyril Concolato wrote: > I think that Christian's example is right and the fact that > it does not work is probably a bug in ETV. Try to ROUTE -1 to > the mediaStartTime field of the MediaControl node, instead of > using the static default value, it should work. I just tried it by routing the TouchSensor.isActive to a Valuator with factor 0 and offset -1 and the output of this valuator to the mediaStartTime. It's value changes but the result is the same - the video doesn't PAUSE correctly. Then I decided to make a test with the envivo content. I extracted every needed stream of the mp4 file with mp4tool (scene, video, audio ...) and and muxed together again to an mp4 without a change. And now it comes: the scene lookes exactly the same BUT is has the SAME Error as my scenes right now. If you press the PAUSE button - the video "is going on in the background". The same scene does different things. What could be the reason for this?? Maybe it's a bug in mp4tool. What did envivio different by creating the scene? I found that the scene time of the envivio sample scene is 20min whereas the time of "my" envivio scene was just as long as the video. Could there be some problem with the time bases resulting in these problems? I will post this problem and the link to the thread in the envivio support forum - maybe they know an answer. Regards Christian Weigel From renaud envivio.com Wed Nov 20 15:07:37 2002 From: renaud envivio.com (renaud) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] MediaControl node problem References: <000a01c29098$5ec08820$0500a8c0@christianpc> Message-ID: <3DDB9729.84557020@envivio.com> Hi, > Maybe it's a bug in mp4tool. What did envivio different by creating the > scene? I found that the scene time of the envivio sample scene is 20min > whereas the time of "my" envivio scene was just as long as the video. > Could there be some problem with the time bases resulting in these > problems? Could be, indeed. If you want to "pause" a stream, you must be sure that it is *not* linked to the same OCR as the BIFS scene (otherwize you would pause the scene itself!) At least this could be an explanantion why the envivio content does not show the same durations for the scene and for the video (despite it could be the same but that makes more sense to have a longuer scene duration in this case) regards Renaud Cazoulat Envivio From michael.ditze c-lab.de Wed Nov 20 16:27:00 2002 From: michael.ditze c-lab.de (Michael Ditze) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] MPEG-4 Reference Implementation Message-ID: <000001c290a9$451bd580$a751ea83@clab.de> Hi, I am wondering whether there is already a full reference implementation of the MPEG-4 framework available? Is there any implementation that implements DMIF for example? Furthermore, is there are list with available MPEG-4 codecs and the profiles they encode? And another newbie question: We would like to extend the DMIF in MPEG-4 by implementing a protocol that allows to do video streaming over the CAN bus though IP. I nthink this should be possible. But how to implement without any reference implementation? Any answers? Thanks, Michael Michael Ditze C-LAB, 33094 Paderborn, Germany F?rstenallee 11, Office FU 304 Voice: +49-5251-60-6124, Fax: +49-5251-60-6065 Email: mailto:michael.ditze@c-lab.de From Guido.Franceschini TILAB.COM Wed Nov 20 18:50:42 2002 From: Guido.Franceschini TILAB.COM (Franceschini Guido) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] MPEG-4 Reference Implementation Message-ID: <3737D9839ED3D3408C73611BDA907A0406EFDF@EXC2K05A.cselt.it> Michael, the IM1 reference software fits the DMIF architecture by means of a DMIF filter and a few examples of DMIF Instances: - FileMP4Instance -for reading MP4 files: builds on top of the MP4library from Apple and has been originally provided with my contribution; Jean Le Feuvre is mantaining it right now - FileTrifInstance -for reading TRIF files (a format used in old IM1 times): provided and mantained by Zvi - RemoteInstance -to contact a remote server using DMIF signaling: provided by Kambiz Asrar Haghighi and Yasser Pourmohammadi (University of British Columbia) Is this what you are looking for? Guido Franceschini TILAB - Multimedia Division Via G.Reiss Romoli 274 I-10148 Torino, Italy tel + 39 011 228 6137 fax + 39 011 228 6299 -----Original Message----- From: Michael Ditze [mailto:michael.ditze@c-lab.de] Sent: Wednesday, November 20, 2023 4:27 PM To: technotes@lists.m4if.org Subject: [M4IF Technotes] MPEG-4 Reference Implementation Hi, I am wondering whether there is already a full reference implementation of the MPEG-4 framework available? Is there any implementation that implements DMIF for example? Furthermore, is there are list with available MPEG-4 codecs and the profiles they encode? And another newbie question: We would like to extend the DMIF in MPEG-4 by implementing a protocol that allows to do video streaming over the CAN bus though IP. I nthink this should be possible. But how to implement without any reference implementation? Any answers? Thanks, Michael Michael Ditze C-LAB, 33094 Paderborn, Germany F?rstenallee 11, Office FU 304 Voice: +49-5251-60-6124, Fax: +49-5251-60-6065 Email: mailto:michael.ditze@c-lab.de _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes ==================================================================== CONFIDENTIALITY NOTICE This message and its attachments are addressed solely to the persons above and may contain confidential information. If you have received the message in error, be informed that any use of the content hereof is prohibited. Please return it immediately to the sender and delete the message. Should you have any questions, please contact us by replying to MailAdmin@tilab.com. Thank you ==================================================================== From christian.weigel epost.de Wed Nov 20 22:43:07 2002 From: christian.weigel epost.de (Christian Weigel) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] MediaControl node problem In-Reply-To: <3DDB9729.84557020@envivio.com> Message-ID: <001001c290dd$d50b3990$0500a8c0@christianpc> Renaud Cazoulat wrote: > Hi, > > > Maybe it's a bug in mp4tool. What did envivio different by creating > > the scene? I found that the scene time of the envivio > sample scene is > > 20min whereas the time of "my" envivio scene was just as > long as the > > video. Could there be some problem with the time bases resulting in > > these problems? > > Could be, indeed. If you want to "pause" a stream, you must > be sure that it is *not* linked to the same OCR as the BIFS > scene (otherwize you would pause the scene itself!) At least > this could be an explanantion why the envivio content does > not show the same durations for the scene and for the video > (despite it could be the same but that makes more sense to > have a longuer scene duration in this case) The more I study the spec the more I believe that this is really the problem. Unfortunately I'm not so familiar with time stamps and time bases. I tried to use an additional OCR stream as time base for the video ES but without success. Maybe someone could give an example for a typical ObjectDescriptor with aggreated ES descriptor and especially SLConfigDescriptor (of course in text description) describing a scene with scene description stream and a video or audio media object embedded in this scene where the streams are not linked to the same time base. Regards Christian Weigel From Cyril.Concolato enst.fr Thu Nov 21 09:02:04 2002 From: Cyril.Concolato enst.fr (Cyril Concolato) Date: Wed Jul 30 14:09:14 2003 Subject: [M4IF Technotes] MediaControl node problem References: <001001c290dd$d50b3990$0500a8c0@christianpc> Message-ID: <3DDC92FC.3E619D68@enst.fr> Hi Christian, > > Could be, indeed. If you want to "pause" a stream, you must > > be sure that it is *not* linked to the same OCR as the BIFS > > scene (otherwize you would pause the scene itself!) At least > > this could be an explanantion why the envivio content does > > not show the same durations for the scene and for the video > > (despite it could be the same but that makes more sense to > > have a longuer scene duration in this case) > The more I study the spec the more I believe that this is really the > problem. Unfortunately I'm not so familiar with time stamps and time > bases. I tried to use an additional OCR stream as time base for the > video ES but without success. Maybe someone could give an example for a > typical ObjectDescriptor with aggreated ES descriptor and especially > SLConfigDescriptor (of course in text description) describing a scene > with scene description stream and a video or audio media object embedded > in this scene where the streams are not linked to the same time base. Here is an example in text format that works fine with the Envivio plug-ins. You can encode it with mp4tool. BR, Cyril -- Ecole Nationale Sup?rieure des T?l?communications, Paris Dept. Comelec 46, rue Barrault 75013 Paris Tel: +33145817991 Fax: +33145804036 -------------- next part -------------- A non-text attachment was scrubbed... Name: enst_mediacontrol.bt Type: application/x-unknown-content-type-ultraedit.bt Size: 4150 bytes Desc: not available Url : /pipermail/mp4-tech/attachments/20021121/da6d4476/enst_mediacontrol.bin From hagai enquad.com Thu Nov 21 12:00:11 2002 From: hagai enquad.com (Hagai Folkman) Date: Wed Jul 30 14:09:15 2003 Subject: [M4IF Technotes] (no subject) References: <20021119160729.23570.qmail@webmail17.rediffmail.com> Message-ID: <003101c29144$ca3b9020$c26e003e@default> > Hi, > > The Sprite coding is used to enable transfer of still images along with the > video stream. > The Sprite stream can represent full frame image or part of the image frame. > The use of Sprite coding can be helpful for Object-Based encoding to > transfer the background > object or the changes in the background as slides of frames. But it can be > used for any still > image transmission over MPEG-4. > Only Main-Profile support Sprite Coding. > > best regards, > Hagai ----- Original Message ----- From: Ramanathan Subramanian To: Sent: éåí ùìéùé 19 ðåáîáø 2002 18:07 Subject: [M4IF Technotes] (no subject) > Hi all > > I am doing some research on MPEG 4 Video object Coding.I am using > the Momusys encoder code to create mp4 video content. I did get > some MPEG 4 masks but I am unable to employ the "use Sprite > coding" option in the .cfg file...as the code does not execute > > Could anybody tell me in what cases (more specifically, streams) > for which sprite coding could be used and what is the required > configuration