Td corrigé VCEG-O01d1.doc - ITU pdf

VCEG-O01d1.doc - ITU

1.1.4 Meeting invitation for this meeting VCEG-N-TD-0 ..... Insert information into the header regarding buffering with multiple leaky bucket support. Theoretical ...




part of the document



ITU - Telecommunications Standardization Sector
STUDY GROUP 16 Question 6
Video Coding Experts Group (VCEG)
_________________
14th Meeting: Santa Barbara, CA, USA, 24-27 Sep., 2001Document VCEG-N-88 draft 1
Filename: VCEG-N88d1.doc
Generated: 27 September ’01
Question:Q.6/SG16 (VCEG)Source:Gary Sullivan, Q.6 Rapporteur Microsoft Corp. One Microsoft Way Redmond, WA 98052 USA
Thomas Wiegand, Q.6 Associated Rapporteur Heinrich Hertz Institute (HHI) Einsteinufer 37 10587 Berlin, Germany Tel: Fax: Email:
Tel: Fax: Email: +1 (425) 703-5308 +1 (425) 936-7329  HYPERLINK "mailto:Gary.Sullivan@itu.int" Gary.Sullivan@itu.int
+49 - (0)30 - 31002 617 +49 - (0)30 - 392 72 00 wiegand@hhi.deTitle:Meeting Report of the 14th VCEG Meeting, Sept. 24-27, 2001Purpose:Information_____________________________
Detailed Agenda and Disposition of Action Items
Also See Annexes:
Annex A: Documents Addressed [VCEG-N00r3] (92 documents in all – an all-time record despite adverse circumstances)
Annex B: List of Attending Experts [VCEG-N02r1] (59 registered attendees – an all-time record despite adverse circumstances)
Annex C: Liaison Letter to MPEG [VCEG-N82] with attachments:
H.26L Project Goals and Proposed Schedule [VCEG-N86]
Proposed Terms of Reference for Joint Work [VCEG-N87]
Rules for Proposals to H.26L and TML [VCEG-L33r5]
IPR Strategy and Guidelines for H.26L [VCEG-N70r3]
Opening plenary
Organizational Items
Meeting logistical information
Generating attendee list HYPERLINK "VCEG-N02.doc"VCEG-N02
Reviewing experts list HYPERLINK "VCEG-N03.doc"VCEG-N03 Remarks: A very healthy number of experts were present, and we were pleased that most people could make it to the meeting, despite recent socio-political circumstances. Expression of regret that some experts not present. Group to endeavor to take this fact into account in its work. In particular, holding a meeting without Gisle Bjøntegaard present is not something we would like to repeat in the future.
Meeting invitation for this meeting HYPERLINK "VCEG-NT0r1.doc"VCEG-N-TD-0
Comments regarding ITU-T patent disclosure policy  HYPERLINK "http://www.itu.int/ITU-Databases/TSBPatent/" http://www.itu.int/ITU-Databases/TSBPatent/ Reference VCEG-L37r2 for VCEG IPR coordination & tracking Report received by email: Telenor citing 2.2 re UVLC Report received verbally by a third party (HHI): Netergy Microsystems citing 2.2 re Multiframe Prediction Report received verbally: Philips citing 2.2 re 2-D VLC run-length+value Some companies are willing to license royalty-free for H.26L (essentially per 2.1) if others agree to do so. We ask if parties holding IPR would be willing to license royalty-free if all others with IPR are willing to do the same. Our members may also wish to investigate the potential performance penalty of not including technology inconsistent with the royalty-free baseline goal. Note that only the baseline of the standard has the royalty-free goal, and inclusion in the standard as a whole should be on technical merit basis only.
Review of previous meeting report
Report of Q.6 work at Porto Seguro SG16 HYPERLINK "VCEG-N01.doc"VCEG-N01
Document review HYPERLINK "VCEG-N00r0.doc"VCEG-N00 N34, N35, N55 were without presenters and were considered by the group in absentia, N64 registration was withdrawn.
Review of meeting plan HYPERLINK "VCEG-NT1r0.doc"VCEG-N-TD-1
Future meeting plans
JVT or VCEG Rapporteur’s Group Meetings To be further discussed during the meeting.
