From lmh ycig.com Wed Mar 1 10:09:00 2006 From: lmh ycig.com (ycigcom) Date: Wed Mar 1 05:46:13 2006 Subject: [Mp4-tech] Has anybody know about Chinese AVS's codec? Message-ID: <001b01c63cd5$1c1f6d00$1201a8c0@USER2D5B206299> Hi, experts! Has anybody know about Chinese AVS codec? Could you give me any comparative results between it and other Codec , eg h.264? Thanks a lot! Best Regards lmh -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/cfa3453f/attachment-0001.html From EricChuang via.com.tw Wed Mar 1 15:59:51 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Wed Mar 1 05:46:17 2006 Subject: [Mp4-tech] [H.264]temporal direct luma motion vector Message-ID: <6A7AB4426712514F9E2AB28C407285FA2849D1@exchtp02.vntek.com.tw> Dear experts, Do you have any idea how often temporal direct luma motion vector is used in H.264 ? Thanks, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/783f2deb/attachment-0001.html From EricChuang via.com.tw Wed Mar 1 16:03:51 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Wed Mar 1 05:46:20 2006 Subject: [Mp4-tech] Question regarding reference list in B Slice Message-ID: <6A7AB4426712514F9E2AB28C407285FA2849D2@exchtp02.vntek.com.tw> Hi YK, Thanks very much for your answering. Regarding 3), it seems like that non-existing reference picture shall not be used in display, right ? Why it needs POC ? Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Ye-Kui.Wang@nokia.com Sent: Friday, February 24, 2024 4:42 AM To: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi Eric, 1. For temporal scalability. The reason is that temporal enhancement layers may contain reference pictures and they may not be present in the bitstream after stream thinning. 2. I think the latter is correct: the encoder has to make it impossible to have non-existing pictures being included in the lists. 3. POC is needed in constructing the initial reference picture lists for B slices and is also needed in the operation of output order DPB (C.4.1-C.4.5). BR, YK _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] Sent: 2006年2月22日 2:18 To: garysull@windows.microsoft.com; mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Question regarding reference list in B Slice Dear experts, In standard, it mentions the non-existing reference frame would be put into DPB when frame number gap is allowed. I have 3 questions about this. 1. What’s the usage of frame_num gap ? Why it’s introduced in standard ? Is there any document on the reason ? 2. When constructing reference list for B Slice, the standard describes that reference pictures marked as non-existing are not included in RefList0 or RefList1, is this exclusion operaton is done specially by decoder, or encoder has make it impossible to happen by using reference picture list reordering ? 3. Since non-existing reference picture cannot be put into RefList0 or RefList1 in B slice, why standard specifies the default value of delta_pic_order_cnt[0] and delta_pic_order_cnt[1] to them in section 8.2.5.2. Is there any place to reference the POC of non-existing pictures ? Thanks in advance, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/b6cfa67f/attachment-0001.html From prasanna.sethuraman patni.com Wed Mar 1 14:01:23 2006 From: prasanna.sethuraman patni.com (Prasanna Sethuraman) Date: Wed Mar 1 05:46:26 2006 Subject: [Mp4-tech] On MB skip in the JM encoder. In-Reply-To: <20060228060954.8262523D21@ws5-3.us4.outblaze.com> Message-ID: <200603010836.k218aS7a025306@spzha1.patni.com> Dear all, I was playing around with the JM encoder, and I find that it skips about 1/3rd of the macro blocks. As far as I understand, a macro block can be signaled as P_SKIP if the MVDs are zero and all the Quant. Coeffs are zero. Now, the JM encoder also sets some of the Quant. Coeffs to zero based on some thresholds. I disabled all that, but even after that, there is still a huge number of MBs that are skipped. Does this mean that the JM encoder also uses some other criteria for MB skip? If so, can someone hint me as to what that would be? Any information is highly appreciated. And thanks to Alexis, Gary and Satendra for enlightening me on motion vectors for chroma. Best regards, Prasanna. http://www.patni.com World-Wide Partnerships. World-Class Solutions. _____________________________________________________________________ This e-mail message may contain proprietary, confidential or legally privileged information for the sole use of the person or entity to whom this message was originally addressed. Any review, e-transmission dissemination or other use of or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you have received this e-mail in error kindly delete this e-mail from your records. If it appears that this mail has been forwarded to you without proper authority, please notify us immediately at netadmin@patni.com and delete this mail. _____________________________________________________________________ From Ye-Kui.Wang nokia.com Wed Mar 1 11:10:07 2006 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Wed Mar 1 05:46:30 2006 Subject: [Mp4-tech] Question regarding reference list in B Slice In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA2849D2@exchtp02.vntek.com.tw> Message-ID: <1C1F3D15859526459B4DD0A7A9B2268B01C0B381@trebe101.NOE.Nokia.com> Eric, The initial reference lists for B slices are constructed using POC. Non-existing pictrue should also be included in the initial list such that the temporal base layer pictures can be decoded identically when the temporal enhancement layer pictures are present (in this case they are NOT "non-exisiting pictures") or not (in this case they are "non-exisiting pictures"). However, POC values of non-existing picures are not needed in the operaiton of output order DPB (C.4.1-C.4.5) - I was wrong about this. BR, YK ________________________________ From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] Sent: 2006$BG/(J3$B7n(J1$BF|(J 10:04 To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi YK, Thanks very much for your answering. Regarding 3), it seems like that non-existing reference picture shall not be used in display, right ? Why it needs POC ? Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Ye-Kui.Wang@nokia.com Sent: Friday, February 24, 2024 4:42 AM To: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi Eric, 1. For temporal scalability. The reason is that temporal enhancement layers may contain reference pictures and they may not be present in the bitstream after stream thinning. 2. I think the latter is correct: the encoder has to make it impossible to have non-existing pictures being included in the lists. 3. POC is needed in constructing the initial reference picture lists for B slices and is also needed in the operation of output order DPB (C.4.1-C.4.5). BR, YK ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] Sent: 2006$BG/(J2$B7n(J22$BF|(J 2:18 To: garysull@windows.microsoft.com; mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Question regarding reference list in B Slice Dear experts, In standard, it mentions the non-existing reference frame would be put into DPB when frame number gap is allowed. I have 3 questions about this. 1. What$B!G(Js the usage of frame_num gap ? Why it$B!G(Js introduced in standard ? Is there any document on the reason ? 2. When constructing reference list for B Slice, the standard describes that reference pictures marked as non-existing are not included in RefList0 or RefList1, is this exclusion operaton is done specially by decoder, or encoder has make it impossible to happen by using reference picture list reordering ? 3. Since non-existing reference picture cannot be put into RefList0 or RefList1 in B slice, why standard specifies the default value of delta_pic_order_cnt[0] and delta_pic_order_cnt[1] to them in section 8.2.5.2. Is there any place to reference the POC of non-existing pictures ? Thanks in advance, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/0d9ce89a/attachment-0001.html From satish_babudv yahoo.com Wed Mar 1 01:21:02 2006 From: satish_babudv yahoo.com (satish babu) Date: Wed Mar 1 05:46:33 2006 Subject: [Mp4-tech] Regarding H.264 video codec Message-ID: <20060301092102.75823.qmail@web54705.mail.yahoo.com> Hi Techies, can anyone tell me the website where i can get QVGA video sequences. Regards, satish --------------------------------- Relax. Yahoo! Mail virus scanning helps detect nasty viruses! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/4cff2c8a/attachment-0001.html From EricChuang via.com.tw Wed Mar 1 17:52:45 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Wed Mar 1 05:46:37 2006 Subject: [Mp4-tech] Question regarding reference list in B Slice Message-ID: <6A7AB4426712514F9E2AB28C407285FA372D87@exchtp02.vntek.com.tw> Hi YK, According to 8.2.4.2.3 and 8.2.4.2.4 in standard, When pic_order_cnt_type is equal to 0, reference pictures that are marked as “non-existing” are not included in either RefPicList0 and RefPicList1. And with your description, it seems that when pic_order_cnt_type not equal to 0, non-existing pictures shall be put into RefList0 or RefList1. I just don’t understand why it differs depends on pic_order_cnt_type. Thanks, Eric -----Original Message----- From: Ye-Kui.Wang@nokia.com [mailto:Ye-Kui.Wang@nokia.com] Sent: Wednesday, March 01, 2024 1:10 AM To: Eric Chuang; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Eric, The initial reference lists for B slices are constructed using POC. Non-existing pictrue should also be included in the initial list such that the temporal base layer pictures can be decoded identically when the temporal enhancement layer pictures are present (in this case they are NOT "non-exisiting pictures") or not (in this case they are "non-exisiting pictures"). However, POC values of non-existing picures are not needed in the operaiton of output order DPB (C.4.1-C.4.5) - I was wrong about this. BR, YK _____ From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] Sent: 2006年3月1日 10:04 To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi YK, Thanks very much for your answering. Regarding 3), it seems like that non-existing reference picture shall not be used in display, right ? Why it needs POC ? Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Ye-Kui.Wang@nokia.com Sent: Friday, February 24, 2024 4:42 AM To: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi Eric, 1. For temporal scalability. The reason is that temporal enhancement layers may contain reference pictures and they may not be present in the bitstream after stream thinning. 2. I think the latter is correct: the encoder has to make it impossible to have non-existing pictures being included in the lists. 3. POC is needed in constructing the initial reference picture lists for B slices and is also needed in the operation of output order DPB (C.4.1-C.4.5). BR, YK _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] Sent: 2006年2月22日 2:18 To: garysull@windows.microsoft.com; mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Question regarding reference list in B Slice Dear experts, In standard, it mentions the non-existing reference frame would be put into DPB when frame number gap is allowed. I have 3 questions about this. 1. What’s the usage of frame_num gap ? Why it’s introduced in standard ? Is there any document on the reason ? 2. When constructing reference list for B Slice, the standard describes that reference pictures marked as non-existing are not included in RefList0 or RefList1, is this exclusion operaton is done specially by decoder, or encoder has make it impossible to happen by using reference picture list reordering ? 3. Since non-existing reference picture cannot be put into RefList0 or RefList1 in B slice, why standard specifies the default value of delta_pic_order_cnt[0] and delta_pic_order_cnt[1] to them in section 8.2.5.2. Is there any place to reference the POC of non-existing pictures ? Thanks in advance, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/8809a668/attachment-0001.html From getsunil_6 yahoo.co.in Wed Mar 1 02:46:54 2006 From: getsunil_6 yahoo.co.in (venkata sunil) Date: Wed Mar 1 08:51:09 2006 Subject: [Mp4-tech] JPEG - standard In-Reply-To: <200510181405.j9IE1fEm028258@lists1.magma.ca> Message-ID: <20060301104654.29012.qmail@web8503.mail.in.yahoo.com> Hi All, I am looking for JPEG related documents, like 1. ITU-T Recommendation T.81 | ISO/IEC 10918-1:1994, Information technology - Digital compression and coding of continuous-tone still images: Requirements and guidelines. 2. ITU-T Recommendation T.88 | ISO/IEC 14492-1, Lossy/lossless coding of bi-level images. 3. ITU-T Recommendation T.83 | ISO/IEC 10918-2:1995, Information technology - Digital compression and coding of continuous-tone still images: Compliance testing. 4. ITU-T Recommendation T.87 | ISO/IEC 14495-1, Lossless and near-lossless compression of continuous-tone still images-baseline. Please let me know where can i get them. Thanks in Advance, Sunil. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From Ye-Kui.Wang nokia.com Wed Mar 1 13:00:25 2006 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Wed Mar 1 08:51:13 2006 Subject: [Mp4-tech] Question regarding reference list in B Slice In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA372D87@exchtp02.vntek.com.tw> Message-ID: <1C1F3D15859526459B4DD0A7A9B2268B01C0B492@trebe101.NOE.Nokia.com> Hi Eric, I guess the reason for a different initial lists construction for pic_order_cnt_type equal to 0 than other values is that, there is no good way to specify a POC value for non-existing pictures when pic_order_cnt_type equasl to 0. BR, YK ________________________________ From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] Sent: 2006$BG/(J3$B7n(J1$BF|(J 11:53 To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi YK, According to 8.2.4.2.3 and 8.2.4.2.4 in standard, When pic_order_cnt_type is equal to 0, reference pictures that are marked as $B!H(Jnon-existing$B!I(J are not included in either RefPicList0 and RefPicList1. And with your description, it seems that when pic_order_cnt_type not equal to 0, non-existing pictures shall be put into RefList0 or RefList1. I just don$B!G(Jt understand why it differs depends on pic_order_cnt_type. Thanks, Eric -----Original Message----- From: Ye-Kui.Wang@nokia.com [mailto:Ye-Kui.Wang@nokia.com] Sent: Wednesday, March 01, 2024 1:10 AM To: Eric Chuang; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Eric, The initial reference lists for B slices are constructed using POC. Non-existing pictrue should also be included in the initial list such that the temporal base layer pictures can be decoded identically when the temporal enhancement layer pictures are present (in this case they are NOT "non-exisiting pictures") or not (in this case they are "non-exisiting pictures"). However, POC values of non-existing picures are not needed in the operaiton of output order DPB (C.4.1-C.4.5) - I was wrong about this. BR, YK ________________________________ From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] Sent: 2006$BG/(J3$B7n(J1$BF|(J 10:04 To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi YK, Thanks very much for your answering. Regarding 3), it seems like that non-existing reference picture shall not be used in display, right ? Why it needs POC ? Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Ye-Kui.Wang@nokia.com Sent: Friday, February 24, 2024 4:42 AM To: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Question regarding reference list in B Slice Hi Eric, 1. For temporal scalability. The reason is that temporal enhancement layers may contain reference pictures and they may not be present in the bitstream after stream thinning. 2. I think the latter is correct: the encoder has to make it impossible to have non-existing pictures being included in the lists. 3. POC is needed in constructing the initial reference picture lists for B slices and is also needed in the operation of output order DPB (C.4.1-C.4.5). BR, YK ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] Sent: 2006$BG/(J2$B7n(J22$BF|(J 2:18 To: garysull@windows.microsoft.com; mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Question regarding reference list in B Slice Dear experts, In standard, it mentions the non-existing reference frame would be put into DPB when frame number gap is allowed. I have 3 questions about this. 1. What$B!G(Js the usage of frame_num gap ? Why it$B!G(Js introduced in standard ? Is there any document on the reason ? 2. When constructing reference list for B Slice, the standard describes that reference pictures marked as non-existing are not included in RefList0 or RefList1, is this exclusion operaton is done specially by decoder, or encoder has make it impossible to happen by using reference picture list reordering ? 3. Since non-existing reference picture cannot be put into RefList0 or RefList1 in B slice, why standard specifies the default value of delta_pic_order_cnt[0] and delta_pic_order_cnt[1] to them in section 8.2.5.2. Is there any place to reference the POC of non-existing pictures ? Thanks in advance, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/dcd73a37/attachment-0001.html From fethi.tlili supcom.rnu.tn Wed Mar 1 12:34:54 2006 From: fethi.tlili supcom.rnu.tn (fethi) Date: Wed Mar 1 08:51:17 2006 Subject: [Mp4-tech] [H.264 Video] D1 Bitsreams Message-ID: <008401c63d24$29cf0d30$1a1010ac@FETHI> Hi All Where could I find H264 bistreams / D1 resolution ? Thanks in advance -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060301/f2bd5323/attachment.html From Alexis.Tourapis dolby.net Wed Mar 1 10:35:30 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Wed Mar 1 18:04:06 2006 Subject: [Mp4-tech] On MB skip in the JM encoder. Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01D698BA@sapphire.dolby.net> Dear Prasanna, I would recommend to you to study how mode decision is performed. Essentially the H.264/AVC reference software uses a Lagrangian based Rate Distortion Optimized (RDO) mode decision method which essentially selects a "best" mode for a Macroblock Mb_i according to the criterion: J(Mb_i,mode) = SSE_luma(Mb_i,mode) + SSE_chroma_u(Mb_i,mode) + SSE_chroma_v(Mb_i,mode) + lambda * rate(Mb_i,mode) The SSE values correspond to the Sum of Square error/difference of the fully reconstructed component given the current mode vs. the original pixel values. The rate(.) term corresponds to the exact rate that the encoder requires to encode the current mode, while lambda is the lagrangian multiplier. The mode J that leads to the smallest such value is selected as the encoding mode for the current MB. Note that lambda in the reference software tends to increase depending on quantization level since rate tends to become a more significant factor in that situation. The basic idea of the above technique is to try and optimize decision jointly for both bitrate and distortion. This has been shown to lead to considerably better compression performance compared to considering only one of these components. I would strongly recommend finding a copy of the Nov 1998 issue of the IEEE Signal Processing Magazine which contains some very important papers on the topic. See in particular the following: A. Ortega, and K. Ramchandran "Rate-distortion methods for image and video compression," in IEEE Signal Processing Magazine, Volume 15, Issue 6, pp. 23-50, Nov. 1998 G.J. Sullivan, T, Wiegand, "Rate-distortion optimization for video compression," in IEEE Signal Processing Magazine, Volume 15, Issue 6, pp. 74-90, Nov. 1998 To go back to the JM, as I have said the encoder considers all possible modes and selects the one the smallest J. Skip mode in particular is usually the mode that leads to the smallest (and at times considerably smaller) rate, but as you can expect it may be worse or have equal quality with other modes. Therefore it is still possible to select skip mode during encoding while you will also observe that the probability/frequency of selecting skip mode increases as bitrate decreases (QP increases). In any case you may check the function RDCost_for_macroblocks() in rdopt.c of the software. I hope the above info helps... Best regards, Alexis PS. Note that in the J computation skip mode could also cause an "abnormal" computation case since skip may even require zero bits to encode at times, therefore disregarding lambda within its computation. This could lead at times to bad decisions (over-selection of skip/direct modes) which, even though it may not appear to hurt objectively (i.e. in terms of PSNR) can hurt subjectively (i.e. creation of blocky artifacts which the deblocker may also not be able to compensate due to the way such works). For this purpose we have actually modified J not to consider rate but instead max(0.5,rate). Note that it is important many times to not rely on PSNR/MSE but to also verify quality with other techniques (especially visually) before making any claims that one technique is better than another. -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Prasanna Sethuraman Sent: Wednesday, March 01, 2024 12:31 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] On MB skip in the JM encoder. Dear all, I was playing around with the JM encoder, and I find that it skips about 1/3rd of the macro blocks. As far as I understand, a macro block can be signaled as P_SKIP if the MVDs are zero and all the Quant. Coeffs are zero. Now, the JM encoder also sets some of the Quant. Coeffs to zero based on some thresholds. I disabled all that, but even after that, there is still a huge number of MBs that are skipped. Does this mean that the JM encoder also uses some other criteria for MB skip? If so, can someone hint me as to what that would be? Any information is highly appreciated. And thanks to Alexis, Gary and Satendra for enlightening me on motion vectors for chroma. Best regards, Prasanna. http://www.patni.com World-Wide Partnerships. World-Class Solutions. _____________________________________________________________________ This e-mail message may contain proprietary, confidential or legally privileged information for the sole use of the person or entity to whom this message was originally addressed. Any review, e-transmission dissemination or other use of or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you have received this e-mail in error kindly delete this e-mail from your records. If it appears that this mail has been forwarded to you without proper authority, please notify us immediately at netadmin@patni.com and delete this mail. _____________________________________________________________________ ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From Alexis.Tourapis dolby.net Wed Mar 1 10:46:54 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Wed Mar 1 18:04:11 2006 Subject: [Mp4-tech] [H.264]temporal direct luma motion vector Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01D698C6@sapphire.dolby.net> Dear Eric, Decision of temporal vs. spatial direct mode is an encoder decision. One may choose to use only temporal direct (i.e. could at times simplify mode decision, good performance at static scenes or when motion is continuous and consistent, extremely low QPs), use only spatial (considerably lower memory requirements, under most cases better performance than temporal, can better handle hierarchical picture coding etc), or optimally determine the best mode for each slice. You may have a look at this paper for the latter, although note that various enhancements in the software (i.e. bug fixing within the RDO and other processes, support for bipredictive ME and Weighted Prediction, hierarchical coding etc) have probably skewed the performance in favor of spatial direct : A.M. Tourapis, F. Wu, S. Li, "Direct mode coding for bipredictive slices in the H.264 standard," in IEEE Transactions on Circuits and Systems for Video Technology, Volume 15, Issue 1, pp.119-126, Jan. 2005 If your question is more in terms of direct modes, occurrence depends on content but also on quantization level/bitrate target and in general can be quite high. Best regards, Alexis ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Wednesday, March 01, 2024 12:00 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264]temporal direct luma motion vector Dear experts, Do you have any idea how often temporal direct luma motion vector is used in H.264 ? Thanks, Eric ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From singer apple.com Wed Mar 1 11:53:31 2006 From: singer apple.com (Dave Singer) Date: Wed Mar 1 18:04:15 2006 Subject: [Mp4-tech] JPEG - standard In-Reply-To: <20060301104654.29012.qmail@web8503.mail.in.yahoo.com> References: <20060301104654.29012.qmail@web8503.mail.in.yahoo.com> Message-ID: At 2:46 -0800 1/03/06, venkata sunil wrote: >Hi All, >I am looking for JPEG related documents, like > >1. ITU-T Recommendation T.81 | ISO/IEC 10918-1:1994, >Information technology - Digital compression >and coding of continuous-tone still images: >Requirements and guidelines. > >2. ITU-T Recommendation T.88 | ISO/IEC 14492-1, >Lossy/lossless coding of bi-level images. > >3. ITU-T Recommendation T.83 | ISO/IEC 10918-2:1995, >Information technology - Digital compression >and coding of continuous-tone still images: Compliance >testing. > >4. ITU-T Recommendation T.87 | ISO/IEC 14495-1, >Lossless and near-lossless compression of >continuous-tone still images-baseline. > >Please let me know where can i get them. Um, maybe from the ITU at ? > >Thanks in Advance, >Sunil. > > >__________________________________________________ >Do You Yahoo!? >Tired of spam? Yahoo! Mail has the best spam protection around >http://mail.yahoo.com >_______________________________________________ >NOTE: Please use clear subject lines for your posts. Include [audio, >[video], [systems], [general] or another apppropriate identifier to >indicate the type of question you have. > >Note: Conduct on the mailing list is subject to the Antitrust >guidelines found at >http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -- David Singer Apple Computer/QuickTime From garysull windows.microsoft.com Wed Mar 1 12:02:07 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Wed Mar 1 18:04:19 2006 Subject: [Mp4-tech] JPEG - standard In-Reply-To: <20060301104654.29012.qmail@web8503.mail.in.yahoo.com> Message-ID: <91D7F2CEE3425A4A9D11311D09FCE24613460F2E@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Sunil et al, These can be found at http://www.itu.int/rec/T-REC-T.81, etc. Note that you can get three ITU standards per year for free using the following link: http://ecs.itu.ch/cgi-bin/ebookshop Since these are also ISO/IEC standards, you can also get them through the ISO publication site, although they will be more expensive there. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of venkata sunil +> Sent: Wednesday, March 01, 2024 2:47 AM +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] JPEG - standard +> +> Hi All, +> I am looking for JPEG related documents, like +> +> 1. ITU-T Recommendation T.81 | ISO/IEC 10918-1:1994, +> Information technology - Digital compression +> and coding of continuous-tone still images: +> Requirements and guidelines. +> +> 2. ITU-T Recommendation T.88 | ISO/IEC 14492-1, +> Lossy/lossless coding of bi-level images. +> +> 3. ITU-T Recommendation T.83 | ISO/IEC 10918-2:1995, +> Information technology - Digital compression +> and coding of continuous-tone still images: Compliance +> testing. +> +> 4. ITU-T Recommendation T.87 | ISO/IEC 14495-1, +> Lossless and near-lossless compression of +> continuous-tone still images-baseline. +> +> +> Please let me know where can i get them. +> +> Thanks in Advance, +> Sunil. +> +> +> __________________________________________________ +> Do You Yahoo!? +> Tired of spam? Yahoo! Mail has the best spam protection around +> http://mail.yahoo.com +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From shiva_chikkalli yahoo.com Wed Mar 1 20:48:59 2006 From: shiva_chikkalli yahoo.com (Shivalingappa Chikkalli) Date: Thu Mar 2 09:46:10 2006 Subject: [Mp4-tech] JPEG - standard In-Reply-To: <20060301104654.29012.qmail@web8503.mail.in.yahoo.com> Message-ID: <20060302044859.98196.qmail@web52509.mail.yahoo.com> Hi Sunil, You can get it ftom ITU site http://www.itu.int/rec/recommendation.asp?type=products&lang=e&parent=T-REC-T by either purchage or as a individual if you regrister, you are allowed to download three standards, Regards, Shiva Chikkalli --- venkata sunil wrote: > Hi All, > I am looking for JPEG related documents, like > > 1. ITU-T Recommendation T.81 | ISO/IEC 10918-1:1994, > Information technology - Digital compression > and coding of continuous-tone still images: > Requirements and guidelines. > > 2. ITU-T Recommendation T.88 | ISO/IEC 14492-1, > Lossy/lossless coding of bi-level images. > > 3. ITU-T Recommendation T.83 | ISO/IEC 10918-2:1995, > Information technology - Digital compression > and coding of continuous-tone still images: > Compliance > testing. > > 4. ITU-T Recommendation T.87 | ISO/IEC 14495-1, > Lossless and near-lossless compression of > continuous-tone still images-baseline. > > > Please let me know where can i get them. > > Thanks in Advance, > Sunil. > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam > protection around > http://mail.yahoo.com > _______________________________________________ > NOTE: Please use clear subject lines for your posts. > Include [audio, [video], [systems], [general] or > another apppropriate identifier to indicate the type > of question you have. > > Note: Conduct on the mailing list is subject to the > Antitrust guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From pankaj_bajpai_iet operamail.com Thu Mar 2 06:00:47 2006 From: pankaj_bajpai_iet operamail.com (pankaj bajpai) Date: Thu Mar 2 09:46:14 2006 Subject: [Mp4-tech] Re: Mp4-tech Digest, Vol 32, Issue 4 Message-ID: <20060302050047.43E8323D22@ws5-3.us4.outblaze.com> Hi Sunil, I think the documents T.81 and T.83 are available at ITU . for T.81 the other link : www.w3.org/Graphics/JPEG/itu-t81.pdf with regards > From: venkata sunil > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] JPEG - standard > Date: Wed, 1 Mar 2024 02:46:54 -0800 (PST) > > > Hi All, > I am looking for JPEG related documents, like > > 1. ITU-T Recommendation T.81 | ISO/IEC 10918-1:1994, > Information technology - Digital compression > and coding of continuous-tone still images: > Requirements and guidelines. > > 2. ITU-T Recommendation T.88 | ISO/IEC 14492-1, > Lossy/lossless coding of bi-level images. > > 3. ITU-T Recommendation T.83 | ISO/IEC 10918-2:1995, > Information technology - Digital compression > and coding of continuous-tone still images: Compliance > testing. > > 4. ITU-T Recommendation T.87 | ISO/IEC 14495-1, > Lossless and near-lossless compression of > continuous-tone still images-baseline. > > > Please let me know where can i get them. > > Thanks in Advance, > Sunil. > > ----- Original Message ----- > From: mp4-tech-request@lists.mpegif.org > To: mp4-tech@lists.mpegif.org > Subject: Mp4-tech Digest, Vol 32, Issue 4 > Date: Wed, 1 Mar 2024 09:07:59 -0500 (EST) > > > Send Mp4-tech mailing list submissions to > mp4-tech@lists.mpegif.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.mpegif.org/mailman/listinfo/mp4-tech > or, via email, send a message with subject or body 'help' to > mp4-tech-request@lists.mpegif.org > > You can reach the person managing the list at > mp4-tech-owner@lists.mpegif.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Mp4-tech digest..." > > Today's Topics: > > 1. JPEG - standard (venkata sunil) > 2. RE: Question regarding reference list in B Slice > (Ye-Kui.Wang@nokia.com) > > From: venkata sunil > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] JPEG - standard > Date: Wed, 1 Mar 2024 02:46:54 -0800 (PST) > > > Hi All, > I am looking for JPEG related documents, like > > 1. ITU-T Recommendation T.81 | ISO/IEC 10918-1:1994, > Information technology - Digital compression > and coding of continuous-tone still images: > Requirements and guidelines. > > 2. ITU-T Recommendation T.88 | ISO/IEC 14492-1, > Lossy/lossless coding of bi-level images. > > 3. ITU-T Recommendation T.83 | ISO/IEC 10918-2:1995, > Information technology - Digital compression > and coding of continuous-tone still images: Compliance > testing. > > 4. ITU-T Recommendation T.87 | ISO/IEC 14495-1, > Lossless and near-lossless compression of > continuous-tone still images-baseline. > > > Please let me know where can i get them. > > Thanks in Advance, > Sunil. > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > > From: Ye-Kui.Wang@nokia.com > To: EricChuang@via.com.tw mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > Date: Wed, 1 Mar 2024 13:00:25 +0200 > > > Hi Eric, > > I guess the reason for a different initial lists construction for > pic_order_cnt_type equal to 0 than other values is that, there is > no good way to specify a POC value for non-existing pictures when > pic_order_cnt_type equasl to 0. > > BR, YK > > ________________________________ > > From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] > Sent: 2006”N3ŒŽ1“? 11:53 > To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > > > Hi YK, > According to 8.2.4.2.3 and 8.2.4.2.4 in standard, When > pic_order_cnt_type is equal to 0, reference pictures that are > marked as �gnon-existing�h are not included in either RefPicList0 > and RefPicList1. And with your description, it seems that when > pic_order_cnt_type not equal to 0, non-existing pictures shall be > put into RefList0 or RefList1. I just don�ft understand why it > differs depends on pic_order_cnt_type. > > Thanks, > Eric > > > -----Original Message----- > From: Ye-Kui.Wang@nokia.com [mailto:Ye-Kui.Wang@nokia.com] > Sent: Wednesday, March 01, 2024 1:10 AM > To: Eric Chuang; mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > > Eric, > > The initial reference lists for B slices are constructed using > POC. Non-existing pictrue should also be included in the initial > list such that the temporal base layer pictures can be decoded > identically when the temporal enhancement layer pictures are > present (in this case they are NOT "non-exisiting pictures") or not > (in this case they are "non-exisiting pictures"). However, POC > values of non-existing picures are not needed in the operaiton of > output order DPB (C.4.1-C.4.5) - I was wrong about this. > > BR, YK > > > ________________________________ > > From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] > Sent: 2006”N3ŒŽ1“? 10:04 > To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > Hi YK, > Thanks very much for your answering. > Regarding 3), it seems like that non-existing reference picture > shall not be used in display, right ? Why it needs POC ? > > Eric > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > Ye-Kui.Wang@nokia.com > Sent: Friday, February 24, 2024 4:42 AM > To: mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > > Hi Eric, > > 1. For temporal scalability. The reason is that temporal > enhancement layers may contain reference pictures and they may not > be present in the bitstream after stream thinning. > > 2. I think the latter is correct: the encoder has to make it > impossible to have non-existing pictures being included in the > lists. > > 3. POC is needed in constructing the initial reference picture > lists for B slices and is also needed in the operation of output > order DPB (C.4.1-C.4.5). > > BR, YK > > > ________________________________ > > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] > Sent: 2006”N2ŒŽ22“? 2:18 > To: garysull@windows.microsoft.com; mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] Question regarding reference list in B Slice > Dear experts, > In standard, it mentions the non-existing reference frame would > be put into DPB when frame number gap is allowed. I have 3 > questions about this. > 1. What�fs the usage of frame_num gap ? Why it�fs > introduced in standard ? Is there any document on the reason ? > 2. When constructing reference list for B Slice, the > standard describes that reference pictures marked as non-existing > are not included in RefList0 or RefList1, is this exclusion > operaton is done specially by decoder, or encoder has make it > impossible to happen by using reference picture list reordering ? > 3. Since non-existing reference picture cannot be put into > RefList0 or RefList1 in B slice, why standard specifies the default > value of delta_pic_order_cnt[0] and delta_pic_order_cnt[1] to them > in section 8.2.5.2. Is there any place to reference the POC of > non-existing pictures ? > > Thanks in advance, > Eric > > _______________________________________________ > Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to > indicate the type of question you have. > > Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > Pankaj Kumar Bajpai Mulitimedia Engineer India -- _______________________________________________ Surf the Web in a faster, safer and easier way: Download Opera 8 at http://www.opera.com Powered by Outblaze From pankaj_bajpai_iet operamail.com Thu Mar 2 06:05:22 2006 From: pankaj_bajpai_iet operamail.com (pankaj bajpai) Date: Thu Mar 2 09:46:18 2006 Subject: [Mp4-tech] Re: Mp4-tech Digest, Vol 32, Issue 2 Message-ID: <20060302050522.9170D7B527@ws5-10.us4.outblaze.com> Hi Prasanna, I am not very sure about JM encoder, but it is general practice to estimate whether the MB can be skip. The one rule is when Coeffs are zero and MVD is also zero. This you know after quantization. But during ME itself if you find that the for MV=0(or MVD=0) if SAD/MAD is very low (below some threshold) then you can classify it as SKIP With regards Pankaj > From: Prasanna Sethuraman > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] On MB skip in the JM encoder. > Date: Wed, 1 Mar 2024 14:01:23 +0530 > > > > Dear all, > > I was playing around with the JM encoder, and I find that it skips about > 1/3rd of the macro blocks. As far as I understand, a macro block can be > signaled as P_SKIP if the MVDs are zero and all the Quant. Coeffs are zero. > Now, the JM encoder also sets some of the Quant. Coeffs to zero based on > some thresholds. I disabled all that, but even after that, there is still a > huge number of MBs that are skipped. > > Does this mean that the JM encoder also uses some other criteria for MB > skip? If so, can someone hint me as to what that would be? Any information > is highly appreciated. > > And thanks to Alexis, Gary and Satendra for enlightening me on motion > vectors for chroma. > > Best regards, > Prasanna. > > > > > http://www.patni.com > World-Wide Partnerships. World-Class Solutions. > _____________________________________________________________________ > > This e-mail message may contain proprietary, confidential or legally > privileged information for the sole use of the person or entity to > whom this message was originally addressed. Any review, e-transmission > dissemination or other use of or taking of any action in reliance upon > this information by persons or entities other than the intended > recipient is prohibited. If you have received this e-mail in error > kindly delete this e-mail from your records. If it appears that this > mail has been forwarded to you without proper authority, please notify > us immediately at netadmin@patni.com and delete this mail. > _____________________________________________________________________ > > From: Ye-Kui.Wang@nokia.com > To: EricChuang@via.com.tw mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > Date: Wed, 1 Mar 2024 11:10:07 +0200 > > > Eric, > > The initial reference lists for B slices are constructed using POC. > Non-existing pictrue should also be included in the initial list > such that the temporal base layer pictures can be decoded > identically when the temporal enhancement layer pictures are > present (in this case they are NOT "non-exisiting pictures") or not > (in this case they are "non-exisiting pictures"). However, POC > values of non-existing picures are not needed in the operaiton of > output order DPB (C.4.1-C.4.5) - I was wrong about this. > > BR, YK > > ________________________________ > > From: ext EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] > Sent: 2006N31 10:04 > To: Wang Ye-Kui (Nokia-NRC/Tampere); mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > > > Hi YK, > Thanks very much for your answering. > Regarding 3), it seems like that non-existing reference picture > shall not be used in display, right ? Why it needs POC ? > > Eric > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > Ye-Kui.Wang@nokia.com > Sent: Friday, February 24, 2024 4:42 AM > To: mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Question regarding reference list in B Slice > > Hi Eric, > > 1. For temporal scalability. The reason is that temporal > enhancement layers may contain reference pictures and they may not > be present in the bitstream after stream thinning. > > 2. I think the latter is correct: the encoder has to make it > impossible to have non-existing pictures being included in the > lists. > > 3. POC is needed in constructing the initial reference picture > lists for B slices and is also needed in the operation of output > order DPB (C.4.1-C.4.5). > > BR, YK > > > ________________________________ > > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] > Sent: 2006N222 2:18 > To: garysull@windows.microsoft.com; mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] Question regarding reference list in B Slice > Dear experts, > In standard, it mentions the non-existing reference frame would > be put into DPB when frame number gap is allowed. I have 3 > questions about this. > 1. Whatfs the usage of frame_num gap ? Why itfs > introduced in standard ? Is there any document on the reason ? > 2. When constructing reference list for B Slice, the > standard describes that reference pictures marked as non-existing > are not included in RefList0 or RefList1, is this exclusion > operaton is done specially by decoder, or encoder has make it > impossible to happen by using reference picture list reordering ? > 3. Since non-existing reference picture cannot be put into > RefList0 or RefList1 in B slice, why standard specifies the default > value of delta_pic_order_cnt[0] and delta_pic_order_cnt[1] to them > in section 8.2.5.2. Is there any place to reference the POC of > non-existing pictures ? > > Thanks in advance, > Eric > > _______________________________________________ > Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to > indicate the type of question you have. > > Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > Pankaj Kumar Bajpai Mulitimedia Engineer India -- _______________________________________________ Surf the Web in a faster, safer and easier way: Download Opera 8 at http://www.opera.com Powered by Outblaze From mykerekes eml.cc Thu Mar 2 20:34:32 2006 From: mykerekes eml.cc (hungary) Date: Fri Mar 3 03:04:06 2006 Subject: [Mp4-tech] Intra Luma and Chroma Modes In-Reply-To: <20060302050522.9170D7B527@ws5-10.us4.outblaze.com> References: <20060302050522.9170D7B527@ws5-10.us4.outblaze.com> Message-ID: <1141360472.6886.255718581@webmail.messagingengine.com> Dear Experts, >From the H.264 standard i can see that there are 3 modes for intra luma coding (Latest standard) 4x4,8x8,16x16, and one intra chroma mode 8x8. Is there any specific reason for not introducing 8x8 and 16x16 for chroma. --Hungary -- hungary mykerekes@eml.cc -- http://www.fastmail.fm - A no graphics, no pop-ups email service From magarwal NeoMagic.com Fri Mar 3 10:41:32 2006 From: magarwal NeoMagic.com (Mohit Agarwal) Date: Fri Mar 3 03:04:11 2006 Subject: [Mp4-tech] [Video:H.264] Message-ID: <003c01c63e80$f17a06c0$931fa8c0@pcmagarwal> Can anybody tell me how to determine frame width and height in a raw .264 file. What I want to know is at what location in the bitstream is this value of width and height stored. Regards Mohit -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060303/2c7daeb4/attachment.html From mehmet.guney vestel.com.tr Fri Mar 3 10:54:32 2006 From: mehmet.guney vestel.com.tr (=?iso-8859-9?Q?Mehmet_G=FCney_=28Design_Engineer_-_VESTEK=29?=) Date: Fri Mar 3 05:40:06 2006 Subject: [Mp4-tech] [Video:H.264] Message-ID: <55E56B164BD80F47A7D46E241BF8B8220B95B6FB@zhavcexch1.zh.corp> Hi Mohit, Width can be extracted from the "pic_width_in_mbs_minus1" parameter of the sequence parameter set. width = (pic_width_in_mbs_minus1 + 1)*16; Similarly, height is included in "pic_height_in_map_units_minus1" parameter in SPS. PicHeightInMapUnits = pic_height_in_map_units_minus1 + 1; frameHeightInMbs = (2-frame_mbs_only_flag) * PicHeightInMapUnits; height = frameHeightInMbs * 16; Mehmet GUNEY mehmet.guney@vestel.com.tr VESTEL - Vestek Electronic R&D Corp. Istanbul/TURKEY ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Mohit Agarwal Sent: Friday, March 03, 2024 7:12 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [Video:H.264] Can anybody tell me how to determine frame width and height in a raw .264 file. What I want to know is at what location in the bitstream is this value of width and height stored. Regards Mohit -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060303/b66086c9/attachment.html From amittal NeoMagic.com Fri Mar 3 17:31:06 2006 From: amittal NeoMagic.com (Archna Mittal) Date: Fri Mar 3 08:34:06 2006 Subject: [Mp4-tech] [Video:H.264] In-Reply-To: <55E56B164BD80F47A7D46E241BF8B8220B95B6FB@zhavcexch1.zh.corp> Message-ID: Hi All, In addition to this, where (in which Atom) the ObjectTypeIndication stored for H.264 Video. and what is its value? If I parse the MP4 file containing AVC using YAMB , it gives ObjectTypeIndication = 0x21. According to ISO/IEC 14496-1:2002 , 0x21 is reserved. or is it I am looking at older version of the standard. Any pointers are welcome. Thanks in Advance, Regards, -Archna _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Mehmet G?ney (Design Engineer - VESTEK) Sent: Friday, March 03, 2024 2:25 PM To: mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] [Video:H.264] Hi Mohit, Width can be extracted from the ?pic_width_in_mbs_minus1? parameter of the sequence parameter set. width = (pic_width_in_mbs_minus1 + 1)*16; Similarly, height is included in ?pic_height_in_map_units_minus1? parameter in SPS. PicHeightInMapUnits = pic_height_in_map_units_minus1 + 1; frameHeightInMbs = (2-frame_mbs_only_flag) * PicHeightInMapUnits; height = frameHeightInMbs * 16; Mehmet GUNEY mehmet.guney@vestel.com.tr VESTEL - Vestek Electronic R&D Corp. Istanbul/TURKEY _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Mohit Agarwal Sent: Friday, March 03, 2024 7:12 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [Video:H.264] Can anybody tell me how to determine frame width and height in a raw .264 file. What I want to know is at what location in the bitstream is this value of width and height stored. Regards Mohit -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060303/b8441cd1/attachment.html From senthilr_pava rediffmail.com Fri Mar 3 13:53:42 2006 From: senthilr_pava rediffmail.com (senthil k) Date: Fri Mar 3 16:46:09 2006 Subject: [Mp4-tech] [MP4-audio]PNS testing Message-ID: <20060303135307.9682.qmail@webmail32.rediffmail.com> Hi experts! I am working on MPEG4 AAC-LC decoder.I need some information regarding testing PNS. I have done PNS decoding, But i don't know how to test it. I have the test vectors for PNS also. Even ISO reference software's output is not matching with reference output given by them? Can anyone tell me how to test the PNS? It will be very very helpful to me. Thanks senthil. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060303/419ea2d6/attachment.html From EricChuang via.com.tw Mon Mar 6 08:45:03 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Mon Mar 6 06:34:06 2006 Subject: [Mp4-tech] [H.264] Motion vector count Message-ID: <6A7AB4426712514F9E2AB28C407285FA372D8A@exchtp02.vntek.com.tw> Dear experts, Could anyone tell me what's the maximum possible motion vector count of a macroblock ? I read a paper, and it reads the maximum is 16. But from the standard, if we split the macroblock to 16 4x4 submacroblock partition with bi-direction prediction, there shall be 32 motion vectors. Does the standard define the limit ? Another question is what's the usage of subMvCnt mentioned in standard ? In deriving motion vectors, the standard mentioned how to compute subMvCnt, but I did not found the usage of it. Thanks very much, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060306/9820aea5/attachment.html From bheemarjun gmail.com Mon Mar 6 12:38:41 2006 From: bheemarjun gmail.com (Bheemarjuna Reddy Tamma) Date: Mon Mar 6 06:34:11 2006 Subject: [Mp4-tech] API to read/write YUV files Message-ID: <3fadd97c0603052308u79d3c566v90a760b4953a899c@mail.gmail.com> Hi, I am looking for an API, which can do the following: yuv_read(frame_num) and yuv_write(frame_write), ie. an API that can read and write YUV files. For example libmpeg3 can _read_ mpeg3 into yuv format (however lacks a similar support for writing into mpeg3 format). I want to be able to read and write YUV files. Is there a preexisting library/api for this ? -- T. Bheemarjuna Reddy Ph.D. Student HPCN Lab Dept. of Computer Science & Engg. IIT Madras, INDIA Mobile: +91-9444206507 http://hpcn.cs.iitm.ernet.in From pbijesh yahoo.co.in Mon Mar 6 09:51:37 2006 From: pbijesh yahoo.co.in (Bijesh) Date: Mon Mar 6 06:34:15 2006 Subject: [Mp4-tech] [MP4-audio]PNS testing In-Reply-To: <20060303135307.9682.qmail@webmail32.rediffmail.com> Message-ID: <20060306095137.13834.qmail@web8411.mail.in.yahoo.com> Hi Senthil, The criteria for PNS conformance is different from others. The implementation has to meet 1) Spectral Conformance 2) Temporal Conformance These are basically tests for the mean, variance and total enregy of the generated noise. You can get the details from ISO-IEC 14496-4. The tools will be available from 'audio/conftools/pns' of Reference Software. Regards, Bijesh --- senthil k wrote: > > Hi experts! > > I am working on MPEG4 AAC-LC decoder.I need some > information regarding testing PNS. > > I have done PNS decoding, But i don't know how to > test it. > > I have the test vectors for PNS also. > > Even ISO reference software's output is not > matching with reference output given by them? > > Can anyone tell me how to test the PNS? > > It will be very very helpful to me. > > Thanks > senthil.