when sprite coding is used?? is it that you get 2 > objects(namely the background and foreground resp) when u do > sprite coding even without giving a .seg input?? > > I hope my questions make sense.. > > would be great if somebody could help.. > > Regards > Ramanathan > > > Ramanathan Subramanian > Project Assistant > Multimedia Systems Lab > Super Computer Education & Research Center > Indian Institute of Science > Bangalore-560 012 > Email: pillayar@mmsl.serc.iisc.ernet.in > Phone: +91-80-3942896 > > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes From christian.weigel epost.de Thu Nov 21 13:08:06 2002 From: christian.weigel epost.de (Christian Weigel) Date: Wed Jul 30 14:09:17 2003 Subject: AW: [M4IF Technotes] MediaControl node problem In-Reply-To: <3DDC92FC.3E619D68@enst.fr> Message-ID: <000a01c29156$b2134ee0$0500a8c0@christianpc> Cyril Concolato wrote: > Here is an example in text format that works fine with the > Envivio plug-ins. You can encode it with mp4tool. > > BR, > > Cyril I'm sorry to say that but I tried the attached bt file together with the extracted media files from the envivio sample and I have still the same problem (PAUSE IS Not working correctly). When I used the mp4tool.exe it outputs following warning: MP4 Manager v020318 (c) ENST 1999-2002 WARN: ESDescriptor -> error fixing ocr idWARN: ESDescriptor -> error fixing ocr id I also tried it with the java version: MP4 Manager v020627 (c) ENST 1999-2002 WARN: ESDescriptor -> error fixing ocr id WARN: ESDescriptor -> error fixing ocr id Because of the reply by Renaud I assume these fixed ocr ID's could be the reason. There are two OCR_ES_ID's (in the scene description stream and the ES of the video) refering to an ES_ID 0. But this ES doesn't exist. Regards Christian Weigel From Cyril.Concolato enst.fr Thu Nov 21 13:33:53 2002 From: Cyril.Concolato enst.fr (Cyril Concolato) Date: Wed Jul 30 14:09:17 2003 Subject: AW: [M4IF Technotes] MediaControl node problem References: <000a01c29156$b2134ee0$0500a8c0@christianpc> Message-ID: <3DDCD2B1.B925A80B@enst.fr> Hi, > I'm sorry to say that but I tried the attached bt file together with the > extracted media files from the envivio sample and I have still the same > problem (PAUSE IS Not working correctly). I've tried it with the current mp4tool that we have internally and it works. It will be released soon on the web. Stay tuned. > Because of the reply by Renaud I assume these fixed ocr ID's could be > the reason. There are two OCR_ES_ID's (in the scene description stream > and the ES of the video) refering to an ES_ID 0. But this ES doesn't > exist. The spec says that OCR_ES_IS 0 means that the stream is explicitly not synchronized with any other stream. Basically in the scene I sent, it explicitely says that the BIFS stream is not synchronized with any other stream. Then, it says that the OD stream is synchronized with the BIFS stream. Finally, it says that the video stream is not synchronized neither with the BIFS stream nor with the OD one. This allows for pausing and stopping of the video indepently from the BIFS and OD streams. Hope it helps, Cyril -- Ecole Nationale Sup?rieure des T?l?communications, Paris Dept. Comelec 46, rue Barrault 75013 Paris Tel: +33145817991 Fax: +33145804036 From christian.weigel epost.de Thu Nov 21 13:58:54 2002 From: christian.weigel epost.de (Christian Weigel) Date: Wed Jul 30 14:09:17 2003 Subject: [M4IF Technotes] MediaControl node problem In-Reply-To: <3DDCD2B1.B925A80B@enst.fr> Message-ID: <000b01c2915d$c3dbe770$0500a8c0@christianpc> Cyril Concolato wrote: > I've tried it with the current mp4tool that we have > internally and it works. It will be released soon on the web. > Stay tuned. That's good news. I can't hardly wait :o). > The spec says that OCR_ES_IS 0 means that the stream is > explicitly not synchronized with any other stream. > Basically in the scene I sent, it explicitely says that the > BIFS stream is not synchronized with any other stream. Then, > it says that the OD stream is synchronized with the BIFS > stream. Finally, it says that the video stream is not > synchronized neither with the BIFS stream nor with the OD > one. This allows for pausing and stopping of the video > indepently from the BIFS and OD streams. Now I understood. Thx. > Hope it helps, Yes, thanks to all people involved in this discussion. I've learned much from it. Regards Christian Weigel From ben interframemedia.com Sat Nov 23 12:29:22 2002 From: ben interframemedia.com (Ben Waggoner) Date: Wed Jul 30 14:09:17 2003 Subject: [M4IF Technotes] GUI editor for BIFS text files? Message-ID: Folks, Say, has anyone made a GUI editor for BIFS text files? Editing them by hand is not proving to be a pleasant experience. Ben Waggoner Compressed Video Consulting, Training, and Encoding My Book: http://www.benwaggoner.com/books.htm Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html Compression Books: http://www.benwaggoner.com/bookshelf.htm From zhaozhd bis.com.cn Sun Nov 24 17:20:05 2002 From: zhaozhd bis.com.cn (zhaozhd) Date: Wed Jul 30 14:09:17 2003 Subject: [M4IF Technotes] decode error Message-ID: <008001c2939a$b6e590e0$3964e29f@bis.com.cn> Hello I encoded a QCIF image (frame number (0-9),) on xvid with framerate 15,bitrate 64 according to the Simple Level0.then I decoded it on