SG16 Meeting, Geneva, Feb 5-15 ‘02
SG16 Meeting, Geneva, Oct 15-25 ‘02
Report of ad-hoc committees
Test model and software HYPERLINK "VCEG-N04.doc"VCEG-N04 Status reported – very significant progress.
H.26L development HYPERLINK "VCEG-N05.doc"VCEG-N05 Remarks: Status reviewed and a number of action items were raised in the report for consideration.
H.26L complexity HYPERLINK "VCEG-N06.doc"VCEG-N06 To be discussed later in the context of related contributions
H.26L profiles & applications HYPERLINK "VCEG-N07.doc"VCEG-N07 No real action reported, current status reported.
H.26L loop filtering  HYPERLINK "VCEG-08.doc" VCEG-N08 To be discussed later in the context of related contributions
H.26L motion rendition  HYPERLINK "VCEG-N09.doc" VCEG-N09 Status reported – little coordinated work, but a significant number of contributions in the technical area. Side activity conducted during the meeting to study the direct vs. subsequent interpolation process in software and document, encoder and decoder and describe that status for group consideration.
Liaison statements and collaborative letters received
From ISO/IEC JTC1/SC29/WG11 HYPERLINK "VCEG-N12.doc"VCEG-N12 (joint video team) HYPERLINK "VCEG-N13.doc"VCEG-N13 (video coding test results) HYPERLINK "VCEG-N14.doc"VCEG-N14 (proposed workplan) HYPERLINK "VCEG-N15.doc"VCEG-N15 (digital cinema) Excellent test results reported along with proposed workplan, scope and requirements, etc.. Remarks: How do we ensure technical alignment on an ongoing basis, without common text? There should be a single master document maintained by an editor within the JVT… Side activity conducted during the meeting to discuss and recommend a course of action in regard to the proposed JVT formation, workplan schedule, goals, requirements…
Coordination with other organizations
With ISO/IEC JTC1/SC29WG11  HYPERLINK "VCEG-N11.doc" VCEG-N11, VCEG-N75 Thank Heiko Schwarz for work in generating the results.
With ITU-T WP2 / SG 16 [no action]
With IETF AVT [no action]
With 3GPP/3GPP2 [no action]
H.26L Project Development
General Status of Work  HYPERLINK "VCEG-N05.doc" VCEG-N05 (ad-hoc report) HYPERLINK "VCEG-N10.doc"VCEG-N10 (test model document)
Patent/IPR Practices  HYPERLINK "VCEG-N66.doc" VCEG-N66, VCEG-N70 Note of problems of prior IPR data collection and tracking efforts (e.g. in JPEG-2000) in terms of collection of statements that may not apply to the actual adopted content or lack of information due to provision of only blanket “general” statements (without reference to any particular standard) or lack of any filing of information until (or after) final approval or filing of statements just in case IPR may be present (without any actual IPR being in a standard). Remarks: Concrete action requested to be taken to address these issues does not seems clear (e.g., in regard to how the proposed form addresses the specific problem issues that were asserted) – side activity conducted to propose a specific course of action.
Performance Evaluation
Compression Performance Without Losses or Errors  HYPERLINK "VCEG-N11.doc" VCEG-N11 (MPEG video coding test entry) HYPERLINK "VCEG-N13.doc"VCEG-N13 (MPEG video coding test results) Excellent results reported for H.26L performance.  HYPERLINK "VCEG-N18.doc" VCEG-N18 (compression tests vs. MPEG-4 and H.263++) N18 contains comparison of H.26L vs. non-optimized MPEG-2 and highly-optimized H.263++ High Latency Profile and highly-optimized MPEG-4 Advanced Simple Profile – similar results to N11.
Performance with Losses / Errors  HYPERLINK "VCEG-N67.doc" VCEG-N67 (various conditions) Tested H.26L performance with errors. Not using multiple slices per picture or extra intra updates, only one data partition. Performing bitstream encoding without optimization for error/loss robustness. Observed that the decoder often stops operation upon encountering some errors in data. (Remark: The decoder is not really written to handle bit errors.) In UVLC and CABAC testing, the CABAC decoder was more likely to stop the decoding process in response to bit errors. There is no real error recovery designed into the present code at present, and there is very little inherent error resilience to the design. Action item: The group should review and clarify its various common conditions descriptions for error/loss resilience testing.