> _______________________________________________ > NOTE: Please use clear subject lines for your posts. > Include [audio, [video], [systems], [general] or > another apppropriate identifier to indicate the type > of question you have. > > Note: Conduct on the mailing list is subject to the > Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php ____________________________________________________ Jiyo cricket on Yahoo! India cricket http://in.sports.yahoo.com/cricket/ From jc sj.co.uk Mon Mar 6 12:33:23 2006 From: jc sj.co.uk (John Cox) Date: Mon Mar 6 15:28:06 2006 Subject: [Mp4-tech] API to read/write YUV files In-Reply-To: <3fadd97c0603052308u79d3c566v90a760b4953a899c@mail.gmail.com> References: <3fadd97c0603052308u79d3c566v90a760b4953a899c@mail.gmail.com> Message-ID: Hi YUV files are so trivial that you can write your own API in an hour or so. They simply go: Frame0 Y Samples (in raster scan order) Frame0 U Samples Frame0 V Samples Frame1 Y Samples Frame1 U Samples Frame1 V Samples Frame2 ... There is no header or anything so you need to know in advance what the size of a frame is. Beware that with 4:2:0 video the U sample block will be a quarter the size of the Y block, likewise V. Hope that helps John Cox SJ Consulting >Hi, >I am looking for an API, which can do the following: > >yuv_read(frame_num) and yuv_write(frame_write), ie. an API that can >read and write YUV files. For example libmpeg3 can _read_ mpeg3 into >yuv format (however lacks a similar support for writing into mpeg3 >format). I want to be able to read and write YUV files. Is there a >preexisting library/api for this ? From garysull windows.microsoft.com Mon Mar 6 11:03:56 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Mon Mar 6 15:28:11 2006 Subject: [Mp4-tech] [H.264] Motion vector count In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA372D8A@exchtp02.vntek.com.tw> Message-ID: <91D7F2CEE3425A4A9D11311D09FCE246135B4906@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Eric et al, The answer is level-dependent and the limit is applied to each two consecutive macroblocks in the bitstream. See the use of the variable MaxMvsPer2Mb in the standard. subMvCnt is summed to compute MvCnt. Then, in some levels, the sum of MvCnt for two consecutive macroblocks is constrained as specified in Annex A. Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Sunday, March 05, 2024 4:45 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] Motion vector count Dear experts, Could anyone tell me what's the maximum possible motion vector count of a macroblock ? I read a paper, and it reads the maximum is 16. But from the standard, if we split the macroblock to 16 4x4 submacroblock partition with bi-direction prediction, there shall be 32 motion vectors. Does the standard define the limit ? Another question is what's the usage of subMvCnt mentioned in standard ? In deriving motion vectors, the standard mentioned how to compute subMvCnt, but I did not found the usage of it. Thanks very much, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060306/12a14cae/attachment.html From EricChuang via.com.tw Tue Mar 7 07:38:52 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Tue Mar 7 12:16:08 2006 Subject: [Mp4-tech] [H.264] Motion vector count Message-ID: <6A7AB4426712514F9E2AB28C407285FA2849DF@exchtp02.vntek.com.tw> Hi Gary, In standard, only level 3~5.1 defines the MaxMvsPer2Mb, but level 1~2.2 doesn't. Does it mean there is no limit for level 1~2.2 ? I don't understand why the max mv count is defined for consecutive 2 Mbs, instead for each Mb ? Could you tell me why ? Thanks, Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gary Sullivan Sent: Monday, March 06, 2024 11:04 AM To: Eric Chuang; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] [H.264] Motion vector count Eric et al, The answer is level-dependent and the limit is applied to each two consecutive macroblocks in the bitstream. See the use of the variable MaxMvsPer2Mb in the standard. subMvCnt is summed to compute MvCnt. Then, in some levels, the sum of MvCnt for two consecutive macroblocks is constrained as specified in Annex A. Best Regards, Gary Sullivan _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Sunday, March 05, 2024 4:45 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] Motion vector count Dear experts, Could anyone tell me what's the maximum possible motion vector count of a macroblock ? I read a paper, and it reads the maximum is 16. But from the standard, if we split the macroblock to 16 4x4 submacroblock partition with bi-direction prediction, there shall be 32 motion vectors. Does the standard define the limit ? Another question is what's the usage of subMvCnt mentioned in standard ? In deriving motion vectors, the standard mentioned how to compute subMvCnt, but I did not found the usage of it. Thanks very much, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060307/3b726769/attachment.html From Danijel.Domazet zg.t-com.hr Tue Mar 7 16:06:38 2006 From: Danijel.Domazet zg.t-com.hr (Danijel Domazet) Date: Tue Mar 7 12:16:11 2006 Subject: [Mp4-tech] [Audio][HE AAC] Concept of "inverse filtering"? References: <3fadd97c0603052308u79d3c566v90a760b4953a899c@mail.gmail.com> Message-ID: <000801c641f8$bc105680$b700a8c0@FERGUSON> Hi, Can someone please explain the concept of "inverse filtering" and how it applies to SBR encoding in HE AAC? Why is it used? How does it improve quality (does it?)? How do we esitamate it's level (needed by sbr decoder)? Any web links are also welcome. Thanks, Daniel From Danijel.Domazet zg.t-com.hr Tue Mar 7 19:10:45 2006 From: Danijel.Domazet zg.t-com.hr (Danijel Domazet) Date: Wed Mar 8 09:10:07 2006 Subject: [Mp4-tech] [Audio][HE AAC] Concept of "inverse filtering"? Message-ID: <002a01c64212$7480a3f0$b700a8c0@FERGUSON> Hi, Can someone please explain the concept of "inverse filtering" and how it applies to SBR encoding used in HE AAC audio codec? Why is it used? How does it improve quality (does it?)? How do we esitamate it's level (needed by sbr decoder)? Any web links are also welcome. Thanks, Daniel From gilda apple.com Tue Mar 7 10:23:12 2006 From: gilda apple.com (Gilda Montesino) Date: Wed Mar 8 09:10:11 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA Message-ID: Apple's Interactive Media Group is hiring. If you or anyone you know sounds like a good match for the position below, please forward a resume directly to gilda@apple.com. We are looking for a bright and talented software engineer to focus on Audio Codec development, especially on AAC. The candidate should have strong programing skills in implementing audio compression algorithms. This individual should also be able to assess the sound quality and analyze the encoded audio, to figure out the cause of artifacts (e.g., frequency dropouts, lack of frequency resolution for harmonics, ringing, etc.), to run effective test to qualify codecs, and to develop tools (in C, scripting languages, or matlab) to perform these tasks. It's a plus but not a requirement for the person to have a solid background in psycho-acoustics or human auditory perception. The candidate should have practical implementation experience, discriminating hearing skills as well as strong analysis and programming skills. - BS or MS in CS or EE, with solid background in the field of Digital Signal Processing. - Expertise in Audio signal analysis and subjective/objective audio quality assessment. - Experience in implementing AAC or other psycho-acoustic model based audio codecs - Excellent coding skills in C, C++, with solid track records of code optimization. ---------------------- From gemini.bdr gmail.com Wed Mar 8 16:19:26 2006 From: gemini.bdr gmail.com (bdr bdr) Date: Wed Mar 8 09:10:15 2006 Subject: [Mp4-tech] hierarchical JPEG codec Message-ID: Dear experts, iam looking for source codes of hierarchical JPEG codec , test images &documents related to hierarchical mode jpeg , plz suggest me some links and test images 2) And also let me know how can we view jpeg-ls images(tool). Thanks Gemini -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060308/865f8285/attachment-0001.html From mykerekes eml.cc Wed Mar 8 06:25:36 2006 From: mykerekes eml.cc (hungary) Date: Wed Mar 8 12:04:06 2006 Subject: [Mp4-tech] Intra Luma and Chroma Modes In-Reply-To: <1141360472.6886.255718581@webmail.messagingengine.com> References: <20060302050522.9170D7B527@ws5-10.us4.outblaze.com> <1141360472.6886.255718581@webmail.messagingengine.com> Message-ID: <1141827936.24314.256127302@webmail.messagingengine.com> Dear Experts, I am putting the question again here. From the latest H.264 standard i can see that there are 3 modes for intra luma coding 4x4,8x8,16x16, and one intra chroma mode 8x8. Is there any specific reason for not introducing 4x4 for chroma. --Hungary -- hungary mykerekes@eml.cc -- http://www.fastmail.fm - A no graphics, no pop-ups email service From Alexis.Tourapis dolby.net Wed Mar 8 10:48:01 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Thu Mar 9 03:46:09 2006 Subject: [Mp4-tech] Intra Luma and Chroma Modes Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01D6A022@sapphire.dolby.net> Dear Hungary, Maybe it is better to call these as prediction classes instead of modes considering that for each "class" or prediction size you have different prediction modes as well. In any case, you should keep in mind that in H.264/AVC intra prediction can be quite cumbersome, especially if such is done using 4x4 block sizes. Also, for 4:2:0 content, chroma components tend to be usually a little bit more "saturated" and less "textured" than luma while the percentage of the bitrate used by chroma tends to be usually considerably low compared to the overall rate. In fact, if you try using only "DC" prediction for such content you may see that usually performance is only slightly affected. Support for extra chroma prediction modes could also add extra overhead in terms of bitrate, obvious additional complexity at the decoder, which may not be justified with gains in prediction. Nevertheless, note that currently for 4:4:4 encoding there is some activity in terms of using similar prediction modes for chroma (or better say other color components since such may involve different color transforms) as is done for luma/primary color component. You may find some interesting research documents especially in the Poznan meeting of last year if you go to the jvt ftp site: http://ftp3.itu.ch/av-arch/jvt-site/ Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of hungary Sent: Wednesday, March 08, 2024 6:26 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Intra Luma and Chroma Modes Dear Experts, I am putting the question again here. From the latest H.264 standard i can see that there are 3 modes for intra luma coding 4x4,8x8,16x16, and one intra chroma mode 8x8. Is there any specific reason for not introducing 4x4 for chroma. --Hungary -- hungary mykerekes@eml.cc -- http://www.fastmail.fm - A no graphics, no pop-ups email service _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From garysull windows.microsoft.com Wed Mar 8 12:38:15 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Thu Mar 9 03:46:14 2006 Subject: [Mp4-tech] Intra Luma and Chroma Modes In-Reply-To: <1141827936.24314.256127302@webmail.messagingengine.com> Message-ID: <91D7F2CEE3425A4A9D11311D09FCE246136AED14@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> +> Is there any specific reason for not introducing 4x4 for chroma. The idea was just to avoid unnecessary features and complexity. Chroma tends to consume a relatively-small percentage of the total bit rate, so there is not much overall gain that can be obtained by throwing more computational resources at it. Notice that we also use sloppier motion-compensation interpolation for chroma, and simplified deblocking too. (These remarks should be taken in the context of an assumption of 4:2:0 chroma format.) Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of hungary +> Sent: Wednesday, March 08, 2024 6:26 AM +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] Intra Luma and Chroma Modes +> +> Dear Experts, +> +> I am putting the question again here. +> +> From the latest H.264 standard i can see that there are 3 modes for +> intra luma +> coding 4x4,8x8,16x16, and one intra chroma mode 8x8. +> +> Is there any specific reason for not introducing 4x4 for chroma. +> +> --Hungary +> -- +> hungary +> mykerekes@eml.cc +> +> -- +> http://www.fastmail.fm - A no graphics, no pop-ups email service +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From kjk mcubeworks.com Thu Mar 9 10:04:48 2006 From: kjk mcubeworks.com (JK(Jeong Kwon) Kim) Date: Thu Mar 9 03:46:18 2006 Subject: [Mp4-tech] Intra Luma and Chroma Modes In-Reply-To: <1141827936.24314.256127302@webmail.messagingengine.com> Message-ID: <200603090112.k291CmwX015438@ns.mcubeworks.com> In compression with prediction information, there is always a trade-off between better prediction and bit-burden of info. The following factors can be considered in this case: - Adoption of 4x4 chroma results in more bits for prediction mode. - Chroma is likely to have the more low-frequency than luma. - So, the prediction gain is not large enough to compensate bits for mode. If other opinions, take easy to mail me. JK -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of hungary Sent: Wednesday, March 08, 2024 11:26 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Intra Luma and Chroma Modes Dear Experts, I am putting the question again here. From the latest H.264 standard i can see that there are 3 modes for intra luma coding 4x4,8x8,16x16, and one intra chroma mode 8x8. Is there any specific reason for not introducing 4x4 for chroma. --Hungary -- hungary mykerekes@eml.cc -- http://www.fastmail.fm - A no graphics, no pop-ups email service _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From abhishek.khandelwal patni.com Thu Mar 9 13:05:15 2006 From: abhishek.khandelwal patni.com (Abhishek Khandelwal) Date: Thu Mar 9 03:46:22 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA In-Reply-To: Message-ID: <200603090733.k297XOsP010590@pcsmail.patni.com> Dear Glida, I understand that this mailing list is for the technical discussions and not for offering "career opportunities", I believe we should use it only for technical discussions. Regards Abhishek Khandelwal -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gilda Montesino Sent: Tuesday, March 07, 2024 11:53 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA Apple's Interactive Media Group is hiring. If you or anyone you know sounds like a good match for the position below, please forward a resume directly to gilda@apple.com. We are looking for a bright and talented software engineer to focus on Audio Codec development, especially on AAC. The candidate should have strong programing skills in implementing audio compression algorithms. This individual should also be able to assess the sound quality and analyze the encoded audio, to figure out the cause of artifacts (e.g., frequency dropouts, lack of frequency resolution for harmonics, ringing, etc.), to run effective test to qualify codecs, and to develop tools (in C, scripting languages, or matlab) to perform these tasks. It's a plus but not a requirement for the person to have a solid background in psycho-acoustics or human auditory perception. The candidate should have practical implementation experience, discriminating hearing skills as well as strong analysis and programming skills. - BS or MS in CS or EE, with solid background in the field of Digital Signal Processing. - Expertise in Audio signal analysis and subjective/objective audio quality assessment. - Experience in implementing AAC or other psycho-acoustic model based audio codecs - Excellent coding skills in C, C++, with solid track records of code optimization. ---------------------- _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From cjj730 hotmail.com Thu Mar 9 09:52:30 2006 From: cjj730 hotmail.com (ChenJianjun) Date: Thu Mar 9 06:34:06 2006 Subject: [Mp4-tech] subMvCnt in temporal B_Direct mode Message-ID: Dear All: I have a question about subMvCnt, in SPEC 8.4.1.2, For Temporal B_Direct mode, it says"If subMbPartIdx is equal to 0, subMvCnt is equal to 2, otherwise(subMbPartIdx is not equal to 0), subMvCnt is set equal to 0" Does it mean only first subMbPartIdx need do inter prediction, and the other subMbPart in this macroblock partition need not do inter prediction since their subMvCnt=0? I think my understanding is probably wrong, but what does subMvCnt mean? Does per 4x4 block in B_Direct macroblock can has different MV value? Thanks _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn From mp4-tech-owner lists.mpegif.org Thu Mar 9 12:22:53 2006 From: mp4-tech-owner lists.mpegif.org (MPEGIF List Bounces) Date: Thu Mar 9 06:40:07 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA In-Reply-To: <200603090733.k297XOsP010590@pcsmail.patni.com> Message-ID: <003f01c6436b$cedd2940$0301a8c0@corp.intertrust.com> While the comment is appreciated, the Board of MPEGIF discussed the issue in the past and decided to allow these posts. Some members of this list may be looking for a job, and this information may be in their interest. Kind Regards, The list admins. > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > Abhishek Khandelwal > Sent: Thursday, 09 March 2024 08:35 > To: 'Gilda Montesino'; mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > Dear Glida, > > I understand that this mailing list is for the technical > discussions and not > for offering "career opportunities", I believe we should use > it only for > technical discussions. > > Regards > > Abhishek Khandelwal > > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gilda > Montesino > Sent: Tuesday, March 07, 2024 11:53 PM > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > > Apple's Interactive Media Group is hiring. If you or anyone you know > sounds like a good match for the position below, please forward a > resume directly to gilda@apple.com. > > We are looking for a bright and talented software engineer to focus > on Audio Codec development, especially on AAC. The candidate should > have strong programing skills in implementing audio compression > algorithms. This individual should also be able to assess the sound > quality and analyze the encoded audio, to figure out the cause of > artifacts (e.g., frequency dropouts, lack of frequency > resolution for > harmonics, ringing, etc.), to run effective test to qualify codecs, > and to develop tools (in C, scripting languages, or matlab) to > perform these tasks. It's a plus but not a requirement for > the person > to have a solid background in psycho-acoustics or human auditory > perception. The candidate should have practical implementation > experience, discriminating hearing skills as well as strong analysis > and programming skills. > > - BS or MS in CS or EE, with solid background in the field of > Digital > Signal Processing. > - Expertise in Audio signal analysis and subjective/objective audio > quality assessment. > - Experience in implementing AAC or other psycho-acoustic > model based > audio codecs > - Excellent coding skills in C, C++, with solid track records > of code > optimization. > > ---------------------- > > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate > identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php > > From mbp dgbmicro.com Thu Mar 9 17:00:37 2006 From: mbp dgbmicro.com (Mahantesh B Patil) Date: Thu Mar 9 09:22:08 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA In-Reply-To: <003f01c6436b$cedd2940$0301a8c0@corp.intertrust.com> Message-ID: then it means member's can post the req ???? -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org]On Behalf Of MPEGIF List Bounces Sent: Thursday, March 09, 2024 4:53 PM To: abhishek.khandelwal@patni.com; 'Gilda Montesino'; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA While the comment is appreciated, the Board of MPEGIF discussed the issue in the past and decided to allow these posts. Some members of this list may be looking for a job, and this information may be in their interest. Kind Regards, The list admins. > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > Abhishek Khandelwal > Sent: Thursday, 09 March 2024 08:35 > To: 'Gilda Montesino'; mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > Dear Glida, > > I understand that this mailing list is for the technical > discussions and not > for offering "career opportunities", I believe we should use > it only for > technical discussions. > > Regards > > Abhishek Khandelwal > > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gilda > Montesino > Sent: Tuesday, March 07, 2024 11:53 PM > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > > Apple's Interactive Media Group is hiring. If you or anyone you know > sounds like a good match for the position below, please forward a > resume directly to gilda@apple.com. > > We are looking for a bright and talented software engineer to focus > on Audio Codec development, especially on AAC. The candidate should > have strong programing skills in implementing audio compression > algorithms. This individual should also be able to assess the sound > quality and analyze the encoded audio, to figure out the cause of > artifacts (e.g., frequency dropouts, lack of frequency > resolution for > harmonics, ringing, etc.), to run effective test to qualify codecs, > and to develop tools (in C, scripting languages, or matlab) to > perform these tasks. It's a plus but not a requirement for > the person > to have a solid background in psycho-acoustics or human auditory > perception. The candidate should have practical implementation > experience, discriminating hearing skills as well as strong analysis > and programming skills. > > - BS or MS in CS or EE, with solid background in the field of > Digital > Signal Processing. > - Expertise in Audio signal analysis and subjective/objective audio > quality assessment. > - Experience in implementing AAC or other psycho-acoustic > model based > audio codecs > - Excellent coding skills in C, C++, with solid track records > of code > optimization. > > ---------------------- > > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate > identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php > > _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From sakthi.narayanan soc-soft.com Thu Mar 9 17:31:13 2006 From: sakthi.narayanan soc-soft.com (sakthi.narayanan@soc-soft.com) Date: Thu Mar 9 09:22:17 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA Message-ID: <4BF47D56A0DD2346A1B8D622C5C5902C014C35E7@soc-mail.soc-soft.com> Hi, There is nothing wrong to mailing the "career opportunities" in the MPEG forum.It will be very useful to know about the current requirements in the Branded company. With Regards, Sakthi -----Original Message----- From: MPEGIF List Bounces [mailto:mp4-tech-owner@lists.mpegif.org] Sent: Thursday, March 09, 2024 4:53 PM To: abhishek.khandelwal@patni.com; 'Gilda Montesino'; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA While the comment is appreciated, the Board of MPEGIF discussed the issue in the past and decided to allow these posts. Some members of this list may be looking for a job, and this information may be in their interest. Kind Regards, The list admins. > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > Abhishek Khandelwal > Sent: Thursday, 09 March 2024 08:35 > To: 'Gilda Montesino'; mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > Dear Glida, > > I understand that this mailing list is for the technical > discussions and not > for offering "career opportunities", I believe we should use > it only for > technical discussions. > > Regards > > Abhishek Khandelwal > > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gilda > Montesino > Sent: Tuesday, March 07, 2024 11:53 PM > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > > Apple's Interactive Media Group is hiring. If you or anyone you know > sounds like a good match for the position below, please forward a > resume directly to gilda@apple.com. > > We are looking for a bright and talented software engineer to focus > on Audio Codec development, especially on AAC. The candidate should > have strong programing skills in implementing audio compression > algorithms. This individual should also be able to assess the sound > quality and analyze the encoded audio, to figure out the cause of > artifacts (e.g., frequency dropouts, lack of frequency > resolution for > harmonics, ringing, etc.), to run effective test to qualify codecs, > and to develop tools (in C, scripting languages, or matlab) to > perform these tasks. It's a plus but not a requirement for > the person > to have a solid background in psycho-acoustics or human auditory > perception. The candidate should have practical implementation > experience, discriminating hearing skills as well as strong analysis > and programming skills. > > - BS or MS in CS or EE, with solid background in the field of > Digital > Signal Processing. > - Expertise in Audio signal analysis and subjective/objective audio > quality assessment. > - Experience in implementing AAC or other psycho-acoustic > model based > audio codecs > - Excellent coding skills in C, C++, with solid track records > of code > optimization. > > ---------------------- > > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate > identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti trust.php > > _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php The information contained in this e-mail message and in any annexure is confidential to the recipient and may contain privileged information. If you are not the intended recipient, please notify the sender and delete the message along with any annexure. You should not disclose, copy or otherwise use the information contained in the message or any annexure. Any views expressed in this e-mail are those of the individual sender except where the sender specifically states them to be the views of SoCrates Software India Pvt Ltd., Bangalore. From kaustubh.patankar vsnl.net Thu Mar 9 20:58:07 2006 From: kaustubh.patankar vsnl.net (kaustubh.patankar@vsnl.net) Date: Thu Mar 9 16:51:08 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA Message-ID: <7ca3b67c777c.7c777c7ca3b6@vsnl.net> Dear All In my view this should not be used as job portal. The forum admin please look into this. with regards Kaustubh ----- Original Message ----- From: sakthi.narayanan@soc-soft.com Date: Thursday, March 9, 2024 5:31 pm Subject: RE: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > Hi, > There is nothing wrong to mailing the "career opportunities" in the > MPEG forum.It will be very useful to know about the current > requirementsin the > Branded company. > > With Regards, > Sakthi > > > -----Original Message----- > From: MPEGIF List Bounces [mp4-tech-owner@lists.mpegif.org] > Sent: Thursday, March 09, 2024 4:53 PM > To: abhishek.khandelwal@patni.com; 'Gilda Montesino'; > mp4-tech@lists.mpegif.org > Subject: RE: [Mp4-tech] Career Opportunity with Apple in > Cupertino, CA > > While the comment is appreciated, the Board of MPEGIF discussed the > issue in > the past and decided to allow these posts. Some members of this > list may > be > looking for a job, and this information may be in their interest. > > Kind Regards, > The list admins. > > > -----Original Message----- > > From: mp4-tech-bounces@lists.mpegif.org > > [mp4-tech-bounces@lists.mpegif.org] On Behalf Of > > Abhishek Khandelwal > > Sent: Thursday, 09 March 2024 08:35 > > To: 'Gilda Montesino'; mp4-tech@lists.mpegif.org > > Subject: RE: [Mp4-tech] Career Opportunity with Apple in > Cupertino, CA > > > > Dear Glida, > > > > I understand that this mailing list is for the technical > > discussions and not > > for offering "career opportunities", I believe we should use > > it only for > > technical discussions. > > > > Regards > > > > Abhishek Khandelwal > > > > -----Original Message----- > > From: mp4-tech-bounces@lists.mpegif.org > > [mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gilda > > Montesino > > Sent: Tuesday, March 07, 2024 11:53 PM > > To: mp4-tech@lists.mpegif.org > > Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA > > > > > > Apple's Interactive Media Group is hiring. If you or anyone you > know > > sounds like a good match for the position below, please forward > a > > resume directly to gilda@apple.com. > > > > We are looking for a bright and talented software engineer to > focus > > on Audio Codec development, especially on AAC. The candidate > should > > have strong programing skills in implementing audio compression > > algorithms. This individual should also be able to assess the > sound > > quality and analyze the encoded audio, to figure out the cause > of > > artifacts (e.g., frequency dropouts, lack of frequency > > resolution for > > harmonics, ringing, etc.), to run effective test to qualify > codecs, > > and to develop tools (in C, scripting languages, or matlab) to > > perform these tasks. It's a plus but not a requirement for > > the person > > to have a solid background in psycho-acoustics or human auditory > > perception. The candidate should have practical implementation > > experience, discriminating hearing skills as well as strong > analysis > > and programming skills. > > > > - BS or MS in CS or EE, with solid background in the field of > > Digital > > Signal Processing. > > - Expertise in Audio signal analysis and subjective/objective > audio > > quality assessment. > > - Experience in implementing AAC or other psycho-acoustic > > model based > > audio codecs > > - Excellent coding skills in C, C++, with solid track records > > of code > > optimization. > > > > ---------------------- > > > > > > > > _______________________________________________ > > NOTE: Please use clear subject lines for your posts. Include [audio, > > [video], [systems], [general] or another apppropriate > > identifier to indicate > > the type of question you have. > > > > Note: Conduct on the mailing list is subject to the Antitrust > > guidelines > > found at > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti > trust.php > > > > > > _______________________________________________ > > NOTE: Please use clear subject lines for your posts. Include > > [audio, [video], [systems], [general] or another apppropriate > > identifier to indicate the type of question you have. > > > > Note: Conduct on the mailing list is subject to the Antitrust > > guidelines found at > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti > trust.php > > > > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to > indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelinesfound at > http://www.mpegif.org/public/documents/vault/mp-out-30042- > Antitrust.php > > > > > > The information contained in this e-mail message and in any > annexure is > confidential to the recipient and may contain privileged > information. If you are not > the intended recipient, please notify the sender and delete the > message along with > any annexure. You should not disclose, copy or otherwise use the > information contained > in the message or any annexure. Any views expressed in this e-mail > are those of the > individual sender except where the sender specifically states them > to be the views of > SoCrates Software India Pvt Ltd., Bangalore. > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include > [audio, [video], [systems], [general] or another apppropriate > identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042- > Antitrust.php From mp4-tech-owner lists.mpegif.org Thu Mar 9 17:40:49 2006 From: mp4-tech-owner lists.mpegif.org (MPEGIF List Admins) Date: Thu Mar 9 16:51:13 2006 Subject: [Mp4-tech] Career Opportunity with Apple in Cupertino, CA Message-ID: <004201c64398$3925b3e0$0301a8c0@corp.intertrust.com> Dear Kaustubh, all, As we have seen on the list today, there are arguments pro and con. Taking into account both sides of the discussion, the pstanding olicy has been determined by MPEGIF's Board some time ago. If we get too much emails like this, the policy can always be revised, but so far there has not been the need to do that. If you feel strongly about this, we recommend that you join MPEGIF and raise your voice as a member. Kind Regards, The admins. ps: This will end the discussion. The purpose of the list is not discussing list policy. > > -----Original Message----- > > From: mailman-bounces@lists.mpegif.org > > [mailto:mailman-bounces@lists.mpegif.org] On Behalf Of > > kaustubh.patankar@vsnl.net > > Sent: Thursday, 09 March 2024 16:58 > > To: sakthi.narayanan@soc-soft.com > > Cc: mp4-tech-owner@lists.mpegif.org; > > abhishek.khandelwal@patni.com; gilda@apple.com; > > mp4-tech@lists.mpegif.org > > Subject: Re: RE: [Mp4-tech] Career Opportunity with Apple in > > Cupertino, CA > > > > Dear All > > > > In my view this should not be used as job portal. The forum admin > > please look into this. > > > > with regards > > > > Kaustubh > > > > ----- Original Message ----- > > From: sakthi.narayanan@soc-soft.com > > Date: Thursday, March 9, 2024 5:31 pm > > Subject: RE: [Mp4-tech] Career Opportunity with Apple in > Cupertino, CA > > > > > > > > Hi, > > > There is nothing wrong to mailing the "career > opportunities" in the > > > MPEG forum.It will be very useful to know about the current > > > requirementsin the > > > Branded company. > > > > > > With Regards, > > > Sakthi > > > > > > > > > -----Original Message----- > > > From: MPEGIF List Bounces [mp4-tech-owner@lists.mpegif.org] > > > Sent: Thursday, March 09, 2024 4:53 PM > > > To: abhishek.khandelwal@patni.com; 'Gilda Montesino'; > > > mp4-tech@lists.mpegif.org > > > Subject: RE: [Mp4-tech] Career Opportunity with Apple in > > > Cupertino, CA > > > > > > While the comment is appreciated, the Board of MPEGIF > discussed the > > > issue in > > > the past and decided to allow these posts. Some members of this > > > list may > > > be > > > looking for a job, and this information may be in their interest. > > > > > > Kind Regards, > > > The list admins. > > > > > > > -----Original Message----- > > > > From: mp4-tech-bounces@lists.mpegif.org > > > > [mp4-tech-bounces@lists.mpegif.org] On Behalf Of > > > > Abhishek Khandelwal > > > > Sent: Thursday, 09 March 2024 08:35 > > > > To: 'Gilda Montesino'; mp4-tech@lists.mpegif.org > > > > Subject: RE: [Mp4-tech] Career Opportunity with Apple in > > > Cupertino, CA > > > > > > > > Dear Glida, > > > > > > > > I understand that this mailing list is for the technical > > > > discussions and not > > > > for offering "career opportunities", I believe we should use > > > > it only for > > > > technical discussions. > > > > > > > > Regards > > > > > > > > Abhishek Khandelwal > > > > > > > > -----Original Message----- > > > > From: mp4-tech-bounces@lists.mpegif.org > > > > [mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gilda > > > > Montesino > > > > Sent: Tuesday, March 07, 2024 11:53 PM > > > > To: mp4-tech@lists.mpegif.org > > > > Subject: [Mp4-tech] Career Opportunity with Apple in > Cupertino, CA > > > > > > > > > > > > Apple's Interactive Media Group is hiring. If you or anyone you > > > know > > > > sounds like a good match for the position below, please forward > > > a > > > > resume directly to gilda@apple.com. > > > > > > > > We are looking for a bright and talented software engineer to > > > focus > > > > on Audio Codec development, especially on AAC. The candidate > > > should > > > > have strong programing skills in implementing audio compression > > > > algorithms. This individual should also be able to assess the > > > sound > > > > quality and analyze the encoded audio, to figure out the cause > > > of > > > > artifacts (e.g., frequency dropouts, lack of frequency > > > > resolution for > > > > harmonics, ringing, etc.), to run effective test to qualify > > > codecs, > > > > and to develop tools (in C, scripting languages, or matlab) to > > > > perform these tasks. It's a plus but not a requirement for > > > > the person > > > > to have a solid background in psycho-acoustics or human > auditory > > > > perception. The candidate should have practical implementation > > > > experience, discriminating hearing skills as well as strong > > > analysis > > > > and programming skills. > > > > > > > > - BS or MS in CS or EE, with solid background in the field of > > > > Digital > > > > Signal Processing. > > > > - Expertise in Audio signal analysis and subjective/objective > > > audio > > > > quality assessment. > > > > - Experience in implementing AAC or other psycho-acoustic > > > > model based > > > > audio codecs > > > > - Excellent coding skills in C, C++, with solid track records > > > > of code > > > > optimization. > > > > > > > > ---------------------- > > > > > > > > > > > > > > > > _______________________________________________ > > > > NOTE: Please use clear subject lines for your posts. > > Include [audio, > > > > [video], [systems], [general] or another apppropriate > > > > identifier to indicate > > > > the type of question you have. > > > > > > > > Note: Conduct on the mailing list is subject to the Antitrust > > > > guidelines > > > > found at > > > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti > > > trust.php > > > > > > > > > > > > _______________________________________________ > > > > NOTE: Please use clear subject lines for your posts. Include > > > > [audio, [video], [systems], [general] or another apppropriate > > > > identifier to indicate the type of question you have. > > > > > > > > Note: Conduct on the mailing list is subject to the Antitrust > > > > guidelines found at > > > > http://www.mpegif.org/public/documents/vault/mp-out-30042-Anti > > > trust.php > > > > > > > > > > > > > > _______________________________________________ > > > NOTE: Please use clear subject lines for your posts. > Include [audio, > > > [video], [systems], [general] or another apppropriate > identifier to > > > indicate the type of question you have. > > > > > > Note: Conduct on the mailing list is subject to the Antitrust > > > guidelinesfound at > > > http://www.mpegif.org/public/documents/vault/mp-out-30042- > > > Antitrust.php > > > > > > > > > > > > > > > > > > The information contained in this e-mail message and in any > > > annexure is > > > confidential to the recipient and may contain privileged > > > information. If you are not > > > the intended recipient, please notify the sender and delete the > > > message along with > > > any annexure. You should not disclose, copy or otherwise use the > > > information contained > > > in the message or any annexure. Any views expressed in > this e-mail > > > are those of the > > > individual sender except where the sender specifically > states them > > > to be the views of > > > SoCrates Software India Pvt Ltd., Bangalore. > > > > > > _______________________________________________ > > > NOTE: Please use clear subject lines for your posts. Include > > > [audio, [video], [systems], [general] or another apppropriate > > > identifier to indicate the type of question you have. > > > > > > Note: Conduct on the mailing list is subject to the Antitrust > > > guidelines found at > > > http://www.mpegif.org/public/documents/vault/mp-out-30042- > > > Antitrust.php > > > > > > From Alexis.Tourapis dolby.net Fri Mar 10 09:45:50 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Fri Mar 10 21:04:07 2006 Subject: [Mp4-tech] subMvCnt in temporal B_Direct mode Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01DDE184@sapphire.dolby.net> Dear ChenJianjun subMvCnt only has to do with the computation of MvCnt in section 8.4 (inter prediction process). MvCnt is essentially a parameter used within Level Conformance to determine the number of motion vectors for a macroblock (in fact the number of motion vectors for two consecutive macroblocks). See sections A.3.1 and A.3.2 for more information. It has nothing to do with how inter prediction is done (i.e. it is not used anywhere within the prediction semantics as you can also observe from the text), but it does relate with the number of motion vectors this process uses (btw do note the constraint in terms of the use of direct_8x8_inference_flag for some levels which affects this process). Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ChenJianjun Sent: Thursday, March 09, 2024 1:53 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] subMvCnt in temporal B_Direct mode Dear All: I have a question about subMvCnt, in SPEC 8.4.1.2, For Temporal B_Direct mode, it says"If subMbPartIdx is equal to 0, subMvCnt is equal to 2, otherwise(subMbPartIdx is not equal to 0), subMvCnt is set equal to 0" Does it mean only first subMbPartIdx need do inter prediction, and the other subMbPart in this macroblock partition need not do inter prediction since their subMvCnt=0? I think my understanding is probably wrong, but what does subMvCnt mean? Does per 4x4 block in B_Direct macroblock can has different MV value? Thanks _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From lanxug 163.com Sat Mar 11 18:55:53 2006 From: lanxug 163.com (Xuguang Lan) Date: Sat Mar 11 14:40:08 2006 Subject: [Mp4-tech] IWICPAS, Springer publication, IJCM special issue Message-ID: <4412ACBA.036216.12605> Dear all, Sorry for sending CFP again. website: unit.xjtu.edu.cn/iwicpas Publication news 1.The conference proceedings for all the papers will be published as a volume in Lecture Notes in Computer Science. 2.A number of selected papers at the International Workshop on Intelligent Computing in Pattern Analysis/Synthesis (IWICPAS) will be invited (an updated version) to publish by International Journal of Computer Mathematics (Taylor & Francis) as a special issue. The journal is abstracted in a dozen of indices, including the ISI Science Citation Index Expanded. Paper submission: April 20,2024 Notification of acceptance: May 20, 2024 (modified): Camera-ready manuscript: June 20, 2023 (modified): IWICPAS 2006 The International Workshop on Intelligent Computing in Pattern Analysis/Synthesis August 26, 2006, Xi'an, China The International Workshop on Intelligent Computing in Pattern Analysis/Synthesis (IWICPAS), sponsored by IAPR and CAA, will be held in Xi'an, China on August 26, 2006. The IWICPAS brings together leading researchers and engineers from the fields of pattern analysis/synthesis around the world. This is an international forum for identifying, encouraging and exchanging new ideas on different topics of pattern analysis/synthesis as well as for promoting novel applications in an attempt to extend the frontiers of this fascinating field of research. The workshops will comprise of oral and poster presentations, and panel discussions. Topics of interest include, but are not limited to Pattern presentation and modeling Visual pattern modeling Object detection, tracking and recognition Statistical Motion analysis Texture analysis/synthesis Spatial-temporal pattern analysis/synthesis Sparse coding Scalable Image/Video coding Markov chain Monte Carlo computing Manifold Learning Partial Differential Equation -------------- next part -------------- A non-text attachment was scrubbed... Name: IWICPAS_CFP.pdf Type: application/octet-stream Size: 64893 bytes Desc: not available Url : /pipermail/mp4-tech/attachments/20060311/d90074b7/IWICPAS_CFP-0001.obj From garima.singh tivr.co.in Sat Mar 11 04:52:26 2006 From: garima.singh tivr.co.in (Garima Singh) Date: Sat Mar 11 14:40:12 2006 Subject: [Mp4-tech] [Audio]: mp4audec Conformance Issues In-Reply-To: Message-ID: <20060311125226.48975.qmail@web211.biz.mail.re2.yahoo.com> Hi All, If I am decoding the mp4 streams available at ftp://mpaudconf:adif2mp4@ftp.iis.fhg.de/ with the mp4audec.exe that I got after building reference software I am not getting output wav files bit exact with reference wav files available at the same location. What could be the reason? Regards, Garima Regards, Garima Singh TIVR Communications - Delivering Efficient Solutions for Mobile Multimedia www.tivr.co.in -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060311/8a5deb95/attachment.html From robin94539 yahoo.com Sun Mar 12 01:17:06 2006 From: robin94539 yahoo.com (Robin Zoo) Date: Sun Mar 12 20:28:07 2006 Subject: [Mp4-tech] MPEG1/2/3 standard and reference c-code In-Reply-To: <200509151607.j8FG5cHx010405@lists1.magma.ca> Message-ID: <20060312091706.34612.qmail@web52404.mail.yahoo.com> Hi All, Could someone tell me where I can find the following MPEG1/2/4 standards and reference c-code for free? ISO/IEC 11172 (part1-3) ISO/IEC 13818 (part1-3) ISO/IEC 14496 (part1-3) thanks ! Robin __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From cjj730 hotmail.com Mon Mar 13 02:54:44 2006 From: cjj730 hotmail.com (ChenJianjun) Date: Mon Mar 13 09:34:07 2006 Subject: [Mp4-tech] compute colocated macroblock Message-ID: Dear All: In JM reference decode code, it will use "RSD()" in compute_colocated() function, I dont know the meaning of RSD(), why should it do like this, I cant find the reason in SPEC. Could anyone tell me about it or is there any document I can read? Another question: In dpb_split_field() function, it will compute the MVs of colocated macroblocks for top_field and bottom_field pictures, but in compute_colocated() function, it will compute again, Is there any difference? Thanks very much. ChenJianjun. _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn From luwang astri.org Mon Mar 13 12:20:59 2006 From: luwang astri.org (Wang Lu) Date: Mon Mar 13 09:34:12 2006 Subject: [Mp4-tech] A question about the frame_num Message-ID: Dear experts, I'm reading the h264 profile, I have several questions: 1: Do the two fields of one frame must be complementary field pair? 2: Can the top field is a non-reference field but the bottom field is a reference field in one frame (non complementary field pair)? Thanks & Regards Lu This message (including any attachments) is for the named addressee(s)'s use only. It may contain sensitive, confidential, private proprietary or legally privileged information intended for a specific individual and purpose, and is protected by law. If you are not the intended recipient, please immediately delete it and all copies of it from your system, destroy any hard copies of it and notify the sender. Any use, disclosure, copying, or distribution of this message and/or any attachments is strictly prohibited. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060313/1222fa1e/attachment.html From audio sukhan.com Mon Mar 13 13:53:44 2006 From: audio sukhan.com (Shafi Ullah Khan) Date: Mon Mar 13 09:34:15 2006 Subject: [Mp4-tech] RMS conformance for 16 bit AAC LC Decoder Message-ID: <5545a8b30603130023y7dffbb0cv787e30d4d92f11a1@mail.gmail.com> I have doubt regarding test vector for RMS conformance of 16-bit AAC LC decoder. Does any AAC LC decoder with 16 bit output able to meet RMS requirements (2^(-15)/sqrt(12) == 8.810e-6 == -101.10 dB) for al00_08.wav and al00_32.wav reference output? For these two test vectors theoretically RMS cannot be less than 8.821e-6 and 8.817e-6 respectively for 16-bit decoder. shafi From Danijel.Domazet zg.t-com.hr Mon Mar 13 10:23:51 2006 From: Danijel.Domazet zg.t-com.hr (Danijel Domazet) Date: Mon Mar 13 13:46:07 2006 Subject: [Mp4-tech] off-topic: each mp4-tech message delivered twice!?! References: <20060311125226.48975.qmail@web211.biz.mail.re2.yahoo.com> Message-ID: <005601c6467f$d76ca3d0$5800a8c0@FERGUSON> Hi, Am I the only one who is getting all the mp4-tech list messages twice? Can someone fix this, please? Daniel From mp4-tech-owner lists.mpegif.org Mon Mar 13 10:27:52 2006 From: mp4-tech-owner lists.mpegif.org (MPEGIF List Admins) Date: Mon Mar 13 13:46:11 2006 Subject: [Mp4-tech] Double emails Message-ID: <000701c64680$67426300$0301a8c0@corp.intertrust.com> All, several people have notified us they are receiving double emails from the list. Many people have filtering of duplicate emails these days, so not everybody will have seen the problem. Anyway, we are looking into the issue (or will be later today; the servers are in Canada); thanks for letting us know. We hope to be able to resolve it soon. Best, the list admins. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060313/61988b7c/attachment.html From Alexis.Tourapis dolby.net Mon Mar 13 02:19:34 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Mon Mar 13 13:46:15 2006 Subject: [Mp4-tech] compute colocated macroblock Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01DDE310@sapphire.dolby.net> Dear ChenJianjun The RSD() function relates to the consideration of the direct_8x8_inference_flag. I would recommend studying JVT-C115r1.doc if you are curious about more details on this and to see a visual representation of what is done and why. This is basically what is done in section 8.4.1.2.1 (Derivation process for the co-located 4x4 sub-macroblock partitions) although without any of the explanations. dpb_split_field() only prepares motion vectors, that is it generates the corresponding field motion vectors from frame information (which can be quite complex for the MBAFF case). Note that this is done only once after a picture is encoded. Compute_colocated() on the other hand is called before a B slice is generated and prepares/selects the appropriate motion information to be used for the encoding of this slice. It is quite important to make this separation since, due to reordering or the capability of specifying any coding order in H.264/AVC, it is possible that any picture is used for the generation of these parameters. These parameters may also be used differently if there is a certain relationship between top and bottom fields and such are obviously unknown during preparation of the field motion info. Btw you could of course do the decisions in dpb_split_field within compute_colocated but it was found a bit cleaner to prepare the motion information in a field vs frame format after a picture is encoded than redoing everything for every new slice that you have to encode (remember also that you may have to do many slices). The code is by far not optimized (especially relating to memory operations), but well ... it works... Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ChenJianjun Sent: Sunday, March 12, 2024 6:55 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] compute colocated macroblock Dear All: In JM reference decode code, it will use "RSD()" in compute_colocated() function, I dont know the meaning of RSD(), why should it do like this, I cant find the reason in SPEC. Could anyone tell me about it or is there any document I can read? Another question: In dpb_split_field() function, it will compute the MVs of colocated macroblocks for top_field and bottom_field pictures, but in compute_colocated() function, it will compute again, Is there any difference? Thanks very much. ChenJianjun. _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From bheemarjun gmail.com Mon Mar 13 23:34:32 2006 From: bheemarjun gmail.com (Bheemarjuna Reddy Tamma) Date: Mon Mar 13 21:51:08 2006 Subject: [Mp4-tech] Can't decode this mp4 file, but mplayer plays it, why? Message-ID: <3fadd97c0603131004g6608a012hcb47067b1a8598b9@mail.gmail.com> Hi All, I am using ffmpeg for a research project in video multicasting over ad hoc wireless networks. I am facing the following problem: I am able to play the highly distorted mp4 files using mplayer, however when I try ffmpeg on them to decode them into yuv format (I need that to compute the psnr). It gives the following output : -- [devesh@mcn test_psnr]$ ~/ffmpeg/ffmpeg -i try_2_0.mp4 try.yuv FFmpeg version CVS, Copyright (c) 2000-2004 Fabrice Bellard configuration: libavutil version: 49.0.0 libavcodec version: 51.2.0 libavformat version: 50.2.1 built on Feb 23 2006 00:09:27, gcc: 3.3.3 20040412 (Red Hat Linux 3.3.3-7) Input #0, mov,mp4,m4a,3gp,3g2, from 'try_2_0.mp4': Duration: 00:00:13.3, start: 0.000000, bitrate: 379 kb/s Stream #0.0(und), 7.50 fps(r): Video: mpeg4, yuv420p Output #0, rawvideo, to 'try.yuv': Stream #0.0, 7.50 fps(c): Video: rawvideo, yuv420p, q=2-31, 200 kb/s Stream mapping: Stream #0.0 -> #0.0 [rawvideo @ 0x82fcd68]dimensions not set Could not write header for output file #0 (incorrect codec parameters ?) [devesh@mcn test_psnr]$ -- These files are available at http://www.cs.iitm.ernet.in/~dvagr/mp4files Mplayer can play each one of them, but ffmpeg doesn't work (ie gives the same error message as shown above), and xvid_decraw segfaults. These files were produced using the etmp4 tool, as given on the site http://www.tkn.tu-berlin.de/research/evalvid/EvalVid/docevalvid.html. However, ffmpeg is able to decode slighly less distorted files comfortably, it only seems to give up on more distorted files. If ffmpeg can't do the job, is there any way I can make mplayer/mencoder to dump yuv420p files from this mp4 video ?. Thanks, eagerly awaiting your reply. Any pointers/advise/help would be highly appreciated. regards Arjun Reddy -- T. Bheemarjuna Reddy Ph.D. Student HPCN Lab Dept. of Computer Science & Engg. IIT Madras, INDIA Mobile: +91-9444206507 http://hpcn.cs.iitm.ernet.in From mp4-tech-owner lists.mpegif.org Tue Mar 14 08:46:08 2006 From: mp4-tech-owner lists.mpegif.org (MPEGIF List Admins) Date: Tue Mar 14 10:04:06 2006 Subject: [Mp4-tech] Double emails In-Reply-To: <000701c64680$67426300$0301a8c0@corp.intertrust.com> Message-ID: <002801c6473b$5ba943e0$0301a8c0@corp.intertrust.com> All, Our mailing list service provider has indicated that the problem with duplicate emails has been resolved. Should you still receive duplicate emails, please send email to mp4-tech-owner@mpegif.org We apologize for the temporary inconvenience. Thanks, the list admins. _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of MPEGIF List Admins Sent: Monday, 13 March 2024 10:28 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Double emails All, several people have notified us they are receiving double emails from the list. Many people have filtering of duplicate emails these days, so not everybody will have seen the problem. Anyway, we are looking into the issue (or will be later today; the servers are in Canada); thanks for letting us know. We hope to be able to resolve it soon. Best, the list admins. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060314/7cfb1064/attachment.html From cjj730 hotmail.com Tue Mar 14 08:52:46 2006 From: cjj730 hotmail.com (ChenJianjun) Date: Tue Mar 14 10:04:11 2006 Subject: [Mp4-tech] compute colocated macroblock Message-ID: Dear Alexis: First, thank you for your reply. I still have doubt for B direct mode: if direct_8x8_inference_flag=1, does four 4x4 blocks in a 8x8 block share the same MV? Since their colocated MV seems same:select the same "x" 4x4 block in the following table: xo|ox oo|oo --|-- oo|oo xo|ox Thanks again! ChenJianjun ---------------------------------------- > Subject: RE: [Mp4-tech] compute colocated macroblock > Date: Mon, 13 Mar 2024 02:19:34 -0800 > From: Alexis.Tourapis@dolby.net > To: cjj730@hotmail.com; mp4-tech@lists.mpegif.org > > Dear ChenJianjun > > The RSD() function relates to the consideration of the direct_8x8_inference_flag. I would recommend studying JVT-C115r1.doc if you are curious about more details on this and to see a visual representation of what is done and why. This is basically what is done in section 8.4.1.2.1 (Derivation process for the co-located 4x4 sub-macroblock partitions) although without any of the explanations. > > dpb_split_field() only prepares motion vectors, that is it generates the corresponding field motion vectors from frame information (which can be quite complex for the MBAFF case). Note that this is done only once after a picture is encoded. Compute_colocated() on the other hand is called before a B slice is generated and prepares/selects the appropriate motion information to be used for the encoding of this slice. It is quite important to make this separation since, due to reordering or the capability of specifying any coding order in H.264/AVC, it is possible that any picture is used for the generation of these parameters. These parameters may also be used differently if there is a certain relationship between top and bottom fields and such are obviously unknown during preparation of the field motion info. Btw you could of course do the decisions in dpb_split_field within compute_colocated but it was found a bit cleaner to prepare the motion information in a field vs frame format after a picture is encoded than redoing everything for every new slice that you have to encode (remember also that you may have to do many slices). > > The code is by far not optimized (especially relating to memory operations), but well ... it works... > > Best regards, > > Alexis > > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ChenJianjun > Sent: Sunday, March 12, 2024 6:55 PM > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] compute colocated macroblock > > Dear All: > > In JM reference decode code, it will use "RSD()" in compute_colocated() function, I dont know the meaning of RSD(), why should it do like this, I cant find the reason in SPEC. Could anyone tell me about it or is there any document I can read? > Another question: In dpb_split_field() function, it will compute the MVs of colocated macroblocks for top_field and bottom_field pictures, but in compute_colocated() function, it will compute again, Is there any difference? > > Thanks very much. > ChenJianjun. > _________________________________________________________________ > 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: > http://messenger.msn.com/cn > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > > ----------------------------------------- > This message (including any attachments) may contain confidential > information intended for a specific individual and purpose. If you > are not the intended recipient, delete this message. If you are > not the intended recipient, disclosing, copying, distributing, or > taking any action based on this message is strictly prohibited. > _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn From Max.Griessl DynaPel.com Tue Mar 14 11:03:53 2006 From: Max.Griessl DynaPel.com (=?ISO-8859-1?Q?Max_Grie=DFl?=) Date: Tue Mar 14 10:28:07 2006 Subject: [Mp4-tech] Can't decode this mp4 file, but mplayer plays it, why? In-Reply-To: <3fadd97c0603131004g6608a012hcb47067b1a8598b9@mail.gmail.com> References: <3fadd97c0603131004g6608a012hcb47067b1a8598b9@mail.gmail.com> Message-ID: <44169509.2080800@DynaPel.com> Using the option -vo yuv4mpeg for mplayer a file stream.yuv is written. This file contains the raw yuv420 images. But every frame has a small header. These headers may be removed with the following command: perl -p -e's/YUV4MPEG.*\n//g' stream.yuv | perl -p -e's/FRAME\n//g' > video.yuv Regards, Max Griessl Bheemarjuna Reddy Tamma wrote: > Hi All, > > I am using ffmpeg for a research project in video multicasting over ad > hoc wireless networks. I am facing the following problem: > > I am able to play the highly distorted mp4 files using mplayer, > however when I try ffmpeg on them to decode them into yuv format (I > need that to compute the psnr). It gives the following output : > -- > [devesh@mcn test_psnr]$ ~/ffmpeg/ffmpeg -i try_2_0.mp4 try.yuv > FFmpeg version CVS, Copyright (c) 2000-2004 Fabrice Bellard > configuration: > libavutil version: 49.0.0 > libavcodec version: 51.2.0 > libavformat version: 50.2.1 > built on Feb 23 2006 00:09:27, gcc: 3.3.3 20040412 (Red Hat Linux 3.3.3-7) > Input #0, mov,mp4,m4a,3gp,3g2, from 'try_2_0.mp4': > Duration: 00:00:13.3, start: 0.000000, bitrate: 379 kb/s > Stream #0.0(und), 7.50 fps(r): Video: mpeg4, yuv420p > Output #0, rawvideo, to 'try.yuv': > Stream #0.0, 7.50 fps(c): Video: rawvideo, yuv420p, q=2-31, 200 kb/s > Stream mapping: > Stream #0.0 -> #0.0 > [rawvideo @ 0x82fcd68]dimensions not set > Could not write header for output file #0 (incorrect codec parameters ?) > [devesh@mcn test_psnr]$ > -- > These files are available at http://www.cs.iitm.ernet.in/~dvagr/mp4files > Mplayer can play each one of them, but ffmpeg doesn't work (ie gives > the same error message as shown above), and xvid_decraw segfaults. > > These files were produced using the etmp4 tool, as given on the site > http://www.tkn.tu-berlin.de/research/evalvid/EvalVid/docevalvid.html. > However, ffmpeg is able to decode slighly less distorted files > comfortably, it only seems to give up on more distorted files. > > If ffmpeg can't do the job, is there any way I can make > mplayer/mencoder to dump yuv420p files from this mp4 video ?. > > Thanks, eagerly awaiting your reply. Any pointers/advise/help would be > highly appreciated. > > regards > Arjun Reddy > > -- > T. Bheemarjuna Reddy > Ph.D. Student > HPCN Lab > Dept. of Computer Science & Engg. > IIT Madras, INDIA > Mobile: +91-9444206507 > http://hpcn.cs.iitm.ernet.in > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -- Max Griessl Software Design Holon Technology GmbH (a DynaPel Company) Nymphenburger Str. 90e, D-80636 Munich, Germany Web: www.DynaPel.com From nitthilan fastmail.fm Tue Mar 14 05:59:15 2006 From: nitthilan fastmail.fm (Nitthilan Kannappan Jayakodi) Date: Tue Mar 14 15:28:07 2006 Subject: [Mp4-tech] Error concealment in H.264 standard Message-ID: <1142344755.14720.256586147@webmail.messagingengine.com> Hi I have come across a number of papers talking about the presence of error concealment algorithm in H.264 standard. Can somebody clarify. Regards Nitthilan -- Nitthilan Kannappan Jayakodi nitthilan@fastmail.fm -- http://www.fastmail.fm - A fast, anti-spam email service. From Alexis.Tourapis dolby.net Tue Mar 14 07:21:49 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Tue Mar 14 20:46:09 2006 Subject: [Mp4-tech] compute colocated macroblock Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01DDE465@sapphire.dolby.net> The answer is yes. Best regards, Alexis -----Original Message----- From: ChenJianjun [mailto:cjj730@hotmail.com] Sent: Tuesday, March 14, 2024 12:53 AM To: Tourapis, Alexis; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] compute colocated macroblock Dear Alexis: First, thank you for your reply. I still have doubt for B direct mode: if direct_8x8_inference_flag=1, does four 4x4 blocks in a 8x8 block share the same MV? Since their colocated MV seems same:select the same "x" 4x4 block in the following table: xo|ox oo|oo --|-- oo|oo xo|ox Thanks again! ChenJianjun ---------------------------------------- > Subject: RE: [Mp4-tech] compute colocated macroblock > Date: Mon, 13 Mar 2024 02:19:34 -0800 > From: Alexis.Tourapis@dolby.net > To: cjj730@hotmail.com; mp4-tech@lists.mpegif.org > > Dear ChenJianjun > > The RSD() function relates to the consideration of the direct_8x8_inference_flag. I would recommend studying JVT-C115r1.doc if you are curious about more details on this and to see a visual representation of what is done and why. This is basically what is done in section 8.4.1.2.1 (Derivation process for the co-located 4x4 sub-macroblock partitions) although without any of the explanations. > > dpb_split_field() only prepares motion vectors, that is it generates the corresponding field motion vectors from frame information (which can be quite complex for the MBAFF case). Note that this is done only once after a picture is encoded. Compute_colocated() on the other hand is called before a B slice is generated and prepares/selects the appropriate motion information to be used for the encoding of this slice. It is quite important to make this separation since, due to reordering or the capability of specifying any coding order in H.264/AVC, it is possible that any picture is used for the generation of these parameters. These parameters may also be used differently if there is a certain relationship between top and bottom fields and such are obviously unknown during preparation of the field motion info. Btw you could of course do the decisions in dpb_split_field within compute_colocated but it was found a bit cleaner to prepare the motion information in a field vs frame format after a picture is encoded than redoing everything for every new slice that you have to encode (remember also that you may have to do many slices). > > The code is by far not optimized (especially relating to memory operations), but well ... it works... > > Best regards, > > Alexis > > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ChenJianjun > Sent: Sunday, March 12, 2024 6:55 PM > To: mp4-tech@lists.mpegif.org > Subject: [Mp4-tech] compute colocated macroblock > > Dear All: > > In JM reference decode code, it will use "RSD()" in compute_colocated() function, I dont know the meaning of RSD(), why should it do like this, I cant find the reason in SPEC. Could anyone tell me about it or is there any document I can read? > Another question: In dpb_split_field() function, it will compute the MVs of colocated macroblocks for top_field and bottom_field pictures, but in compute_colocated() function, it will compute again, Is there any difference? > > Thanks very much. > ChenJianjun. > _________________________________________________________________ > 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: > http://messenger.msn.com/cn > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.ph > p > > > ----------------------------------------- > This message (including any attachments) may contain confidential > information intended for a specific individual and purpose. If you > are not the intended recipient, delete this message. If you are not > the intended recipient, disclosing, copying, distributing, or taking > any action based on this message is strictly prohibited. > _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn From mssrihari tataelxsi.co.in Wed Mar 15 17:24:34 2006 From: mssrihari tataelxsi.co.in (MSSrihari) Date: Wed Mar 15 12:51:09 2006 Subject: [Mp4-tech] Query about H.264 Encoder Message-ID: <001201c64827$3afbc220$f609320a@telxsi.com> Hi all, Can any one tell me "do any of the H.264 Encoders developed support SEI & VUI Features" If they are, can you please tell me from where can I get the code & other things. Regards, Srihari The information contained in this electronic message and any attachments to this message are intended for the exclusive use of the addressee(s)and may contain confidential or privileged information. If you are not the intended recipient, please notify the sender or administrator@tataelxsi.co.in From Wesley.DeNeve UGent.be Wed Mar 15 20:41:36 2006 From: Wesley.DeNeve UGent.be (Wesley De Neve) Date: Thu Mar 16 09:40:06 2006 Subject: [Mp4-tech] Query about H.264 Encoder References: <001201c64827$3afbc220$f609320a@telxsi.com> Message-ID: <003601c64868$78b982a0$0300a8c0@persephone> Dear Srihari, all, As an example, the H.264/AVC trailers as available on the site of Apple (www.apple.com/trailers) contain VUI metadata, as well as buffering period SEI messages. Best regards, Wesley De Neve MSSrihari wrote: > Hi all, > > Can any one tell me "do any of the H.264 Encoders developed support > SEI & > VUI Features" > If they are, can you please tell me from where can I get the code & > other > things. > > Regards, > Srihari > > The information contained in this electronic message and any > attachments to this message are intended for the exclusive use of the > addressee(s)and may contain confidential or privileged information. > If you are not the intended recipient, please notify the sender or > administrator@tataelxsi.co.in > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to > indicate the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust > guidelines found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php From bheemarjun gmail.com Thu Mar 16 10:39:49 2006 From: bheemarjun gmail.com (Bheemarjuna Reddy Tamma) Date: Thu Mar 16 09:40:11 2006 Subject: [Mp4-tech] Need a Multiple Description Coding (MDC) codec Message-ID: <3fadd97c0603152109i7cf919c2i9fcdcc2c030e8a4c@mail.gmail.com> Hi All, is there any open source codec that supoorts MD coding? Thanx inadvance. regards arjun reddy -- T. Bheemarjuna Reddy Ph.D. Student HPCN Lab Dept. of Computer Science & Engg. IIT Madras, INDIA Mobile: +91-9444206507 http://hpcn.cs.iitm.ernet.in From tummalakalyani gmail.com Thu Mar 16 17:34:52 2006 From: tummalakalyani gmail.com (tummala kalyani) Date: Thu Mar 16 14:16:09 2006 Subject: [Mp4-tech] Re: Reg. DCT-based subpixel motion compensation Message-ID: <13936ad50603160404o58bdcd00ie9bec516bfd36a78@mail.gmail.com> Hi All, I have some doubts regarding dct-domain subpixel motion compensation. I need to convert a H.264 P frame to an I frame without intra-prediction in dct-domain. If the reference frame is in dct form, how can I get a predicted block using 6-tap FIR filter sub pel motion compensation?. I found a paper on "*DCT based sub pixel motion compensation*" by *Ut Va Koc and K. J Ray Liu*, but they mentioned only for subpel motion compensation by bilinear and cubic interpolation in dct domain. But here in H.264, I need to do by 6-tap FIR filter. Can some one point me to some resources on using 6 tap filters in DCT domain. Thanking you in advance kalyani -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060316/45333a29/attachment.html From tummalakalyani gmail.com Thu Mar 16 17:36:13 2006 From: tummalakalyani gmail.com (tummala kalyani) Date: Thu Mar 16 14:16:13 2006 Subject: [Mp4-tech] Reg. DCT-based subpixel motion compensation In-Reply-To: <13936ad50603160404o58bdcd00ie9bec516bfd36a78@mail.gmail.com> References: <13936ad50603160404o58bdcd00ie9bec516bfd36a78@mail.gmail.com> Message-ID: <13936ad50603160406w14bcf78fif23ff871a9aaa7a6@mail.gmail.com> Hi All, I have some doubts regarding dct-domain subpixel motion compensation. I need to convert a H.264 P frame to an I frame without intra-prediction in dct-domain. If the reference frame is in dct form, how can I get a predicted block using 6-tap FIR filter sub pel motion compensation?. I found a paper on "*DCT based sub pixel motion compensation*" by *Ut Va Koc and K. J Ray Liu*, but they mentioned only for subpel motion compensation by bilinear and cubic interpolation in dct domain. But here in H.264, I need to do by 6-tap FIR filter. Can some one point me to some resources on using 6 tap filters in DCT domain. Thanking you in advance kalyani -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060316/92e8edd3/attachment.html From Ilan.Daniel Adimos.com Thu Mar 16 16:24:36 2006 From: Ilan.Daniel Adimos.com (Ilan Daniel) Date: Thu Mar 16 19:04:09 2006 Subject: [Mp4-tech] Re: Reg. DCT-based subpixel motion compensation Message-ID: <2A01AFD68E343242BA2C748A5E1A733DB28F77@jerry.Adimos.com> I read the paper you mentioned a while a go. I do not have a copy of the paper near by, but, as I recall, the bilinear and cubic interpolation are only examples. For 6 tap filter, there is a need of course to build new Mtrices. Regards, Ilan Daniel ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of tummala kalyani Sent: ? 16 ??? 2006 14:05 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Re: Reg. DCT-based subpixel motion compensation Hi All, I have some doubts regarding dct-domain subpixel motion compensation. I need to convert a H.264 P frame to an I frame without intra-prediction in dct-domain. If the reference frame is in dct form, how can I get a predicted block using 6-tap FIR filter sub pel motion compensation?. I found a paper on "DCT based sub pixel motion compensation" by Ut Va Koc and K. J Ray Liu, but they mentioned only for subpel motion compensation by bilinear and cubic interpolation in dct domain. But here in H.264, I need to do by 6-tap FIR filter. Can some one point me to some resources on using 6 tap filters in DCT domain. Thanking you in advance kalyani -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060316/61df097d/attachment.html From mp4-tech-owner lists.mpegif.org Thu Mar 16 19:55:48 2006 From: mp4-tech-owner lists.mpegif.org (MPEGIF List Admins) Date: Fri Mar 17 13:46:08 2006 Subject: FW: [Mp4-tech] A question about the frame_num in H.264 Message-ID: <004f01c6492b$3d7b4cf0$0301a8c0@corp.intertrust.com> Forwarding post form non-member address _____ From: bhaskell [mailto: ] Sent: Thursday, 16 March 2024 19:02 To: Wang Lu Cc: mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] A question about the frame_num in H.264 <>1: Must the two fields of one frame be a complementary field pair? No. But with non-complementary coded field pictures, defining frames may be problematic. 