xvid but the decoder died.So I decoded it on another decoder that can only decodes video of Simple Level0,But not every decoded frame is fine,there were errors at the end of frame 3 and the last frame 9.Can you tell me why. Thanks. Zhao Zhidong -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021124/3387fea1/attachment.html From yshain neomagic.com Sun Nov 24 21:10:53 2002 From: yshain neomagic.com (Yossi Shain) Date: Wed Jul 30 14:09:17 2003 Subject: [M4IF Technotes] Downsample filters In-Reply-To: <200211181704.gAIH4OPf004948@mx3.magma.ca> Message-ID: Gentlemen, Looking at the filter taps for downsampling the luminance componenet of an image by 1/2 or 1/4, the mpeg4 Pisa document (ISO/IEC JTC1/SC29/WG11 N3908, January 2001/Pisa, Section 2.2.2, page 17) recommends using a 13-tap or 15-tap filter. I wanted to know whether these filters offered noticable improvement in quality over other filters for qcif video? If yes, was this a result of a crisp downsampled image, or a reduction in noise that enabled better compression? From OPHIR il.ibm.com Tue Nov 26 11:53:15 2002 From: OPHIR il.ibm.com (Ophir Azulai) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] Block boundary mirroring in quarter pel Message-ID: Hi Block boundary mirroring is defined as symmetrically extension of the block boundaries (Section 7.6.2.2. of the visual (14496-2) ). Does this definition also apply for half pel and quarter pel calculation? e.g. for calculating the 'd' half pel or 'k' quarter pel (defined in section 7.6.2.2.1 figure V2-18b) at the top edge of a block. Regards, Ophir Azulai IBM Research Laboratory in Haifa From nico.oorts xmt.be Tue Nov 26 15:47:31 2002 From: nico.oorts xmt.be (Nico Oorts) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] mp4tool References: <571B06D35309794BA9204B5A090CC1010101E3CA@lanmhs50.rd.francetelecom.fr> <3D75C002.D2644FA4@enst.fr> <3D75C3D9.7030507@xmt.be> <3D75C94F.C957274A@enst.fr> <3D75CBE3.3000206@xmt.be> <3D75D2F1.BE803EFC@enst.fr> <3D76013D.5060201@xmt.be> <3D7783BA.4090605@enst.fr> Message-ID: <3DE38983.90400@xmt.be> Dear Jean-Claude, What I like to do with your mp4tool is to convert a .bt into a scene track and an OD track. After this I would like to include them into an mp4 file. It would be easy if you can supply us with some library and a little bit of documentation. Is there already some simple license available? Thanks. Best regards, Nico Oorts VRT research Jean-Claude Dufourd wrote: > nico.oorts wrote: > >> I think you convinced us to keep using the mp4tool in the form of a >> dll or a protected jar file. >> Can you tell me how we obtain such a license? > > > If you need just a version of the existing mp4tool as a obfuscated > jar, with just enough doc to call the correct function directly rather > than use the command-line, and if you need this for internal work, we > do not really need a licence other than the one for mp4tool.exe (on > the web site). We just need to agree on the way to aknowledge the use > of mp4tool when you publish a paper or if you distribute your tool > with the mp4tool jar in it. > I'll start our legal people on drafting some simple licence saying this. > > Meanwhile, please describe which options you are using, so that we > package the main class for use not as a "main()" but as a library. > We will do just the minimum useful for you. > > Best regards > JC From Wesley.DeNeve rug.ac.be Tue Nov 26 15:58:28 2002 From: Wesley.DeNeve rug.ac.be (Wesley De Neve) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] GUI editor for BIFS text files? References: Message-ID: <00ee01c2955c$47195e90$ba53c19d@Poulenc> Ben Waggoner wrote: > Folks, > > Say, has anyone made a GUI editor for BIFS text files? Editing > them by hand is not proving to be a pleasant experience. I'd like to second this one. Kind regards, Wesley De Neve -- Ghent University ELIS / Multimedia Lab From nixon fokus.gmd.de Tue Nov 26 17:59:12 2002 From: nixon fokus.gmd.de (Lyndon J B Nixon) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] GUI editor for BIFS text files? References: <00ee01c2955c$47195e90$ba53c19d@Poulenc> Message-ID: <3DE3A860.3080109@fokus.fhg.de> I can third that :) I would think BIFS is not intended to be the format in which files would be manually edited, for that you should look into using the XML representation for MPEG-4, 'XMT' (which comes in two flavours. If you want to remain close to BIFS, there's XMT-A, and for a more readable and more abstract syntax which is much like SMIL, there's XMT-O.) Then you can utilise any of the multitude of XML editors out there. There is also reference software to transform from XMT-O to XMT-A, from XMT-A to BIFS and to encode XMT into a mp4 file. lyndon >Ben Waggoner wrote: > > >>Folks, >> >> Say, has anyone made a GUI editor for BIFS text files? Editing >>them by hand is not proving to be a pleasant experience. >> From hanxy magima.com.cn Wed Nov 27 09:10:45 2002 From: hanxy magima.com.cn (hanxy) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] TS Testcast??? Message-ID: <001301c295b1$d1a3d400$b800000a@sh.magima.com> Hi, Can somebody tell me where to find example file of transport stream, which I want to use for my SW TS Decoder. Thanks in advance! Speer Han -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021127/0812d2fd/attachment.html