Performance with RTP/IP on 3GPP/3GPP2  HYPERLINK "VCEG-N37.doc" VCEG-N37 (simulation conditions) Some modifications of M77 test conditions are proposed. Change of overhead for CDMA-2000 and UMTS. New software included to be included in test conditions (not previously available). Adopt test condition changes with new software as proposed.  HYPERLINK "VCEG-N38.doc" VCEG-N38 (simulation results) Compare different data partitioning schemes, entropy coding methods, packet length. Damaged RTP packets ignored. PSNR calculated for all frames of original sequence. One encoding, 10 different decoding starting point. I-frame assumed error free. Copy MB in case of losses. D+L*R optimization with MSE distortion over N different decodings (ref: Wiegand thesis, straightforward trick robust to some difficult-to-analyze aspects of design such as ¼-pel motion and loop filter). Used a fixed L constant times step size (different constant than in current H.26L). For regular MB intra updates, short slices help, no data partitioning better than data partitioning. For optimized updates, 4 partitions per slice seems best with one slice per frame. Data runs were too short for really firm conclusions. Remark: delay impact of one slice per frame. Conclusions: Customizing the coding for the conditions can help significantly. For regular intra updates, DP is inferior to no DP? Appropriate payload length selection is important. Would be good to have a simpler error-resilient MB mode selection method. Put this multi-decoding technique in the reference software? We should have reference software capability to generate more robust encoding (now have ability to choose packet length in macroblocks or bits and to apply regular intra update patterns). Note that this depends on the error concealment in the decoder. (Decoder currently has copy MB for concealment.) Significant improvement shown from the parallel-decoding technique. Try the Zhang and Rose method? Also see N39 and N50 discussions with similar optimization method. Concern that multi-decoding method may not be the most practical. What about using this with errors in intra frame and flexible MB number in slice and different error concealment? Disposition: Adopt conditioned on checking and reporting on operation with new error concealment method and slice size based on number of bytes threshold. [End of Subject Areas Addressed on Monday]
Performance for H.324/M Error-Prone Environments  HYPERLINK "VCEG-N39.doc" VCEG-N39 Similar techniques of N38 and N50. Comparison of data partitioning schemes, entropy coding, packet length. Use AL3 SDUs, ignore damages SDUs. Most conditions similar to N38. Results: Gives reasonable indication, but not statistically consistent, for regular intra update small slice lengths are advantageous. No DP outperforms DP with regular intra update. Optimized parallel-decoding model provides significant gain. At low rates, the optimized model performs very slightly worse in avg PSNR than regular intra update – perhaps because criterion is not avg PSNR but rather minimum MSE of each frame? Remark: The encoder may not be aware of the channel quality in the H.324/M environment.  HYPERLINK "VCEG-N52.doc" VCEG-N52 Aspects of coding with error resilience. Primary distinction between mobile environment and internet packet environment is maximum packet size (500-1500 bytes on IP vs. ~254 on mobile) and the practical expectation of the presence of bit errors within packet payloads. The ideal is to have context-free packet decoding capability. Picture header is a problem for that, and a method of conveying picture header out of band was presented, using a concept of parameter sets sent in advance, with packets referencing these parameter sets. Temporal reference is available in the packet header on RTP but needs to be put in the packets on H.223 mux. Advocating against a fixed picture clock rate assumption. Proposal to unify the two current data partitioning schemes into one. Data partitioning scheme with three packets per slice proposed 1) slice header and MVs, 2) intra MB data coeff’s, 3) inter MB data coeff’s. Each partition gets its own RTP packet. Any empty packets are not sent (e.g. for inter MB coeff’s in an intra slice). Possibility to switch coding modes (e.g. I vs. P vs. B) on slice basis. Picture ID parameter included. Out of order slices not allowed. SDP syntax proposed for sequence-level information. Need expressed for H.263 Annex N-style feedback messages. Use of UVLC for coding of header information as opposed to fixed-length approach. Side Activity conducted to refine and converge with other proposals and document the informative vs. normative changes to be proposed.