2: Can the top field be a non-reference field, but the bottom field be a reference field in one frame (non complementary field pair)? The simple answer is yes. A more complicated answer requires us to ask "What is a frame?" We have no definition for "non complementary field pair". For example, currently is is possible to label all non-paired fields as top fields. Some of us are trying to tighten up the definitions to enable the signalling of "normal" interlace through the fixed_frame_rate_flag. Stay tuned. -- Barry G. Haskell tel +1 408 974 6333 Apple Computer Inc. fax " " " 1756 2 Infinite Loop MS: 302-3KS Cupertino, CA 95014 bhaskell@apple.com (also B.Haskell@ieee.org) -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060316/b3c3db50/attachment.html From publiclist gmail.com Fri Mar 17 14:06:27 2006 From: publiclist gmail.com (James Hensey) Date: Fri Mar 17 13:46:12 2006 Subject: [Mp4-tech][Audio][AAC] RMS conformance for 16 bit AAC LC Decoder Message-ID: Hello Experts, Adding to a previous question by Shafi. The Conformance Standard ISO/IEC 14496-4, states that the maximum allowed RMS should be calculated using the equation RMS = 2^(-K+1)/sqrt(12) where, K is the required accuracy level. Using the above equation, for K=16, Maximum Allowed RMS = 8.8096 e-6 If the number of output bits per sample is 16, then as per the conformance standard and the conformance software, bits 17 to 24 are zero padded. When bits 17 to 24 are zero padded and assuming that all these 8 bits are in error with respect to the reference output, we get the value of RMS error as Calculated RMS= 2^(-16)+ 2^(-17)+ 2^(-18)+ 2^(-19) + 2^(-20)+ 2^(-21)+ 2^(-22)+ 2^(-23) = 30.39e-6 This calculation assumes that the first 16 bits of the generated output are the same as the reference output. The calculated value of RMS is clearly more than the maximum RMS error allowed. The calculated value of RMS can be less than the maximum allowed RMS only when, for an accuracy level of 16 bits the first 18 bits of every sample are the same as the first 18 bits of the reference output. Is this an accurate interpretation of the text in the Conformance Standard ? If yes, then as Shafi identified, RMS criterion cannot be satisfied for an accuracy level of 16 bits when the output bits per sample is also 16 bits. Regards, James > -----Original Message----- > From: mp4-tech-bounces@lists.mpegif.org > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Shafi > Ullah Khan > Sent: Monday, March 13, 2024 1:54 PM > To: MP4-Tech > Subject: [Mp4-tech] RMS conformance for 16 bit AAC LC Decoder > > I have doubt regarding test vector for RMS conformance of > 16-bit AAC LC decoder. > > Does any AAC LC decoder with 16 bit output able to meet RMS > requirements (2^(-15)/sqrt(12) == 8.810e-6 == -101.10 dB) for > al00_08.wav and al00_32.wav reference output? > > For these two test vectors theoretically RMS cannot be less > than 8.821e-6 and 8.817e-6 respectively for 16-bit decoder. > > shafi > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060317/91583602/attachment.html From princeofpersia79 gmail.com Fri Mar 17 15:20:30 2006 From: princeofpersia79 gmail.com (viraj ambetkar) Date: Sat Mar 18 14:51:07 2006 Subject: [Mp4-tech] Extracting MPEG-2 from a MPEG-4 stream Message-ID: <77f040c10603171520g60530a34sca89be520ae75721@mail.gmail.com> Hello Experts, Is it possible to extract a MPEG-2 stream from a MPEG-4 stream (without going through a transcoder)? To make myself clear, the MPEG-4 stream contains motion and texture information. Is it possible to somehow utilize this information to intelligently obtain a MPEG-2 stream without going through the time-consuming transcoding process? Any help is greatly appreciated, Viraj -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060317/3e3db303/attachment-0001.html From garysull windows.microsoft.com Fri Mar 17 16:34:45 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Sat Mar 18 14:51:12 2006 Subject: [Mp4-tech] [H.264] Motion vector count In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA2849DF@exchtp02.vntek.com.tw> Message-ID: <91D7F2CEE3425A4A9D11311D09FCE246139FC578@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> That's correct, there is no limit for levels 1 to 2.2. The idea of imposing a limit per 2 MBs instead of for each MB is to allow some macroblocks to use more than half the limit while constraining the total quantity within a small window. Best Regards, Gary Sullivan ________________________________ From: EricChuang@via.com.tw [mailto:EricChuang@via.com.tw] Sent: Monday, March 06, 2024 3:39 PM To: Gary Sullivan; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] [H.264] Motion vector count Hi Gary, In standard, only level 3~5.1 defines the MaxMvsPer2Mb, but level 1~2.2 doesn't. Does it mean there is no limit for level 1~2.2 ? I don't understand why the max mv count is defined for consecutive 2 Mbs, instead for each Mb ? Could you tell me why ? Thanks, Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Gary Sullivan Sent: Monday, March 06, 2024 11:04 AM To: Eric Chuang; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] [H.264] Motion vector count Eric et al, The answer is level-dependent and the limit is applied to each two consecutive macroblocks in the bitstream. See the use of the variable MaxMvsPer2Mb in the standard. subMvCnt is summed to compute MvCnt. Then, in some levels, the sum of MvCnt for two consecutive macroblocks is constrained as specified in Annex A. Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Sunday, March 05, 2024 4:45 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] Motion vector count Dear experts, Could anyone tell me what's the maximum possible motion vector count of a macroblock ? I read a paper, and it reads the maximum is 16. But from the standard, if we split the macroblock to 16 4x4 submacroblock partition with bi-direction prediction, there shall be 32 motion vectors. Does the standard define the limit ? Another question is what's the usage of subMvCnt mentioned in standard ? In deriving motion vectors, the standard mentioned how to compute subMvCnt, but I did not found the usage of it. Thanks very much, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060317/990d78ca/attachment-0001.html From garysull windows.microsoft.com Fri Mar 17 16:53:10 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Sat Mar 18 14:51:17 2006 Subject: [Mp4-tech] A question about the frame_num in H.264 In-Reply-To: <004f01c6492b$3d7b4cf0$0301a8c0@corp.intertrust.com> Message-ID: <91D7F2CEE3425A4A9D11311D09FCE246139FC5B1@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> Re "Can the top field be a non-reference field, but the bottom field be a reference field in one frame (non complementary field pair)?" Barry says "Yes". I say "No". But we agree with each other completely. :-) It really depends on what you mean by "in one frame". If you mean it to refer to the frame storage memory area in the hypothetical DPB, then the answer is No. Fields are not considered paired (i.e., "complementary") for storage into the DPB unless they are either both reference fields or both non-reference fields. Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of MPEGIF List Admins Sent: Thursday, March 16, 2024 10:56 AM To: mp4-tech@lists.mpegif.org Cc: bhaskell@apple.com Subject: FW: [Mp4-tech] A question about the frame_num in H.264 Forwarding post form non-member address ________________________________ From: bhaskell [mailto: ] Sent: Thursday, 16 March 2024 19:02 To: Wang Lu Cc: mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] A question about the frame_num in H.264 <>1: Must the two fields of one frame be a complementary field pair? No. But with non-complementary coded field pictures, defining frames may be problematic. 2: Can the top field be a non-reference field, but the bottom field be a reference field in one frame (non complementary field pair)? The simple answer is yes. A more complicated answer requires us to ask "What is a frame?" We have no definition for "non complementary field pair". For example, currently is is possible to label all non-paired fields as top fields. Some of us are trying to tighten up the definitions to enable the signalling of "normal" interlace through the fixed_frame_rate_flag. Stay tuned. -- Barry G. Haskell tel +1 408 974 6333 Apple Computer Inc. fax " " " 1756 2 Infinite Loop MS: 302-3KS Cupertino, CA 95014 bhaskell@apple.com (also B.Haskell@ieee.org) -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060317/95b03d8c/attachment-0001.html From cjj730 hotmail.com Sat Mar 18 04:36:27 2006 From: cjj730 hotmail.com (ChenJianjun) Date: Sat Mar 18 14:51:21 2006 Subject: [Mp4-tech] Calculate the MIPS of h.264 sw decoder Message-ID: Dear Experts: I want to evalute the MIPS that h.264 software decoder need, especially for those HD h.264 clips. Is there any tool can use or some other documents can reference ? Thanks ChenJianjun _________________________________________________________________ 迵薊儂腔攬衭輛俴蝠霜ㄛ③妏蚚 MSN Messenger: http://messenger.msn.com/cn From A00774790 itesm.mx Sun Mar 19 13:25:55 2006 From: A00774790 itesm.mx (=?ISO-8859-1?Q?Roberto_Rivera_Jim=E9nez?=) Date: Sun Mar 19 23:04:07 2006 Subject: [Mp4-tech] H.264 DCT coefficients Message-ID: <441DB043.5060702@itesm.mx> Hello moving picture experts group, I'm trying to print in a file the 8x8 DCT coefficients (before and after quantization) of different H.264 encoded videos. In the configuration file (encoder.cfg) I modified the lines: UseHadamard = 0 Transform8x8Mode = 2 And then modified the file transform8x8.c in the routine int dct_luma8x8(int b8,int *coeff_cost, int intra) to print the DCT coefficients. I found out that sometimes this routine is called more than once for the same block (the same img->number, img->current_mb_nr and b8). I assumed that the "right" coefficients for each block were those from the last call. So, to save the array img->m7[][] (between the vertical transform and the quantization http://iphome.hhi.de/suehring/tml/doc/lenc/html/transform8x8_8c-source.html#l01610). I added the lines: for (coeff_ctr=0;coeff_ctr < 64;coeff_ctr++) { if (is_field_mode) { // Alternate scan for field coding i=FIELD_SCAN8x8[coeff_ctr][0]; j=FIELD_SCAN8x8[coeff_ctr][1]; } else { i=SNGL_SCAN8x8[coeff_ctr][0]; j=SNGL_SCAN8x8[coeff_ctr][1]; } DCT_coefficients[img->current_mb_nr][b8][coeff_ctr]=img->m7[i][j]; } For the quantized coefficients I added the line: Q_DCT_coefficients[img->current_mb_nr][b8][coeff_ctr]=level; just before the instruction in: http://iphome.hhi.de/suehring/tml/doc/lenc/html/transform8x8_8c-source.html#l01691 I modified the file image.c in the routine int encode_one_frame() to print the DCT_coefficients and the Q_DCT_coefficients frame by frame. In order to test my assumption, I modified the file mode_decision.c at the end of the routine int encode_one_macroblock() I added the lines: for( j=0; j<4; j++) { c=0; for( i=0; (i<64 && img->cofAC[j][0][0][i]!=0 ); i++) { for (r=0; r < img->cofAC[j][0][1][i]; r++) { fprintf(DCT_coeff_file, "0\t"); c++; } fprintf(DCT_coeff_file, "%d\t", img->cofAC[j][0][0][i]); c++; } while (c<64) { fprintf(DCT_coeff_file, "0\t"); c++; } fprintf(DCT_coeff_file, "\n"); } // cofAC[8x8block][4x4block][level/run][scan_pos] to print cofAC (I used the 8x8block parameter and left the 4x4block parameter of cofAC always equal to 0, is it correct?). I compared the results of cofAC and Q_DCT_coefficients, and found that my assumption was wrong. For the blocks where dct_luma8x8 is called once cofAC and Q_DCT_coefficients are the same, but for the other blocks cofAC can take the values of any the dct_luma8x8 calls for that block (not only the last call, as I previously assumed). What parameter does the encoder use to choose wich call to dct_luma8x8 will be the "good one" to save the coefficients? Thank you in advance for your help. Best Regards, Roberto Rivera From Alexis.Tourapis dolby.net Sun Mar 19 16:15:27 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Mon Mar 20 13:51:09 2006 Subject: [Mp4-tech] H.264 DCT coefficients Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01DDE906@sapphire.dolby.net> Dear Roberto, First of all, UseHadamard has nothing to do with the transform and is only a motion estimation option. With regards now to the calls you are seeing, these relate to how the encoder performs mode decision, i.e. selecting the best possible mode for the current macroblock. You should instead check files rdopt.c and mode_decision.c to understand how this process is done. You should check in particular void SetCoeffAndReconstruction8x8 () and void set_stored_macroblock_parameters and how the cofAC array within the img structure (img->cofAC) is used (assuming you are using RDOOptimization = 1. Best regards, Alexis -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Roberto Rivera Jim?nez Sent: Sunday, March 19, 2024 11:26 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] H.264 DCT coefficients Hello moving picture experts group, I'm trying to print in a file the 8x8 DCT coefficients (before and after quantization) of different H.264 encoded videos. In the configuration file (encoder.cfg) I modified the lines: UseHadamard = 0 Transform8x8Mode = 2 And then modified the file transform8x8.c in the routine int dct_luma8x8(int b8,int *coeff_cost, int intra) to print the DCT coefficients. I found out that sometimes this routine is called more than once for the same block (the same img->number, img->current_mb_nr and b8). I assumed that the "right" coefficients for each block were those from the last call. So, to save the array img->m7[][] (between the vertical transform and the quantization http://iphome.hhi.de/suehring/tml/doc/lenc/html/transform8x8_8c-source.html#l01610). I added the lines: for (coeff_ctr=0;coeff_ctr < 64;coeff_ctr++) { if (is_field_mode) { // Alternate scan for field coding i=FIELD_SCAN8x8[coeff_ctr][0]; j=FIELD_SCAN8x8[coeff_ctr][1]; } else { i=SNGL_SCAN8x8[coeff_ctr][0]; j=SNGL_SCAN8x8[coeff_ctr][1]; } DCT_coefficients[img->current_mb_nr][b8][coeff_ctr]=img->m7[i][j]; } For the quantized coefficients I added the line: Q_DCT_coefficients[img->current_mb_nr][b8][coeff_ctr]=level; just before the instruction in: http://iphome.hhi.de/suehring/tml/doc/lenc/html/transform8x8_8c-source.html#l01691 I modified the file image.c in the routine int encode_one_frame() to print the DCT_coefficients and the Q_DCT_coefficients frame by frame. In order to test my assumption, I modified the file mode_decision.c at the end of the routine int encode_one_macroblock() I added the lines: for( j=0; j<4; j++) { c=0; for( i=0; (i<64 && img->cofAC[j][0][0][i]!=0 ); i++) { for (r=0; r < img->cofAC[j][0][1][i]; r++) { fprintf(DCT_coeff_file, "0\t"); c++; } fprintf(DCT_coeff_file, "%d\t", img->cofAC[j][0][0][i]); c++; } while (c<64) { fprintf(DCT_coeff_file, "0\t"); c++; } fprintf(DCT_coeff_file, "\n"); } // cofAC[8x8block][4x4block][level/run][scan_pos] to print cofAC (I used the 8x8block parameter and left the 4x4block parameter of cofAC always equal to 0, is it correct?). I compared the results of cofAC and Q_DCT_coefficients, and found that my assumption was wrong. For the blocks where dct_luma8x8 is called once cofAC and Q_DCT_coefficients are the same, but for the other blocks cofAC can take the values of any the dct_luma8x8 calls for that block (not only the last call, as I previously assumed). What parameter does the encoder use to choose wich call to dct_luma8x8 will be the "good one" to save the coefficients? Thank you in advance for your help. Best Regards, Roberto Rivera _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From pt_david yahoo.com Sun Mar 19 22:31:24 2006 From: pt_david yahoo.com (Prasanth) Date: Mon Mar 20 13:51:13 2006 Subject: [Mp4-tech] FREXT vs Main Profile anomaly Message-ID: <20060320063124.76113.qmail@web37903.mail.mud.yahoo.com> Hi Experts, I appreciate the help which I have recieved through this forum previously and also the time taken by the people to answer the emails. I want some help with regards to Frext. I was dong some experiments with Frext(with CABAC) and I put in 10 multiple ref frames with weighted prediction for P,B and also weighted ME in the settings and I obtained a increase in bitrate for the same sequence I did with Main(2 multiple ref with CABAC and no weighted ref.) this is really surprising. My sequence is 1280x720p set to be encoded at 30fps and GOP is IBBPBBPBBPBBI I see in the paper H.264/MPEG_4AVC Fidelity Range Xtensions:Tools,Profiles,performance and App Areas..by .Marpe,Weigand et al.... that at least 10% coding gain is obtained over MP.. can anyone explain this strange result. Is it true that if Multiref is increased beyond 5 or so the bitrate increases? thanks, Prasanth --------------------------------- Yahoo! Mail Bring photos to life! New PhotoMail makes sharing a breeze. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060319/d433d2a8/attachment.html From Andreas.Schneider codingtechnologies.com Mon Mar 20 09:56:48 2006 From: Andreas.Schneider codingtechnologies.com (Andreas Schneider) Date: Mon Mar 20 13:51:17 2006 Subject: [Mp4-tech] Coding Technologies is hiring Message-ID: *Summary*: Coding Technologies offers several MPEG related positions. See http://www.codingtechnologies.com/company/careers.htm for details. As many of you probably know: Coding Technologies is one of the primary contributors to various MPEG audio standards and is also one of the most active organizations driving the actual use of open standards in the mobile, broadcasting, and PC/Internet markets. Based on the continuing success of our company we are increasing our business and technical staff. There are several open positions including Senior Research, System Engineering, and Business Development positions. For details please see our website at http://www.codingtechnologies.com/company/careers.htm Within Coding Technologies we combine the flexibility and attractiveness of a start-up company with the financial security of a profitable company. Coding Technologies offers you the opportunity to work on an international team of world class researchers and business developers. If you are interested, please e-mail your application to jobs@CodingTechnologies.com. From EricChuang via.com.tw Mon Mar 20 17:57:33 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Mon Mar 20 13:51:22 2006 Subject: [Mp4-tech] H.264 slice group Message-ID: <6A7AB4426712514F9E2AB28C407285FA372D9B@exchtp02.vntek.com.tw> Dear experts, I have a question regarding the slice and slice group. If slice group number is 1 and data partitioning is not used, does it mean there is exactly only one slice (one NAL) for each coded picture ?? Thanks in advance, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060320/1412e74d/attachment.html From ralph.sperschneider iis.fraunhofer.de Sun Mar 19 12:01:48 2006 From: ralph.sperschneider iis.fraunhofer.de (Ralph Sperschneider) Date: Mon Mar 20 19:10:07 2006 Subject: [Mp4-tech] Re: EP Tool In-Reply-To: <03d001c649a7$4d9bcca0$69e6c68a@amk.st.com> References: <03d001c649a7$4d9bcca0$69e6c68a@amk.st.com> Message-ID: <441D3A1C.7000500@iis.fraunhofer.de> Yao XUE wrote: > Dear Dr. Sperschneider, > > I got your email address from mp4-tech. I'm looking at mpeg4 ep tool. As > I know you are an expert in mp4 audio, I would greatly appreciate if you > could drop me a few lines for my questions! > > My questions are: > 1. Can EP Tool be used as stand-alone, or it has to be combined with ER > objects (e.g ER AAC, ER CELP)? All conformance vectors with EP seem > to go together with ER. ISO/IEC14496-3 specifies the usage of the EP-Tool only with the ER AOTs. The EP-Tool is intended to provide unequal error protection. Particularly for codecs with variable frame lengths (e.g. AAC), only the ER AOTs provide the necessary interface (by means of instances of various error sensitivity categories) to allow for unequal error protection. The EP-Tool reference software can be used as a stand-alone tool. > 2. If EP Tool can be stand alone, then to generate, say, an AAC stream > with EP, should I use a raw AAC as input, or should I use other formats > (mp4, ass...)? The interface between the audio codec and the ep tool is based on access units. The stand-alone EP tool reference software encoder requires raw ER AAC LC input, whereas this raw ER AAC LC input is provided by the aacErrRobTrans transcoder. This transcoder generates ER AAC LC bitstreams based on AAC LC bitstreams (it rearranges the bitstream payload based on the ESC definitions and optionally applies the ER AAC tools). It furthermore generates the required framing info (length of the ER AAC LC frames, lengths of the individual ep classes). > 3. Where can I find more information about writing the "input side > information file" which is required for the EP encoding in the reference > software? I'm not able to find any example in the readme. See second part of the response to question 2. The stand-alone EP encoder requires an "ER AAC raw_data_stream" (not defined in a standard, just a software interface) and framing information. The framing information should look as follows (assuming a configuration with three classes, as is happens for ER AAC LC mono if the RVLC tool is not enabled): The CRC results are out of scope on encoder side, but the same interface files were used on decoder side during the development phase, where the CRC results were used from the decoder to decide whether the frame could be decoded or had to be concealed. > Thanks a lot and best regards, > Xue Yao > > Technical Staff III > STMicroelectronics > R&D Centre - AMK1 > 7 Serangoon North Avenue 5 > Singapore 554574 > DID: +65 6216 5154 Best regards, Ralph -- Dipl.-Ing. Ralph Sperschneider | Phone: +49 9131 776 344 Fraunhofer IIS | Fax: +49 9131 776 398 Am Wolfsmantel 33 | mailto:ralph.sperschneider@iis.fraunhofer.de D 91058 Erlangen | http://www.iis.fraunhofer.de/amm/ From birgera rdr.com Mon Mar 20 10:33:04 2006 From: birgera rdr.com (Alex Birger) Date: Mon Mar 20 19:10:12 2006 Subject: [Mp4-tech] Calculate the MIPS of h.264 sw decoder In-Reply-To: Message-ID: There is an Elecard DS filter called "Chegepuga". You may look for it here: http://forum.doom9.org/showthread.php?p=703999#post703999 Good luck! Alex Birger _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of ChenJianjun Sent: Friday, March 17, 2024 11:36 PM To: Subject: [Mp4-tech] Calculate the MIPS of h.264 sw decoder Importance: Low Dear Experts: I want to evalute the MIPS that h.264 software decoder need, especially for those HD h.264 clips. Is there any tool can use or some other documents can reference ? Thanks ChenJianjun _________________________________________________________________ $(ASkA*;z5DEsSQ=xPP=;Aw#,GkJ9SC(T MSN Messenger: http://messenger.msn.com/cn _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060320/12716d76/attachment.html From Alexis.Tourapis dolby.net Mon Mar 20 09:14:40 2006 From: Alexis.Tourapis dolby.net (Tourapis, Alexis) Date: Mon Mar 20 19:10:16 2006 Subject: [Mp4-tech] FREXT vs Main Profile anomaly Message-ID: <7272EE229DA1AA48B47EBDC47EB0C68B01DDE92F@sapphire.dolby.net> Dear Prasanth, First an important clarification. Weighted prediction and multiple references beyond 2 are also supported from Main (depending obviously on Level). The key difference however, sticking with only High profile, is support for 8x8 Intra prediction and intra transform, but also quantization matrices (note though that the later results usually in more of a subjective improvement and could result in worse RD performance so I would suggest not to use it for your comparisons at this stage). In any case, if you are using the JM reference software, please make also sure that you are making use of the following parameters to properly exploit Weighted Prediction: RDPictureDecision=1 GenerateMultiplePPS=1 UseWeightedReferenceME=1 I would also strongly recommend to use bipredictive ME, i.e.. : BiPredMotionEstimation = 1 BiPredMERefinements = 3 BiPredMESearchRange = 32 BiPredMESubPel = 2 I am, of course, also assuming that you are running the encoder in the high Complexity RDO mode (RDOptimization = 1) while you are also evaluating bitrate and quality at the same time. With regards to the number of references, entropy coding should be able to handle these many references and what you are seeing in my opinion, and if you have set up the encoder properly, should have nothing to do with it. Best regards, Alexis ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Prasanth Sent: Sunday, March 19, 2024 10:31 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] FREXT vs Main Profile anomaly Hi Experts, I appreciate the help which I have recieved through this forum previously and also the time taken by the people to answer the emails. I want some help with regards to Frext. I was dong some experiments with Frext(with CABAC) and I put in 10 multiple ref frames with weighted prediction for P,B and also weighted ME in the settings and I obtained a increase in bitrate for the same sequence I did with Main(2 multiple ref with CABAC and no weighted ref.) this is really surprising. My sequence is 1280x720p set to be encoded at 30fps and GOP is IBBPBBPBBPBBI I see in the paper H.264/MPEG_4AVC Fidelity Range Xtensions:Tools,Profiles,performance and App Areas..by .Marpe,Weigand et al.... that at least 10% coding gain is obtained over MP.. can anyone explain this strange result. Is it true that if Multiref is increased beyond 5 or so ! the bitrate increases? thanks, Prasanth ________________________________ Yahoo! Mail Bring photos to life! New PhotoMail makes sharing a breeze. ----------------------------------------- This message (including any attachments) may contain confidential information intended for a specific individual and purpose. If you are not the intended recipient, delete this message. If you are not the intended recipient, disclosing, copying, distributing, or taking any action based on this message is strictly prohibited. From mikel_m01 yahoo.com Mon Mar 20 11:09:20 2006 From: mikel_m01 yahoo.com (M M) Date: Mon Mar 20 20:46:06 2006 Subject: [Mp4-tech] I need help to get vitooki Message-ID: <20060320190920.63034.qmail@web37512.mail.mud.yahoo.com> Hello Experts, I need help to get open source software vitooki, I tried the site on sourceforge but it doesn't seem to be there. Rgds, Michael --------------------------------- Yahoo! Mail Use Photomail to share photos without annoying attachments. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060320/0fbbd53f/attachment-0001.html From garysull windows.microsoft.com Mon Mar 20 11:59:12 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Mon Mar 20 20:46:11 2006 Subject: [Mp4-tech] H.264 slice group In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA372D9B@exchtp02.vntek.com.tw> Message-ID: <91D7F2CEE3425A4A9D11311D09FCE246139FCCD0@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com> No. Each "slice group" can contain many slices. Best Regards, -Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Monday, March 20, 2024 1:58 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] H.264 slice group Dear experts, I have a question regarding the slice and slice group. If slice group number is 1 and data partitioning is not used, does it mean there is exactly only one slice (one NAL) for each coded picture ?? Thanks in advance, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060320/c856a916/attachment-0001.html From EricChuang via.com.tw Tue Mar 21 09:21:08 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Tue Mar 21 19:46:07 2006 Subject: [Mp4-tech] [H.264] An trivial POC question Message-ID: <6A7AB4426712514F9E2AB28C407285FA372DA0@exchtp02.vntek.com.tw> Dear experts, I had one trivial questions regarding POC, could you help me clarify. In spec. 8.2.1, it defines a range for POC from -2^31 to 2^31-1, is it necessary to have such large of range ? Since the IDR have POC as 0, and it seems to me that none of other pictures of the same sequence be displayed before this IDR, so POC of the others shall be greater than 0, why negative value is allowed ? Thanks, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060321/a81f31bb/attachment.html From audio sukhan.com Tue Mar 21 09:23:36 2006 From: audio sukhan.com (Shafi Ullah Khan) Date: Tue Mar 21 19:46:11 2006 Subject: [Mp4-tech][Audio][AAC] RMS conformance for 16 bit AAC LC Decoder In-Reply-To: References: Message-ID: <5545a8b30603201953l3556eb9bkd94abccb3ba9f356@mail.gmail.com> My query is related to only two reference test vectors (al00_08.wav and al00_32.wav) for LC decoder. For these references wav lower bound of RMS (for 16-bit decoder) is more than required RMS. Lower bound is calculated assuming 24-bit internal output is exactly matching to reference wav (RMS for 24-bit accuracy 0), which is rounded to 16 bits giving the lower bound for 16-bit decoder. shafi On 3/17/06, James Hensey wrote: > Hello Experts, > > Adding to a previous question by Shafi. > The Conformance Standard ISO/IEC 14496-4, states that the maximum allowed > RMS should be calculated using the equation > > RMS = 2^(-K+1)/sqrt(12) > where, K is the required accuracy level. > > Using the above equation, for K=16, > Maximum Allowed RMS = 8.8096 e-6 > > If the number of output bits per sample is 16, then as per the conformance > standard and the conformance software, bits 17 to 24 are zero padded. > When bits 17 to 24 are zero padded and assuming that all these 8 bits are > in error with respect to the reference output, we get the value of RMS error > as > Calculated RMS= 2^(-16)+ 2^(-17)+ 2^(-18)+ 2^(-19) + 2^(-20)+ 2^(-21)+ > 2^(-22)+ 2^(-23) > = 30.39e-6 > This calculation assumes that the first 16 bits of the generated output are > the same as the reference output. > > The calculated value of RMS is clearly more than the maximum RMS error > allowed. > > The calculated value of RMS can be less than the maximum allowed RMS only > when, for an accuracy level of 16 bits the first 18 bits of every sample are > the same as the first 18 bits of the reference output. > > Is this an accurate interpretation of the text in the Conformance Standard > ? > > If yes, then as Shafi identified, RMS criterion cannot be satisfied for an > accuracy level of 16 bits when the output bits per sample is also 16 bits. > > Regards, > James > > > > -----Original Message----- > > From: mp4-tech-bounces@lists.mpegif.org > > [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of > Shafi > > Ullah Khan > > Sent: Monday, March 13, 2024 1:54 PM > > To: MP4-Tech > > Subject: [Mp4-tech] RMS conformance for 16 bit AAC LC Decoder > > > > I have doubt regarding test vector for RMS conformance of > > 16-bit AAC LC decoder. > > > > Does any AAC LC decoder with 16 bit output able to meet RMS > > requirements (2^(-15)/sqrt(12) == 8.810e-6 == -101.10 dB) for > > al00_08.wav and al00_32.wav reference output? > > > > For these two test vectors theoretically RMS cannot be less > > than 8.821e-6 and 8.817e-6 respectively for 16-bit decoder. > > > > shafi > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > > -- with warm regards shafi ullah khan From giri_tammana yahoo.co.in Tue Mar 21 14:08:17 2006 From: giri_tammana yahoo.co.in (Giridhar) Date: Tue Mar 21 19:46:16 2006 Subject: [Mp4-tech] H264-VC1 Comparision Message-ID: Hi, Does anybody have any technical comparison between standards H.264 and VC1. Which one is better in which way, Which one is going to take lead for HD content? -Giridhar -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060321/71c5ca2f/attachment.html From sma com.dtu.dk Tue Mar 21 13:09:38 2006 From: sma com.dtu.dk (Shankar Manuel Aghito) Date: Tue Mar 21 19:46:22 2006 Subject: [Mp4-tech] [Video] [H.264] channel feedback with jvt reference software Message-ID: I want to simulate use of channel feedback for error resilience, using JVT reference software. The simplest solution is not using frames that are damaged for reference. Example: I0 P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 ....... Assume 5 reference frames are used. P6 is reported damaged when the encoder is about to process P10. Therefore P7 P8 P9 are also potentially damaged. I want to force the encoder to use only P5 as reference for coding P10. Please advice my on how to modify the reference software in order to do that 1)should I act on the loop in PartitionMotionSearch? How? 2)should I try to mark the frames that I do not want to use as "unused for reference"? How? 3)Other methods? Best regards and thanks for your time, Shankar Manuel Aghito From garysull windows.microsoft.com Tue Mar 21 13:04:08 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Wed Mar 22 09:22:08 2006 Subject: [Mp4-tech] [H.264] An trivial POC question In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA372DA0@exchtp02.vntek.com.tw> Message-ID: <03CB47D9C3F8074498E4653F814D6E8F0BA32F@WIN-MSG-20.wingroup.windeploy.ntdev.microsoft.com> Eric, What makes you think that non of the pictures in the sequence are displayed before the IDR picture? That can happen. -Gary ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Monday, March 20, 2024 5:21 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] An trivial POC question Dear experts, I had one trivial questions regarding POC, could you help me clarify. In spec. 8.2.1, it defines a range for POC from -2^31 to 2^31-1, is it necessary to have such large of range ? Since the IDR have POC as 0, and it seems to me that none of other pictures of the same sequence be displayed before this IDR, so POC of the others shall be greater than 0, why negative value is allowed ? Thanks, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060321/eaa37658/attachment-0001.html From EricChuang via.com.tw Wed Mar 22 07:07:35 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Wed Mar 22 09:22:13 2006 Subject: [Mp4-tech] [H.264] An trivial POC question Message-ID: <6A7AB4426712514F9E2AB28C407285FA372DA2@exchtp02.vntek.com.tw> Hi Gary, Oh, that's why it's negative. I just thought the IDR is the first picture of a sequence, and concludes it shall be displayed first, and I was wrong about this. Thank you for correct my mistake, Eric -----Original Message----- From: Gary Sullivan [mailto:garysull@windows.microsoft.com] Sent: Tuesday, March 21, 2024 1:04 PM To: Eric Chuang; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] [H.264] An trivial POC question Eric, What makes you think that non of the pictures in the sequence are displayed before the IDR picture? That can happen. -Gary _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of EricChuang@via.com.tw Sent: Monday, March 20, 2024 5:21 PM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] An trivial POC question Dear experts, I had one trivial questions regarding POC, could you help me clarify. In spec. 8.2.1, it defines a range for POC from -2^31 to 2^31-1, is it necessary to have such large of range ? Since the IDR have POC as 0, and it seems to me that none of other pictures of the same sequence be displayed before this IDR, so POC of the others shall be greater than 0, why negative value is allowed ? Thanks, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060322/ad44a3f3/attachment-0001.html From princeofpersia79 gmail.com Tue Mar 21 15:30:50 2006 From: princeofpersia79 gmail.com (viraj ambetkar) Date: Wed Mar 22 09:22:17 2006 Subject: [Mp4-tech] Regarding the extraction of MPEG-2 from MPEG-4 Message-ID: <77f040c10603211530r447bf764r38a1112d508fc929@mail.gmail.com> Hello Experts, Is it theoretically (and practically) possible to extract a MPEG-2 stream from a MPEG-4/H.264 video stream (without using a transcoder)? To make myself clear, a MPEG-4 stream already contains motion vectors and texture information. Is it possible to somehow utilize this information to intelligently obtain a MPEG-2 stream without going through the time-consuming transcoding process? Any help is greatly appreciated, Viraj Ambetkar Software Engineer Path 1 Network Technologies, Inc. 6215 Ferris Square, Suite 140, San Diego, CA 92121 Email: viraj.ambetkar@path1.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060321/c9e48e18/attachment.html From manish.sikarwar st.com Wed Mar 22 10:06:01 2006 From: manish.sikarwar st.com (Manish SIKARWAR) Date: Wed Mar 22 09:22:21 2006 Subject: [Mp4-tech] VBV control in MPEG-4 In-Reply-To: <200603202049.k2KKn6Zp011886@lists1.magma.ca> Message-ID: <00a601c64d6a$1f8e00f0$312dc70a@dlh.st.com> Hello Experts, MPEG-2 controls VBV (13818-2, Annex C) using vbv_delay, max bit rate, and frame rate and all these parameters are specified in MPEG-2 elementary stream. For VBR operation, the VBV buffer is filled at max bit rate and overflow is not an error. MPEG-4 controls VBV (14496-2, Annex D) using Rvol(t), max bit rate, and vop time increment. I have following questions regarding VBV control in MPEG-4 1. At what rate is VBV buffer filled in VBR mode ? 2. Does standard defines Rvol(t) (instantaneous VOL channel bit rate) in elementary stream ? 3. Standard mentions that VBV mechanism is same as MPEG-2 (under some constraints), but also defines overflow as an error condition. Is this correct ? Regards, Manish From Ye-Kui.Wang nokia.com Wed Mar 22 10:52:14 2006 From: Ye-Kui.Wang nokia.com (Ye-Kui.Wang@nokia.com) Date: Wed Mar 22 09:22:26 2006 Subject: [Mp4-tech] [H.264] An trivial POC question In-Reply-To: <6A7AB4426712514F9E2AB28C407285FA372DA0@exchtp02.vntek.com.tw> Message-ID: <1C1F3D15859526459B4DD0A7A9B2268B01DA756F@trebe101.NOE.Nokia.com> It is possible for an IDR picture to be coded earlier than pictures earlier in output/display order. BR, YK ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] Sent: 21 March, 2006 03:21 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] An trivial POC question Dear experts, I had one trivial questions regarding POC, could you help me clarify. In spec. 8.2.1, it defines a range for POC from -2^31 to 2^31-1, is it necessary to have such large of range ? Since the IDR have POC as 0, and it seems to me that none of other pictures of the same sequence be displayed before this IDR, so POC of the others shall be greater than 0, why negative value is allowed ? Thanks, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060322/059f9e60/attachment.html From EricChuang via.com.tw Wed Mar 22 18:19:09 2006 From: EricChuang via.com.tw (EricChuang@via.com.tw) Date: Wed Mar 22 10:51:08 2006 Subject: [Mp4-tech] [H.264] An trivial POC question Message-ID: <6A7AB4426712514F9E2AB28C407285FA284A13@exchtp02.vntek.com.tw> Dear experts, Thanks very much for your correction, Eric -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Ye-Kui.Wang@nokia.com Sent: Wednesday, March 22, 2024 12:52 AM To: mp4-tech-bounces@lists.mpegif.org; mp4-tech@lists.mpegif.org Subject: RE: [Mp4-tech] [H.264] An trivial POC question It is possible for an IDR picture to be coded earlier than pictures earlier in output/display order. BR, YK _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] Sent: 21 March, 2006 03:21 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] [H.264] An trivial POC question Dear experts, I had one trivial questions regarding POC, could you help me clarify. In spec. 8.2.1, it defines a range for POC from -2^31 to 2^31-1, is it necessary to have such large of range ? Since the IDR have POC as 0, and it seems to me that none of other pictures of the same sequence be displayed before this IDR, so POC of the others shall be greater than 0, why negative value is allowed ? Thanks, Eric -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060322/c21586d0/attachment-0001.html From michael.zufferey itec.uni-klu.ac.at Wed Mar 22 12:01:00 2006 From: michael.zufferey itec.uni-klu.ac.at (Michael Zufferey) Date: Wed Mar 22 14:22:09 2006 Subject: WG: WG: [Mp4-tech] I need help to get vitooki Message-ID: <006101c64d9f$e9789350$d77acd8f@nbdell11> Hello, Please read the answer on how to get vitooki here below. Thanks. Best wishes, Michael Z. -----Urspr?ngliche Nachricht----- Von: Michael Kropfberger [mailto:michael@kropfberger.com] Gesendet: Mittwoch, 22. M?rz 2006 11:49 An: Michael Zufferey Betreff: Re: WG: [Mp4-tech] I need help to get vitooki I know there were some outages at Sourceforge directly, but now (Mar 22, 11:47) it worked fine: unset CVS_RSH cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/vitooki login cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/vitooki co vitooki for further problems/requests/.... please join the vitooki-devel@lists.sourceforge.org! regards, Mike > Hello Experts, > > I need help to get open source software vitooki, I tried the site on > sourceforge but it doesn't seem to be there. > > > > Rgds, > Michael > > > > > > > _____ > > Yahoo! Mail > Use > . yahoo.com> Photomail to share photos without annoying attachments. -- < When in doubt, THINK! > <----------------------------------------------------------------------> < DI Dr. Michael Kropfberger A-9074 Keutschach, Rauth 119 > < +43(650)6996000 michael _A T_ kropfberger.com http://www.kropfberger.com > <----------------------------------------------------------------------> From lancelot_lin himax.com.tw Thu Mar 23 19:17:58 2006 From: lancelot_lin himax.com.tw (lancelot_lin@himax.com.tw) Date: Thu Mar 23 15:22:08 2006 Subject: [Mp4-tech] A question about H.264 transform and scaling? Message-ID: Dear experts, Could anyone tell me how to calculate the weightScale (i,j) and what does it mean in spec 8.5.7. => LevelScale( m, i, j) = weightScale( i, j) * normAdjust( m, i, j) ( 8-312) Since it only appear in the latest spec (2005/03). Does this impact to the inverse transform and scaling? Best Regards , Lancelot ========================= Lancelot Lin 林展世 Himax Technologies Inc. Tel#: +886-2-33930877 Ext.22305 Fax#: +886-2-2341-3063 e-mail: lancelot_lin@himax.com.tw ========================= ************************************************************************** ***************************** This transmission is for the use of the intended recipient(s) only. The information contained in this communication may be business proprietary or confidential. If you have received this transmission in error, please notify the sender immediately and then delete it. If you are not the intended recipient, you must not keep, use, disclose, copy, or distribute this transmission without the author's prior permission. ************************************************************************** ***************************** -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060323/8da549ce/attachment.html From garysull windows.microsoft.com Thu Mar 23 14:26:25 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Fri Mar 24 19:34:09 2006 Subject: [Mp4-tech] A question about H.264 transform and scaling? In-Reply-To: Message-ID: <03CB47D9C3F8074498E4653F814D6E8F137B95@WIN-MSG-20.wingroup.windeploy.ntdev.microsoft.com> Good Lancelot, or good Lin, or good Lancelot Lin - Take heed, honest Lancelot; take heed honest Lin, Subclause 8.5.7 says how to calculate weightScale(i, j). This is specified in the text just above Equation 8-312. If you are implementing one of the original profiles of the standard (i.e., Baseline, Main, or Extended), you will eventually figure out that 1) weightScale(i, j) is always equal to 16 in those cases. 2) after multiplying by weightScale(i, j), the result is later shifted to the right by 4 bits more than what was specified in the prior version. Therefore the result will be the same (when using the original profiles of the standard) regardless of whether you use the new equations that involve weightScale(i, j) or the old ones that don't. When using a FRExt profile, weightScale(i, j) can have other values at the discretion of the encoder as specified in ScalingList4x4. Best Regards, Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of lancelot_lin@himax.com.tw Sent: Thursday, March 23, 2024 3:18 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] A question about H.264 transform and scaling? Dear experts, Could anyone tell me how to calculate the weightScale (i,j) and what does it mean in spec 8.5.7. => LevelScale( m, i, j) = weightScale( i, j) * normAdjust( m, i, j) ( 8-312) Since it only appear in the latest spec (2005/03). Does this impact to the inverse transform and scaling? Best Regards , Lancelot ========================= Lancelot Lin ??? Himax Technologies Inc. Tel#: +886-2-33930877 Ext.22305 Fax#: +886-2-2341-3063 e-mail: lancelot_lin@himax.com.tw ========================= ******************************************************************************************************* This transmission is for the use of the intended recipient(s) only. The information contained in this communication may be business proprietary or confidential. If you have received this transmission in error, please notify the sender immediately and then delete it. If you are not the intended recipient, you must not keep, use, disclose, copy, or distribute this transmission without the author's prior permission. ******************************************************************************************************* -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060323/910b7c8e/attachment.html From luwang astri.org Fri Mar 24 17:42:52 2006 From: luwang astri.org (Wang Lu) Date: Fri Mar 24 19:34:14 2006 Subject: [Mp4-tech] a question about POC Message-ID: Dear experts, I'm wondering about why does it need to define delta_pic_order_cnt_bottom in decoding process for picture order count type0? In which condition the encoder needs to use this variable to define the BottomFieldOrderCnt. Thanks a lot! Lu This message (including any attachments) is for the named addressee(s)'s use only. It may contain sensitive, confidential, private proprietary or legally privileged information intended for a specific individual and purpose, and is protected by law. If you are not the intended recipient, please immediately delete it and all copies of it from your system, destroy any hard copies of it and notify the sender. Any use, disclosure, copying, or distribution of this message and/or any attachments is strictly prohibited. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060324/9094a0d6/attachment.html From anjalikmahajan gmail.com Fri Mar 24 10:49:48 2006 From: anjalikmahajan gmail.com (Anjali Mahajan) Date: Fri Mar 24 19:34:18 2006 Subject: [Mp4-tech] a query about the MB partitions Message-ID: <5a2e52310603241049x21c37c6fua237a49102c35ce@mail.gmail.com> Hi Experts, I have a question about the macroblock partitions. I want to get the size of the partitions and their x,y co-ordinates after the macroblock is encoded. >From the trace file, I can get the number of partitions and the motion vectors for each partition. Could anybody tell me how and where to print this information in JM encoder? How to get the same inofrmation about the neighbouring macroblocks? Thanks in advance, Anjali -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060324/07f83085/attachment.html From bam iis.fraunhofer.de Fri Mar 24 22:35:21 2006 From: bam iis.fraunhofer.de (Oliver Baum) Date: Sat Mar 25 00:22:07 2006 Subject: [Mp4-tech] [Systems] BIFS scene generation with IM1 tools BifsEnc and MP4Enc Message-ID: <44246619.9030206@iis.fraunhofer.de> Hi all, I have a question concerning the generation of BIFS scenes (MP4 files) with IM1's BifsEnc and MP4enc: How do I embed/mux an audio stream into the MP4 file containing the scene description? And if I want to use AAC in which format do I need the audio data (ADIF, ADTS, raw)? And do I need to specify StreamFormat in MuxInfo ("AAC" seems to exist, while "PCM", which is used in the AABIFS conformance sequences, is *not* recognized by IM1)? Some details on what I tried up to now: ---8<---------------------=scene.txt=--------------- OrderedGroup { children [ Sound2D { intensity 1.0 spatialize FALSE source AudioClip { url [ "5" ] startTime 0 } } ] } UPDATE OD [ { ObjectDescriptorID 5 muxScript scene.src //// do I need this here?!? } ] --->8---------------------=scene.txt=--------------- and in scene.src (not pasting the IOD stuff): ---8<---------------------=scene.src=--------------- [...] { objectDescriptorID 5 esDescr[ { ES_ID 5 muxInfo MuxInfo { fileName L1R2.adif StreamFormat AAC } decConfigDescr DecoderConfigDescriptor { objectTypeIndication 64 // =AAC streamType 5 // AudioStream bufferSizeDB 5000 } slConfigDescr SLConfigDescriptor { } } ] } --->8---------------------=scene.src=--------------- I would appreciate *any* hints on this topic. Thanks in advance for replies! Best regards, Oliver Baum -- Dipl.