From jean-claude.dufourd enst.fr Wed Nov 27 09:14:38 2002 From: jean-claude.dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] mp4tool References: <571B06D35309794BA9204B5A090CC1010101E3CA@lanmhs50.rd.francetelecom.fr> <3D75C002.D2644FA4@enst.fr> <3D75C3D9.7030507@xmt.be> <3D75C94F.C957274A@enst.fr> <3D75CBE3.3000206@xmt.be> <3D75D2F1.BE803EFC@enst.fr> <3D76013D.5060201@xmt.be> <3D7783BA.4090605@enst.fr> <3DE38983.90400@xmt.be> Message-ID: <3DE47EEE.8060304@enst.fr> Dear Nico, Nico Oorts wrote: > What I like to do with your mp4tool is to convert a .bt into a scene > track and an OD track. After this I would like to include them into > an mp4 file. It would be easy if you can supply us with some library > and a little bit of documentation. Is there already some simple > license available? Thanks. Sorry for not answering you sooner. Though it would not be so difficult for us to produce such a package from mp4tool, we are very busy and even giving software away creates *always* a lot of work. And there is the question of licencing. I'll work on it on the technical side, but I welcome suggestions for the distribution: is there a way to say in the mp4tool library licence that we have not paid the MPEG licence so you have to pay it ? Basically, we cannot both give the software away and pay the licence. We may have to withdraw mp4tool from free download. Open source is not a possibility for us, and the free distribution of mpeg4ip is only tolerated, they do not really have the right. So help on a free (or cheap) distribution scheme appreciated. Best regards JC From jean-claude.dufourd enst.fr Wed Nov 27 09:37:35 2002 From: jean-claude.dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] GUI editor for BIFS text files? References: Message-ID: <3DE4844F.8050600@enst.fr> Dear Ben, Ben Waggoner wrote: > Say, has anyone made a GUI editor for BIFS text files? Editing them by >hand is not proving to be a pleasant experience. > > There is, to the best of my knowledge: - the Envivio suite, which is quite expensive because it does a lot more than just BIFS... and BIFS is not so easy to author with EBS either. - a tool by a Korean company called Net&TV, which is very Flash-like on the GUI side. We saw a demo at the Shanghai MPEG meeting. I am not sure it is out yet, but soon. - a tool by Hong Kong university (look for papers to WEMP) - a tool by a university in Taiwan (same) with templates and 3D - a tool demonstrated by HHI for 3D, not sure if it is a product yet. - and last but not least, TDK just announced a product called WonderCreator in Japan (based on ENST, TILAB and Optibase technology), which is a template-based, 2D GUI for the Japanese market, creating BIFS scenes with a bias toward school usage in the available templates, to be released in Japan only on the 6th of December. This tool hides all of the BIFS complexity from the user, but of course in return you are constrained in what you can do, for example: in this version, no BIFS updates. However, no English version is planned yet for commercial release. Since the TDK project, we have stopped doing GUIs. Our expertise being on the MPEG-4 authoring and not on the GUI, we lost too much time on this side, so we are now sticking to doing the "engines" and would leave GUI to others. We have translation from a subset of Flash to BIFS, but that means no video. We have translation from a large subset of SVG to BIFS, but that means no updates. So to conclude this long answer, we do not have the right answer to your problem. Best regards JC From jean-claude.dufourd enst.fr Wed Nov 27 09:46:39 2002 From: jean-claude.dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] GUI editor for BIFS text files? References: <00ee01c2955c$47195e90$ba53c19d@Poulenc> Message-ID: <3DE4866F.2090107@enst.fr> Dear Wesley Wesley De Neve wrote: >Ben Waggoner wrote: > > >>Folks, >> >> Say, has anyone made a GUI editor for BIFS text files? Editing >>them by hand is not proving to be a pleasant experience. >> >> > > >I'd like to second this one. > > From this and your private email, I deduce that your first need is for a tutorial, or at least for some simple samples with media and a clear procedure to produce the MP4s. This is a reasonable request. We will try to expand our mp4tool website to include such basic tutorials. If we do not do it, please remind us regularly. ;-) Best regards JC -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20021127/6fc34359/attachment.html From Jean-Claude.Dufourd enst.fr Wed Nov 27 13:47:39 2002 From: Jean-Claude.Dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] Quick BIFS primer Message-ID: <3DE4BEEB.6020002@enst.fr> Or how to create your first BIFS scenes... Ten samples, complete with text, scripts, exe, media, all tested with the current MPEG-4 Systems software (Osmose). http://www.comelec.enst.fr/~dufourd/mpeg-4/Bifs_Primer/primer.html Bug reports welcome. Enjoy. Best regards JC PS: Most files do not work with Envivio, for various reasons: - no G723 or H263 decoder - no texturing of circle - bug in their mpeg-4 video on non-multiple of 8 sizes (uninitialized padding) hence the green fog patches -- Jean-Claude Dufourd @======================================@ ENST, Dept COMELEC The wing, over the big rock... 46, rue Barrault @======================================@ 