Performance for H.323/Internet  HYPERLINK "VCEG-N50.doc" VCEG-N50 Similar techniques of N39 and N38. Data partitioning with three slices or three partitions per N52. No optimization of packet length. Lost packets ignored. Rather than encoding 4000 frames, just encoded once and shift the error pattern starting point. (Should we change the common conditions use this concept?) DP seemed worse than non-DP (didn’t try things like repeating short high-priority packets). Remark: Channel-quality feedback (e.g., RTCP error reports) might be a more accurate model of actual system operation and provide ability to adapt to non-stationarity of statistics.  HYPERLINK "VCEG-N52.doc" VCEG-N52 [see above]  HYPERLINK "VCEG-N63.doc" VCEG-N63 Investigation of TML 8.4 with addition of error concealment method of N62. Used Q15-I-61 (the latest delivered common conditions for internet). Slices—one macroblock line each. No data partitioning. Instead of 4000 coded frames, used different starting positions of packet loss pattern for single encoding. Applied regular (new non-propagating style) intra GOB updates. Used interleaving of gob rows per prior Stephan H.263 design. Picture header repeated in every slice. Constant QP. Calculated PSNR for every frame of original although not all were coded (not directly comparable to TUM results). How to handle issue of different reference frames in neighbor MBs? Ans: It is handled, but details not remembered exactly. Results: 0.5 to 1.0 dB improvement over simply MB copy method (some relatively stationary sequences showed no gain). Adopted.
Network Adaption Layer and Error Resilience
Constrained Intra Prediction  HYPERLINK "VCEG-N56.doc" VCEG-N56 Information on constrained intra prediction. It was pointed out that the restriction on Intra prediction could be also achieved by an encoder decision. In the constrained intra prediction mode, “gaps” are generated in the VLC table due to the existence of forbidden prediction modes. Hence, the VLC table is adjusted accordingly for test of inefficiency of current approach. The decision taken was to not alter the syntax because of the very minor gains achieved with the alteration of the VLC table. It was suggested to add a non-predicted intra block type, which is a dc prediction mode with a constant value of 128, and this will be the dc prediction mode (the mode with the shortest codeword) at the start of a slice (as is the case now). We request experimental data as to whether never predicting intra blocks from neighboring P blocks will significantly harm R-D performance. It was suggested to conduct an experiment to have only one intra macroblock type in the design.
Handling of Erroneous Bitstreams  HYPERLINK "VCEG-N61.doc" VCEG-N61 Example: May have recording of session for later use, with errors/losses in received file. The later use might be in an environment not robust to errors (e.g. a terminal that operates in a generally error-free environment). What to do about losses and bit errors in such a scenario? Proposal that all decoders must be capable of return to normal decoding eventually after a loss (not necessarily to error-free or even meaningful images). After loss of a slice decoder should start from next slice. Devices storing bitstreams should put pictures in order, but slices may be out of order. Devices storing bitstreams should support a per-slice indicator flag to indicate bit errors within each slice content. Should also be able to put the bad slice indicator in the slice header itself. Require display of the recovered pictures? (presumably not) Shouldn’t a decoder have the right to consider some bitstreams simply unacceptable? Defer normative decoding requirements to later consideration. How about the bad slice indicator bit – does the group support that? Defer to header discussion.
Header Issue
Picture sync  HYPERLINK "VCEG-N21.doc" VCEG-N21 No immediate action required for software. It was mentioned that the document does not match the software. For the picture synchronization codes, it was decided to match the TML document to the software (Stephan Wenger).
Picture and Slice Headers  HYPERLINK "VCEG-N59.doc" VCEG-N59, VCEG-N72, VCEG-N73 Repetition of picture header info, badSliceIndicator, .. Similar spirit to N52 in some aspects. Side activity conducted to refine and converge with other proposals.
VBV/HRD [Tues. – RM]
Enhanced Buffer Model  HYPERLINK "VCEG-N58.doc" VCEG-N58r1 Insert information into the header regarding buffering with multiple leaky bucket support. Theoretical analysis showing use of piecewise-linear relation between minimum buffering delay and minimum rate for playback. 1+8N Bytes transmitted as binary numbers for a sequence with N being the points on the R-B curve. Comment: buffer information should not go into the video syntax. Comment: consistent with MPEG vbv design as opposed to H.263 HRD design in terms of timing preservation. Adopt vbv multiple leaky buffer compliance feature. Encourage development of a rate control.