-Ing. Oliver Baum Multimedia Transport Group, Audio Department Fraunhofer Institute for Integrated Circuits IIS Am Wolfsmantel 33 91058 Erlangen Germany Phone: +49 9131 776-319 Fax: +49 9131 776-398 From eechyang gmail.com Mon Mar 27 21:59:42 2006 From: eechyang gmail.com (Yang Cheng) Date: Mon Mar 27 15:10:11 2006 Subject: [Mp4-tech] How does JM 10.1 implements error concealment Message-ID: <2e282fb70603270559l484040a6naf2a86ec03da05bf@mail.gmail.com> Hi, everyone: Will your please recommend some material that introduce how JM decoder implements error concealment, thanks a lot! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060327/07e475fd/attachment.html From skb3 buffalo.edu Mon Mar 27 11:22:41 2006 From: skb3 buffalo.edu (skb3@buffalo.edu) Date: Tue Mar 28 15:22:07 2006 Subject: [Mp4-tech] How does JM 10.1 implements error concealment Message-ID: <1143476561.442811514ec36@mail3.buffalo.edu> Probably one of the best references is JM contribution on "Text Description of Joint Model Reference Encoding Methods & Decoding Concealment Methods". Best regards, Saurav Bandyopadhyay ________________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech- bounces@lists.mpegif.org] On Behalf Of Yang Cheng Sent: Monday, March 27, 2024 9:00 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] How does JM 10.1 implements error concealment Hi, everyone: Will your please recommend some material that introduce how JM decoder implements error concealment, thanks a lot! From research.264 gmail.com Tue Mar 28 00:57:13 2006 From: research.264 gmail.com (Anand Paul) Date: Tue Mar 28 15:22:13 2006 Subject: [Mp4-tech] test sequence for HDTV size ? Message-ID: hello everyone, where can i get standard test sequence (like foreman, news, stephen...etc) for HDTV / SDTV and EDTV size ? appreciate any links or information in this regard Anand -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060328/5d0d894a/attachment.html From sma com.dtu.dk Tue Mar 28 11:40:59 2006 From: sma com.dtu.dk (Shankar Manuel Aghito) Date: Tue Mar 28 15:22:17 2006 Subject: [Mp4-tech] How does JM 10.1 implements error concealment Message-ID: A very good reference is this http://ftp3.itu.ch/av-arch/jvt-site/2005_01_HongKong/N046r1.doc Recently it has been added the possibility of concealing entire pictures, as described in http://ftp3.itu.ch/av-arch/jvt-site/2005_07_Poznan/JVT-P072.zip best regards, Manuel -----Original Message----- From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Yang Cheng Sent: 27. marts 2006 16:00 To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] How does JM 10.1 implements error concealment Hi, everyone: Will your please recommend some material that introduce how JM decoder implements error concealment, thanks a lot! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060328/aca69d01/attachment.html From Ravi.Bulusu portalplayer.com Tue Mar 28 09:45:45 2006 From: Ravi.Bulusu portalplayer.com (Ravi Bulusu) Date: Tue Mar 28 23:51:08 2006 Subject: [Mp4-tech] H264 Transform Precision Requirement Message-ID: <4FBF117B12EDCC4ABDE1D4531719FCCD08B027@jupiter.portalplayersc.com> Hi All, I've asked this question earlier with not much luck. I'm hoping some expert will answer my question this time. I have been trying to understand the data precision requirement for H264 transform. There are a few intermediate stages in the transform computation in H264. The output of every stage has a constraint that the range of data should not exceed the values from -2^15 to 2^15 - 1. Since the same constraint is applicable for all the four (broadly classified) stages of transform computation, it does feel like these constaints are very loosely specified. Can someone tell me what would be a "strict" requirement on the input coefficients to the transform process given the intermediate constraints that have already been placed by the specification? In other words, if I mention that a 15 bit precision (i.e, -2^14 to 2^14-1) on the input coefficients is good enough to meet H264 transform compliance, would there be a challenging argument (or a proof) to disprove it? Another issue I have is with the precision on the output of this transform. Currently the specification restricts the output data in the range of [-512, 511]. However, since this is the residual data which is added to the motion compensated data (which happens to be the 8-bit unsigned pixel values), does it not suffice for the output of the transform to be restricted to [-256, 255]? I am unable to understand the reason behind the extra bit precision requirement at the output. thx, Ravi -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060328/cfc1030c/attachment.html From eechyang gmail.com Wed Mar 29 03:32:43 2006 From: eechyang gmail.com (Yang Cheng) Date: Tue Mar 28 23:51:12 2006 Subject: [Mp4-tech] testing slice mode in jm98 In-Reply-To: References: Message-ID: <2e282fb70603281132hbcc711agf7bfbdbb4043107a@mail.gmail.com> When slice mode are turn on one picture are encoded to be 30 slices with your parameters. Using slices will degrade the encoding efficiency due to reduced prediction within one frame. For example, spation intra prediction and motion vector is not allowed over slice boundries. 2006/3/1, David Oskardmay : > > Hello, > > I have been testing the performance of the JM 9.8 reference encoder with > various features enabled/disabled and I noticed a substantial quality > degradation when I turned on slice mode with a fixed number of Mb's per > slice. > > Config details: resolution: 640x480, 30 fps, high motion test video, > 1Mbps > SliceMode = 1 (fixed #mb in slice) > SliceArgument = 40 > > All other slice options are off. > > The artifacts during high motion periods are much worse with slice mode > enabled than with it off. Any idea why this is so? Is this a possible > bug? > Pilot error? > > Thanks in advance for any info. > Best Regards, > david > > > _______________________________________________ > NOTE: Please use clear subject lines for your posts. Include [audio, > [video], [systems], [general] or another apppropriate identifier to indicate > the type of question you have. > > Note: Conduct on the mailing list is subject to the Antitrust guidelines > found at > http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060329/723cb35e/attachment.html From stephane.pechard univ-nantes.fr Wed Mar 29 10:29:43 2006 From: stephane.pechard univ-nantes.fr (=?iso-8859-15?q?St=E9phane_P=E9chard?=) Date: Wed Mar 29 11:51:08 2006 Subject: [Mp4-tech] test sequence for HDTV size ? In-Reply-To: References: Message-ID: <200603291029.52950.stephane.pechard@univ-nantes.fr> > hello everyone, hello > where can i get standard test sequence (like foreman, news, stephen...etc) > for HDTV / SDTV and EDTV size ? some are freely available here : ftp://ftp.ldv.e-technik.tu-muenchen.de/pub/test_sequences/ st?phane -- ~ st?phane p?chard ~ phd student | irccyn-ivc | france ~ http://www.irccyn.ec-nantes.fr/~pechard ~~~~~~ pgp : hkp://subkeys.pgp.net ~~~~~~ -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 309 bytes Desc: not available Url : /pipermail/mp4-tech/attachments/20060329/ab19222d/attachment.bin From sma com.dtu.dk Wed Mar 29 11:05:25 2006 From: sma com.dtu.dk (Shankar Manuel Aghito) Date: Wed Mar 29 11:51:13 2006 Subject: [Mp4-tech] test sequence for HDTV size ? Message-ID: I am not sure that the sequences you are looking for exist in HD resolution. I think there is a good chance that these HD sequences will be used a lot in the future. http://www.ldv.ei.tum.de/liquid.php?page=70 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060329/2af1a119/attachment.html From research.264 gmail.com Wed Mar 29 18:46:05 2006 From: research.264 gmail.com (Anand Paul) Date: Wed Mar 29 11:51:17 2006 Subject: [Mp4-tech] test sequence for HDTV size ? In-Reply-To: <44297704.308ccbea.473b.5eee@mx.gmail.com> References: <44297704.308ccbea.473b.5eee@mx.gmail.com> Message-ID: hi Vinay, Thanks for that information, i went thru the procedure to access mini-movie trailers from StEM and available materials (mini-movie with trailers) good that they have HDTV size sequence, but i doubt will that be one of the standard test sequence like foreman,stephen,news...etc..! (as i am looking for one of those). I wanna make sure of this as the procedure to access those sequence is quite complicated. or if anyone out there have any standard video sequence for HDTV/SDTV/EDTV plz let me some info.. regards Anand On 3/29/06, Vinay Sathe wrote: > > Among other sequences, I suggest getting hold of StEM sequence from > Digital Cinema Initiative, LLC. This sequence was shot by Hollywood studios > specifically for digital video community. > > ------------------------------ > *From:* mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] > *On Behalf Of *Anand Paul > *Sent:* Monday, March 27, 2024 8:57 AM > *To:* mp4-tech@lists.mpegif.org > *Subject:* [Mp4-tech] test sequence for HDTV size ? > > > hello everyone, > > where can i get standard test sequence (like foreman, news, stephen...etc) > for HDTV / SDTV and EDTV size ? > > appreciate any links or information in this regard > > Anand > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060329/a7eaf3f0/attachment.html From oelbaum tum.de Wed Mar 29 14:25:07 2006 From: oelbaum tum.de (Tobias Oelbaum) Date: Wed Mar 29 14:04:07 2006 Subject: AW: [Mp4-tech] test sequence for HDTV size ? In-Reply-To: <200603291029.52950.stephane.pechard@univ-nantes.fr> Message-ID: <006801c6532b$d0d42120$5769bb81@ldvoelbaum> Dear List Members > > where can i get standard test sequence (like foreman, news, > stephen...etc) > > for HDTV / SDTV and EDTV size ? Please have a look at http://www.ldv.ei.tum.de/page70 You can download some HD sequences there, but even more interesting sequences were produceds by SVT (Swedish Television) and are currently distributed. To read more about this sequences get the following PDF document: http://www.ldv.ei.tum.de/media/files/homes/oelbaum/forschung/mpeg/SVT_Mult iFormat_v10.pdf Up to now, there are only a few 'standard' test sequences in HD meaning HD test sequences everybody uses. Most used ones are called 'city', 'crew', 'harbour' and 'soccer', these are available at least for everybody involved in JVT work. Some companies also sell HD test material, one can be found at www.videatis.com Greetings Tobias ----- Dipl.Ing. Tobias Oelbaum Lehrstuhl f?r Datenverarbeitung Institute for Data Processing Technische Universit?t M?nchen Munich University of Technology Tel: +49 (0)89 289 23620 Fax: +49 (0)89 289 23600 Email: oelbaum@tum.de Homepage: http://www.ldv.ei.tum.de/page50 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 4907 bytes Desc: not available Url : /pipermail/mp4-tech/attachments/20060329/209deb8a/smime-0001.bin From pankaj_bajpai_iet operamail.com Wed Mar 29 14:34:17 2006 From: pankaj_bajpai_iet operamail.com (pankaj bajpai) Date: Wed Mar 29 14:04:13 2006 Subject: [Mp4-tech] Query regarding Mpeg4 - part2 Message-ID: <20060329133417.6D91423D27@ws5-3.us4.outblaze.com> Dear All, I have following query: 1. The standard (part-2, video) says following regarding gob_resync_marker gob_resync_marker: This is a fixed length code of 17 bits having the value ‘0000 0000 0000 0000 1 which may optionally be inserted at the beginning of each gob_layer(). Its purpose is to serve as a type of resynchronization marker for error recovery in the bitstream. The gob_resync_marker codes may (and should) be byte aligned by inserting zero to seven zero-valued bits in the bitstream just prior to the gob_resync_marker in order to obtain byte alignment. Now from this I conclude that gob_resync_marker should appear at byte align boundary. This seems logical also as for resynchronization the decoder should check codes from byte. But in description of pseudocode for gob_layer(): gob_layer() { gob_header_empty = 1 if(gob_number != 0) { if (next_bits() == gob_resync_marker) { gob_header_empty = 0 gob_resync_marker gob_number gob_frame_id quant_scale } } for(i=0; i Message-ID: <03CB47D9C3F8074498E4653F814D6E8F20DAC5@WIN-MSG-20.wingroup.windeploy.ntdev.microsoft.com> When we use the word "shall" we are saying that something MUST be done. When we use the word "may" we are saying that something CAN be done (but doesn't have to be done). When we use the word "should" we are saying that we would prefer that something be done (but it doesn't have to be done). So the summary is that GOB resynch markers can appear at non-byte-aligned positions, but it is preferred that they are not. This is allowed, but discouraged, behavior. Best Regards, Gary Sullivan +> -----Original Message----- +> From: mp4-tech-bounces@lists.mpegif.org +> [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of pankaj bajpai +> Sent: Wednesday, March 29, 2024 5:34 AM +> To: mp4-tech@lists.mpegif.org +> Subject: [Mp4-tech] Query regarding Mpeg4 - part2 +> +> Dear All, +> I have following query: +> +> 1. The standard (part-2, video) says following regarding +> gob_resync_marker +> gob_resync_marker: This is a fixed length code of 17 bits +> having the value +> ‘0000 0000 0000 0000 1’ which may optionally +> be inserted at the beginning +> of each gob_layer(). Its purpose is to serve as a type of +> resynchronization +> marker for error recovery in the bitstream. The +> gob_resync_marker codes may +> (and should) be byte aligned by inserting zero to seven +> zero-valued bits in the +> bitstream just prior to the gob_resync_marker in order to +> obtain byte alignment. +> +> Now from this I conclude that gob_resync_marker should +> appear at byte align +> boundary. This seems logical also as for resynchronization +> the decoder should check +> codes from byte. +> +> But in description of pseudocode for gob_layer(): +> gob_layer() { +> gob_header_empty = 1 +> if(gob_number != 0) { +> if (next_bits() == gob_resync_marker) { +> gob_header_empty = 0 +> gob_resync_marker +> gob_number +> gob_frame_id +> quant_scale +> } +> } +> for(i=0; i macroblock() +> if(next_bits() != gob_resync_marker && +> nextbits_bytealigned() == gob_resync_marker) +> while(!bytealigned()) +> zero_bit +> gob_number++ +> } +> +> It seems that +> (next_bits() != gob_resync_marker && nextbits_bytealigned() +> == gob_resync_marker) +> goob_resync_marker can appear from non-byte align position. +> +> These 2 descriptions seems contradictory. I was testing the +> version_1 simple profile +> sequences and I come across these for some of the sequences +> like vcon-ge16-L1.cmp +> and some others. +> +> Please clearify this matter. +> +> 2. Is it true of all other parts or specific to +> ShortVideoHeader mode. +> I find next_bits() at many places , but except for SVH, +> most of the +> places have some mechanism (like next_start_code()) to byte align +> before start code or marker. +> +> With regards +> +> +> -- +> _______________________________________________ +> Surf the Web in a faster, safer and easier way: +> Download Opera 8 at http://www.opera.com +> +> Powered by Outblaze +> +> _______________________________________________ +> NOTE: Please use clear subject lines for your posts. Include +> [audio, [video], [systems], [general] or another +> apppropriate identifier to indicate the type of question you have. +> +> Note: Conduct on the mailing list is subject to the +> Antitrust guidelines found at +> http://www.mpegif.org/public/documents/vault/mp-out-30042-Ant +> itrust.php +> From lattecafe gmail.com Thu Mar 30 14:29:31 2006 From: lattecafe gmail.com (Latte Coffee) Date: Thu Mar 30 14:46:12 2006 Subject: [Mp4-tech] Stardard/official MPEG4 video bitstream for Conformance Test Message-ID: Dear all, Does anyone know where can I download/purchase some Stardard/official MPEG4 video bitstream for Conformance Test? I am looking for MPEG4 Simple profile Level 0 bitstream. I have downloaded couples of bitstream from ISO ftp but those bitstream's header show they are Main Profile actually. I am glad if anyone can point me to obtain some standard bitstream. Thx. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060330/84dab238/attachment.html From subhac noida.interrasystems.com Thu Mar 30 13:57:21 2006 From: subhac noida.interrasystems.com (Subhadeep Chatterjee) Date: Thu Mar 30 14:46:16 2006 Subject: [Mp4-tech] Unrestricted motion compensation in interlaced VOPs in MPEG4 Video Message-ID: <442B9669.8020108@noida.interrasystems.com> Hi All, In the MPEG4 video Specification (ISO/IEC 14496-2:1999/Amd.1:2000(E)) it is specifed in section 7.6.1.5 that "Macroblocks of interlaced VOP (interlaced = 1) are padded according to subclauses 7.6.1.1 through 7.6.1.3. The vertical padding of the luminance component, however, is performed for each field independently. A sample outside of a VOP is therefore filled with the value of the nearest boundary sample of the same field." But the reference decoder from microsoft seems to not follow this and implements the padding process for interlaced VOPs in the same manner as in the case of progressive VOPs. Can anyone tell me which is the right way to implement unrestricted motion compensation in the case of interlaced sequence? Thanks and regards, Subhadeep Chatterjee -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060330/030f4c2f/attachment.html From birgera rdr.com Thu Mar 30 09:23:18 2006 From: birgera rdr.com (Alex Birger) Date: Thu Mar 30 14:46:21 2006 Subject: [Mp4-tech] test sequence for HDTV size ? In-Reply-To: Message-ID: Hi Anand, Muenchen Technical University carries some HD test sequences. You may find more on the issue there: http://www.ldv.ei.tum.de/liquid.php?page=70 Regards, Alex Birger _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Anand Paul Sent: Wednesday, March 29, 2024 5:46 AM To: Vinay Sathe ; mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] test sequence for HDTV size ? Importance: Low hi Vinay, Thanks for that information, i went thru the procedure to access mini-movie trailers from StEM and available materials (mini-movie with trailers) good that they have HDTV size sequence, but i doubt will that be one of the standard test sequence like foreman,stephen,news...etc..! (as i am looking for one of those). I wanna make sure of this as the procedure to access those sequence is quite complicated. or if anyone out there have any standard video sequence for HDTV/SDTV/EDTV plz let me some info.. regards Anand On 3/29/06, Vinay Sathe < vsathe@gmail.com> wrote: Among other sequences, I suggest getting hold of StEM sequence from Digital Cinema Initiative, LLC. This sequence was shot by Hollywood studios specifically for digital video community. _____ From: mp4-tech-bounces@lists.mpegif.org [mailto: mp4-tech-bounces@lists.mpegif.org] On Behalf Of Anand Paul Sent: Monday, March 27, 2024 8:57 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] test sequence for HDTV size ? hello everyone, where can i get standard test sequence (like foreman, news, stephen...etc) for HDTV / SDTV and EDTV size ? appreciate any links or information in this regard Anand _______________________________________________ NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have. Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060330/5e1c4816/attachment.html From research.264 gmail.com Fri Mar 31 11:11:10 2006 From: research.264 gmail.com (Anand Paul) Date: Fri Mar 31 09:40:06 2006 Subject: [Mp4-tech] Test Sequence for Super HDTV Size video ! Message-ID: hello everybody, i appreciate all the useful information and links with regards to HDTV size video sequence. - Do any one of you know - where can i find Test sequence for "Super HDTV size" video. re-searching with you Anand P. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060331/294b6209/attachment.html From garysull windows.microsoft.com Fri Mar 31 02:34:53 2006 From: garysull windows.microsoft.com (Gary Sullivan) Date: Fri Mar 31 11:16:06 2006 Subject: [Mp4-tech] Unrestricted motion compensation in interlaced VOPs inMPEG4 Video In-Reply-To: <442B9669.8020108@noida.interrasystems.com> Message-ID: <03CB47D9C3F8074498E4653F814D6E8F2C26B6@WIN-MSG-20.wingroup.windeploy.ntdev.microsoft.com> I am surprised to hear you make that quote from the text. My recollection was that the padding was done in the same manner as a "progressive" case. We should check the spec on this issue. Best Regards, -Gary Sullivan ________________________________ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Subhadeep Chatterjee Sent: Thursday, March 30, 2024 12:27 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] Unrestricted motion compensation in interlaced VOPs inMPEG4 Video Hi All, In the MPEG4 video Specification (ISO/IEC 14496-2:1999/Amd.1:2000(E)) it is specifed in section 7.6.1.5 that "Macroblocks of interlaced VOP (interlaced = 1) are padded according to subclauses 7.6.1.1 through 7.6.1.3. The vertical padding of the luminance component, however, is performed for each field independently. A sample outside of a VOP is therefore filled with the value of the nearest boundary sample of the same field." But the reference decoder from microsoft seems to not follow this and implements the padding process for interlaced VOPs in the same manner as in the case of progressive VOPs. Can anyone tell me which is the right way to implement unrestricted motion compensation in the case of interlaced sequence? Thanks and regards, Subhadeep Chatterjee -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/mp4-tech/attachments/20060331/5b4622fe/attachment.html From mp4-tech-owner lists.mpegif.org Fri Mar 31 18:09:20 2006 From: mp4-tech-owner lists.mpegif.org (MPEGIF List Admins) Date: Sat Apr 1 17:04:06 2006 Subject: [Mp4-tech] FW: [Video] Mp4-tech Digest, Vol 32, Issue 36 Message-ID: <004b01c654dd$78a14ba0$0301a8c0@corp.intertrust.com> Forward of auto-discarded message -----Original Message----- From: thomas.dove@exgate.tek.com [mailto:thomas.dove@exgate.tek.com] Sent: Friday, 31 March 2024 15:42 To: mp4-tech@lists.mpegif.org Subject: RE: [Video] Mp4-tech Digest, Vol 32, Issue 36 Anand, Our company, Vqual, sells a wide range of HD test streams which cover a wide range of content to test and stress encoders. Please have a look at: http:\\www.vqual.com\Vclips.html Click on 'List of Clips' in the left hand menu, then look at stream numbers - VC-013 (up-sampled HD) - VC-015 Synthetic HD with known motion, etc. - VC-031 true HD, filmed in USA - VC-032 true HD, filmed in Asia These are supplied on hard disk (as they are up to 170 GB in size) You can download extracts from the user manuals for each, to see what they contain and the level of documentation. Thomas Dove Senior Manager, Compressed Video Vqual Ltd. (a Tektronix company) Trym Lodge, 1 Henbury Road, Bristol BS9 3HQ UK E-mail: thomas.dove@tektronix.com Direct: +44 (0)117 373 6255 Mobile: +44 (0)7771 560 799 Sales fax: +44 (0)117 373 6269 Web: www.vqual.com ************************************************************ The information in this message is confidential and may be privileged. It is intended for the addressee alone. If you are not the intended recipient it is prohibited to disclose, use or copy this information. Please contact the sender immediately should this message have been transmitted incorrectly. ************************************************************ _____ From: mp4-tech-bounces@lists.mpegif.org [mailto:mp4-tech-bounces@lists.mpegif.org] On Behalf Of Anand Paul Sent: Wednesday, March 29, 2024 5:46 AM To: Vinay Sathe ; mp4-tech@lists.mpegif.org Subject: Re: [Mp4-tech] test sequence for HDTV size ? Importance: Low hi Vinay, Thanks for that information, i went thru the procedure to access mini-movie trailers from StEM and available materials (mini-movie with trailers) good that they have HDTV size sequence, but i doubt will that be one of the standard test sequence like foreman,stephen,news...etc..! (as i am looking for one of those). I wanna make sure of this as the procedure to access those sequence is quite complicated. or if anyone out there have any standard video sequence for HDTV/SDTV/EDTV plz let me some info.. regards Anand On 3/29/06, Vinay Sathe < vsathe@gmail.com> wrote: Among other sequences, I suggest getting hold of StEM sequence from Digital Cinema Initiative, LLC. This sequence was shot by Hollywood studios specifically for digital video community. _____ From: mp4-tech-bounces@lists.mpegif.org [mailto: mp4-tech-bounces@lists.mpegif.org] On Behalf Of Anand Paul Sent: Monday, March 27, 2024 8:57 AM To: mp4-tech@lists.mpegif.org Subject: [Mp4-tech] test sequence for HDTV size ? hello everyone, where can i get standard test sequence (like foreman, news, stephen...etc) for HDTV / SDTV and EDTV size ? appreciate any links or information in this regard Anand