75013 Paris Tel: +33145817807 Fax: +33145804036 From wmdeneve bilbo.elis.rug.ac.be Wed Nov 27 17:30:00 2002 From: wmdeneve bilbo.elis.rug.ac.be (Wesley De Neve) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] Quick BIFS primer In-Reply-To: <3DE4BEEB.6020002@enst.fr> Message-ID: Hi Jean-Claude, all first of all, thanks for the swift reply ;-). On Wed, 27 Nov 2002, Jean-Claude Dufourd wrote: > Or how to create your first BIFS scenes... > Ten samples, complete with text, scripts, exe, media, all tested with > the current MPEG-4 Systems software (Osmose). > > http://www.comelec.enst.fr/~dufourd/mpeg-4/Bifs_Primer/primer.html > > Bug reports welcome. I've had a first (and rather quick) look at it. I was able to compile all sequences with BIFSEnc and MP4Enc. I've tried to play them with OSMOSE and IM1 but this only succeeded for the scenes containing only BIFS or JPEGS. For the other scenes a message box popped up saying "UNKNOWN DECODER TYPE". I think this has something to do with registering the necessary DLLs? I tried this already by doubleclicking on IM1.reg but after having looked at the values in the reg file, it seems that some paths are not fully correct. I will have a look at this tomorrow. Regarding the samples that are supposed to be compiled with MP4Tool, I think that some media files are missing, like cat3.{info, media} and audio.media? I have found back audio.media in the conformance bitstreams but after placing it in the streams directory, MP4Tool was still complaining that it couldn't find the file in question. Would it also be possible to put a sample online that works with a visual MPEG-4 elementary stream (sorry for nagging)? I'm looking forward to your answer. Many thanks to all who have offered help! With kind regards, Wesley De Neve From Jean-Claude.Dufourd enst.fr Wed Nov 27 18:23:59 2002 From: Jean-Claude.Dufourd enst.fr (Jean-Claude Dufourd) Date: Wed Jul 30 14:09:18 2003 Subject: [M4IF Technotes] Quick BIFS primer References: Message-ID: <3DE4FFAF.1010003@enst.fr> Wesley De Neve wrote: > For the other scenes a message box popped up saying "UNKNOWN DECODER > TYPE". I think this has something to do with registering the necessary > DLLs? I tried this already by doubleclicking on IM1.reg but after having > looked at the values in the reg file, it seems that some paths are not > fully correct. I will have a look at this tomorrow. Yes, that is it. You need to register the decoders. > Regarding the samples that are supposed to be compiled with MP4Tool, I > think that some media files are missing, like cat3.{info, media} and > audio.media? These files are all on the web site, either in the all.zip or in the directory directly. Beware, the streams directory is different in the mp4tool and in BIFSENC_MP4ENC directories... > Would it also be possible to put a sample online that works with a visual > MPEG-4 elementary stream (sorry for nagging)? cat3, in directory m4v, is one such sample. Best regards JC -- Jean-Claude Dufourd @======================================@ ENST, Dept COMELEC The wing, over the big rock... 46, rue Barrault @======================================@ 75013 Paris Tel: +33145817807 Fax: +33145804036 From alisson dsc.ufcg.edu.br Wed Nov 27 14:33:27 2002 From: alisson dsc.ufcg.edu.br (Alisson Vasconcelos de Brito / Pos. COPIN) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Mp4 file's quality measurement. (fwd) Message-ID: ---------- Forwarded message ---------- Date: Wed, 27 Nov 2023 14:32:36 -0300 (-03:00) From: Alisson Vasconcelos de Brito / Pos. COPIN To: technotes@lists.m4if.com Subject: Mp4 file's quality measurement. Hi all, I'd like to know if anybody knows any tool that can be used measure an mpeg-4 video file quality (PSNR, for example). I have to compare a modified MPEG-4 file with its original. I am working in my master course and it's extremely important to me. Thanks, Alisson. From the_ether btinternet.com Wed Nov 27 18:13:52 2002 From: the_ether btinternet.com (the_ether) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Mp4 file's quality measurement. (fwd) References: Message-ID: <006701c29640$c3a97050$0100a8c0@DesktopGC> At the end of the day there's no better judge of the quality than humans. There has been some research on automated tools to try and assess human perceptions. Indeed, the U.S. government is promoting a software tool developed by an agency of the Department of Commerce that can measure digital video picture quality. See the following article: http://www.eetimes.com/story/OEG20021118S0027 > Hi all, > I'd like to know if anybody knows any tool that can be used > measure an mpeg-4 video file quality (PSNR, for example). I have to > compare a modified MPEG-4 file with its original. I am working in > my master course and it's extremely important to me. > Thanks, > Alisson. > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > From rob.koenen m4if.org Wed Nov 27 11:57:34 2002 From: rob.koenen m4if.org (Rob Koenen) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Mp4 file's quality measurement. (fwd) In-Reply-To: <006701c29640$c3a97050$0100a8c0@DesktopGC> Message-ID: <3C124172E7FDD511B510000347426D59011B42CB@exchange.epr.com> The interesting topic of automated quality evaulations comes up on these lists from time to time. I have