Comments  HYPERLINK "VCEG-N68.doc" VCEG-N68 Reference made to H.263 HRD. Description of problems in that HRD design. Decoding speed can exceed proper display rate. Requires post-decoding buffering in model in some cases if timing preservation is to be preserved. Has a problem with being consistent with maximum frame rate requirements of profile/level or other negotiated max frames/sec. Should take into account post decoder buffering when designing HRD for H.26L. Comment: real decoder would rather store frames in compressed domain. But sometimes it is important to have decoded frames available. IPR: no
Complexity [Not Monday - MH]  HYPERLINK "VCEG-N06.doc" VCEG-N06 (ad-hoc report) No coordinated activity since Austin. Status reported. Presence of a number of contributions in the area noted. Emphasis on practical real-time encoding and decoding goal.
Benchmark analysis  HYPERLINK "VCEG-N23.doc" VCEG-N23 Complexity analysis of decoder on a RISC 16-bit DSP platform. Cycles measurement from implementation. Emphasis on worst-case analysis. Intra and Inter macroblock decoding operations analyzed. Decoder needs factor of 4-5 more cycles than MPEG-4 SP. Primary Bottlenecks: sub-pel motion interpolation (36% of cycles and 8 kbytes of RAM for a macroblock) and de-blocking filter (24% of cycles). Additional considerations worth attention: transform and VLD. Comparison to MPEG-4 ASP suggested (since that is a higher-complexity, enhanced-compression profile).
Transform and Quantization [Tues. – RM] Side activity conducted to investigate toward harmonization. [End of Tuesday Subject Areas Addressed]
Complexity Minimization
Kerofsky & Lei  HYPERLINK "VCEG-N20.doc" VCEG-N20 Decoder needs 32-bit arithmetic. Split multiplication of a big number into multiplication with a mantissa and shifting by an exponent. Normalization shows sufficient accuracy. Mantissa is merged with transform matrix resulting in 6 matrices. Luma DC transform using Hadamard. No differences in PSNR. Rather, minor improvements for chroma.
Zhou  HYPERLINK "VCEG-N22.doc" VCEG-N22 Divide forward transform by a factor to reduce precision required. Luma DC transform using Hadamard. No difference for high QP values for Intra and Inter. No tests for low QP values regarding coding efficiency. Comment: results at Austin meeting in VCEG-M72, common conditions in VCEG-M38
Topiwala  HYPERLINK "VCEG-N24.doc" VCEG-N24 Follow-up of the Austin contribution VCEG-M16. See Austin meeting report.
Hallapuro  HYPERLINK "VCEG-N43.doc" VCEG-N43 Factorization of forward transform. No degradation in performance. 3 quantization tables are used.
Malvar  HYPERLINK "VCEG-N44.doc" VCEG-N44 Construction of a transform matrix using multiplication of DCT matrix and rounding. 16 bit architecture. The chosen Matrix has different norms across the vectors requiring 3 quantization tables. Very small subjective differences, if any.
VCEG-N69 Procedure to assess trade-offs. Common conclusion of several proposals: 16 bit is possible without loss in performance Iterated Hadamard to be tested Create an AdHoc Group chaired by L. Kerofsky, Compare proposals with the method outlined in VCEG-N69 for various architectures, separate analysis of coder, decoder, transform, quantization with regards to R-D performance and complexity, analyze varying bit-depths, lossless and near-lossless coding, and coarser quantization in the transform domain
Adaptive Block Transforms  HYPERLINK "VCEG-N49.doc" VCEG-N49 Combination with CABAC is analyzed. Different binarization than CABAC. Mobile: slightly better on the right, more weird artifacts on the left, Mobile left looked better Paris: tie was better on the right but Gary’s neck looked weird, other parts better on the left Container: water looked better on the left (smoother) left: TML-8.0, right: ABT 16 bit implementation believed to be possible Based on shown results at this meeting and in the past, it was noted that this feature is unlikely to be adopted in baseline if performance doesn’t change significantly. Investigation of usage in a possible high-res profile was encouraged.