spent a few million of my previous employer's dollars (euros) on researching the issue. We obtained world class results. But in the end, intensive tests could not find that the objective measure we developed differed from SNR in a statistically significant way. None of the measures in that same test (carried out by ITU's Video Quality Experts Group, vqeg, www.vqeg.org) performed better than good old SNR as a matter of fact. So I'd like to caution people that want to use objective measures and extend the conclusions to quality as observed by humans. These measures are great for monitoring, they may help you in your coding research, but they will not allow reasonable conclusions on how codecs compare. Rob > -----Original Message----- > From: the_ether [mailto:the_ether@btinternet.com] > Sent: Wednesday, November 27, 2023 10:14 > To: technotes@lists.m4if.org > Subject: Re: [M4IF Technotes] Mp4 file's quality measurement. (fwd) > > > At the end of the day there's no better judge of the quality > than humans. > > There has been some research on automated tools to try and > assess human perceptions. Indeed, the U.S. government is > promoting a software tool developed by an agency of the > Department of Commerce that can measure digital video picture > quality. See the following article: > http://www.eetimes.com/story/OEG20021118S0027 > Hi all, > I'd like to know if anybody knows any tool that can be used measure an > mpeg-4 video file quality (PSNR, for example). I have to compare a > modified MPEG-4 file with its original. I am working in my master > course and it's extremely important to me. Thanks, > Alisson. > > > _______________________________________________ > Technotes mailing list > Technotes@lists.m4if.org > http://lists.m4if.org/mailman/listinfo/technotes > _______________________________________________ Technotes mailing list Technotes@lists.m4if.org http://lists.m4if.org/mailman/listinfo/technotes From n_habili hotmail.com Thu Nov 28 05:49:29 2002 From: n_habili hotmail.com (Nariman Habili) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] MP4 rate control Message-ID: I'm currently using the Microsoft MPEG-4 Video Reference Software, Version: Microsoft-FDAM1-2.3-001213. I have segmented a video sequence (in QCIF format) into two VOPs (background and foreground). I would like to compress each VOP separately, at different bit-rates, using binary shape coding. The length of the sequence is 8 frames. I'm using MP4 rate control. I've set Texture.QuantStep.IVOP and Texture.QuantStep.PVOP manually, as suggested in the User Manual. However, I find that the same visual quality is obtained for each VOP after decoding using significantly different bit-rates (e.g., 64,000 bps and 8,000 bps); moreover, the average quantizer value is the same for the different bit-rates. I've even tried setting the bit-rate to 0, and the encoder still works! I would be grateful if anyone could suggest ways to remedy this problem. Is there any other software that I can use for the same purpose? regards Nariman -- Nariman Habili Visiting research fellow School of Computer Engineering Nanyang Technological University Singapore _________________________________________________________________ Add photos to your e-mail with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail From oamato wanadoo.fr Thu Nov 28 11:22:18 2002 From: oamato wanadoo.fr (Olivier Amato) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Mp4 file's quality measurement. (fwd) References: Message-ID: <00e801c296c8$0d194260$f340fea9@beck> > I'd like to know if anybody knows any tool that can be used > measure an mpeg-4 video file quality (PSNR, for example). I have to > compare a modified MPEG-4 file with its original. I am working in > my master course and it's extremely important to me. Some companies claim to provide "objective video quality measurement", but I don't know if they can manage MPEG-4 Visual : - Sarnoff's JNDMetrix ( http://www.jndmetrix.com/ ). - Genista's Video PQoS ( http://www.genista.co.jp/Products/VideoSolutions.htm ). Regards, Olivier From xxh soft.thtf.com.cn Thu Nov 28 22:43:57 2002 From: xxh soft.thtf.com.cn (Xue Xuhong) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] About Quarter sample mode interpolation Message-ID: <007b01c296ec$95903760$32d06fa6@xuexh> hi, all in ISO/IEC 14496?2:1999/Amd.1:2000(E), subclause 7.6.2.2 Quarter sample mode interpolation(P207), it is described as following, [ In quarter sample mode interpolation, for each block of size MxN in the reference VOP which position is defined by the decoded motion vector for the block to be predicted, a reference block of size (M+1)x(N+1) biased in the direction of the half or quarter sample position is read from the reconstructed and padded reference VOP. Then, this reference block is symmetrically extended at the block boundaries by three samples using block boundary mirroring according to Figure V2 - 17. Figure V2 - 17 -- block boundary mirroring ] Does Figure V2 - 17 provide an method to do quarter sample interpolation, I can not understand, would anyone please say some words on it? Thank you. Best Regards, Xue xuhong From onural ee.bilkent.edu.tr Fri Nov 29 00:02:24 2002 From: onural ee.bilkent.edu.tr (Levent Onural) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Embeddable mpeg-4 codec Message-ID: <200211282202.gASM2OQ27417@dalyan.ee.bilkent.edu.tr> Hello All, I appreciate some info regarding immediately available software mpeg-4 codec for x86 platforms for variable bit-rate operation; decoder embedded in IE and/or Java. Pointers, technical specs and price info (for volume purchase) are welcome. Thanks... Levent *************************************************** Levent Onural Professor Bilkent University l.onural@ieee.org onural@bilkent.edu.tr http://www.ee.bilkent.edu.tr/~onural *************************************************** From d98rolb stud.hh.se Fri Nov 29 19:06:45 2002 From: d98rolb stud.hh.se (Roland Bengtsson) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Average compression Message-ID: <5.1.1.6.0.20021129185258.00b68a48@studpop.hh.se> I wonder how much compression MPEG-4 can achieve. Of course this is a very general questions and the compression depends on many things like the movie, tha algorithm that is used etc. But I still want to know: I have seen a number around 17x compression. Maybe there is some documentation or paper that I can read further about this subject? Here's my thought of the factors that may affect the compression: The movie, is it slow or a action movie? The algorithm for encoding, I would like to hear some numbers about the new H.26L/AVC algorithm. What profile is used, for example how effective is FGS against simple profile? The cpu-power available, this is of course important. Generally harder compression require more MIPS. The amount of error-correction, have the medium for transport of the movie hight biterror rate? Thoughts please! Regards Roland Bengtsson Student at Halmstad University Master of Computer Science -------------- next part -------------- --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.419 / Virus Database: 235 - Release Date: 2023-11-13 From chl math.uni-bonn.de Fri Nov 29 19:47:10 2002 From: chl math.uni-bonn.de (Christoph Lampert) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Average compression In-Reply-To: <5.1.1.6.0.20021129185258.00b68a48@studpop.hh.se> Message-ID: On Fri, 29 Nov 2002, Roland Bengtsson wrote: > I wonder how much compression MPEG-4 can achieve. Of course this is a very > general questions and the compression depends on many things like the > movie, tha algorithm that is used etc. But I still want to know: I have > seen a number around 17x compression. That's a rather low number, I guess. For normal material (not too noise, etc.) you get rather good quality with less than 0.2 bits per encoded pixel, which is compression ratio of 1:60 over uncompressed YUV 4:2:0 material (12 bits per pixel), and even 1:120 over uncompressed 24 bit RGB (24 bits per pixel). Even DVD-quality MPEG-2 (with up to 10MBps at 720x576 25fps) reaches a ratio of 1:12 or 1:24. Christoph -- Christoph H. Lampert chl@math,uni-bonn,de | Diese Signature wurde maschi- Beringstr. 6, Zi. 15, 53115 Bonn, Germany | nell erstellt und bedarf Tel. (0228) 73-4708 Fax. +49 228 73-7916 | keiner Unterschrift. AZ 27B-6 From ben interframemedia.com Fri Nov 29 12:18:43 2002 From: ben interframemedia.com (Ben Waggoner) Date: Wed Jul 30 14:09:19 2003 Subject: [M4IF Technotes] Average compression In-Reply-To: <5.1.1.6.0.20021129185258.00b68a48@studpop.hh.se> Message-ID: on 11/29/02 10:06, Roland Bengtsson at d98rolb@stud.hh.se wrote: > I wonder how much compression MPEG-4 can achieve. Of course this is a very > general questions and the compression depends on many things like the > movie, tha algorithm that is used etc. But I still want to know: I have > seen a number around 17x compression. I've certainly had useable output at much, much higher rates than that (beyond 100x compared to the uncompressed 4:2:0 source at the same resolution). The real question isn't how low you can go, but how low you can go with "good enough" quality, where "good enough" varies a lot from case to case. > Maybe there is some documentation or paper that I can read further about > this subject? > > Here's my thought of the factors that may affect the compression: > > The movie, is it slow or a action movie? Major effect. Maybe 5x or more between the easiest and hardest kinds of typical content. > The algorithm for encoding, I would like to hear some numbers about the new > H.26L/AVC algorithm. The profile matters a lot, and different encoders targeting the same profile can vary by at least 3x. > What profile is used, for example how effective is FGS against simple profile? Significant factor. > The cpu-power available, this is of course important. Generally harder > compression require more MIPS. Significant. Doesn't matter much with offline encodes, since it just goes slower, but real-time encoders may have to make sacrifices in quality in order to be real-time. Also, they can't do 2-pass. > The amount of error-correction, have the medium for transport of the movie > hight biterror rate? Slight effect at most. But the profiles chosen for lossy environments like FGS may be less efficient than those chosen for lossless environments. Ben Waggoner Compressed Video Consulting, Training, and Encoding My Book: http://www.benwaggoner.com/books.htm Cleaner Tutorial: http://www.saferseas.com/navseries/adclean.html Compression Books: http://www.benwaggoner.com/bookshelf.htm