SP-Frames  HYPERLINK "VCEG-N42.doc" VCEG-N42 Information about applications for SP-frames including switching of pre-coded bitstreams at varying bit-rates and resolutions, attachment of Intra frames to pre-coded bitstreams, resynchronization in video redundancy coding, error recovery via resynchronization. Request for the insertion of a non-propagating dc=128 intra prediction.
Scalable Coding
SNR with Leaky Prediction  HYPERLINK "VCEG-N53.doc" VCEG-N53 An SNR scalable scheme is presented with a leaky prediction method. A comment was raised regarding the temporal behavior. The group indicated that the method is interesting and encourages further work with the target of an application with the associated profile.
Coding of Synchronized Streams  HYPERLINK "VCEG-N35.doc" VCEG-N35 Three loops run with the same motion vectors and their differences are transmitted. The group indicated that the method is interesting and encourages further work with the target of an application with the associated profile.
Motion Rendition  HYPERLINK "VCEG-N09.doc" VCEG-N09 (ad-hoc report)
Motion Coding with Global Motion  HYPERLINK "VCEG-N16.doc" VCEG-N16 Distinct from well-known global motion compensation, i.e., no warping of pictures. Transmit a global motion model and code skip and 16x16 relative to it. News: no visible quality difference Foreman: no visible quality difference Tempete: no visible quality difference Silent: no visible quality difference Coastgard: significantly better visual quality for the proposed method, even though a lower bit-rate was used for the proposed method Bit rate savings average approx 7.6% overall (20%), approx 19% of MV rate (up to more than 30% on coastguard sequence) when tested on Foreman, Mobile, Coastguard, and Tempete. Conduct core experiment for an application with the corresponding conditions. CE partner is RWTH Aachen. Document description VCEG-N77.
Multi-Hypothesis Motion  HYPERLINK "VCEG-N40.doc" VCEG-N40 Information on search techniques for B-frames in the generalized Annex U syntax. Employ multi-hypothesis mode as a generalized B-frame mode. Use of several future reference frames. Further work towards integration into the software is urged.
Long-Term Prediction Syntax  HYPERLINK "VCEG-N48.doc" VCEG-N48 Similar to Annex U, a frame buffering syntax is proposed. An AhG was formed to finally incorporate the Annex U syntax. Chair: T. Wiegand.
Low-Overhead Prediction HYPERLINK "VCEG-N45.doc"VCEG-N45 16x16 mode + 14 flexible MB segmentation patterns using 2 MVs. 16x8 and 8x16 are similar, others differ. Assignment of ref frame parameter to each mb part. Did not encode all test sequences. Issues was raised regarding results for higher definition material. Other material available from VQEG. Encourage further work. Several people indicated interest.
Interpolation Methods VCEG-N71, VCEG-N74
Analysis  HYPERLINK "VCEG-N55.doc" VCEG-N55 For the case of half-pel motion, bi-linear interpolation is sufficient. For quarter-pel, 6-tap filters are needed. All sequences tested are in QCIF resolution and 10 Hz. Not the complete test set. Provides potential for encoder complexity scalability – could allow encoder to select short filters and less accurate motion vectors to minimize encoder complexity
Quality-Improved Filtering  HYPERLINK "VCEG-N28.doc" VCEG-N28 Estimates filters. Uses 12 bit precision. Uses 30 Hz sequences. Word length of filter coefficients. Estimation process incurs delay but coefficients from previous frame could be used. The group finds this work very promising and encourages further investigations.
Complexity-Reduced Filtering  HYPERLINK "VCEG-N51.doc" VCEG-N51 Did only analyze motion compensation part. MC from original sequences. Results with TML needed. Further work encouraged.  HYPERLINK "VCEG-N19.doc" VCEG-N19 Analysis of existing code. Adopt filtering outside of the picture (L. Kerofsky).  HYPERLINK "VCEG-N31.doc" VCEG-N31 Efficient subsequent interpolation method. Claim to be faster than direct interpolation. Modification of interpolation of corner position of the ¼-pel grid. Proposal to exchange with existing method. AhG was generated, chair: Thomas Wedi, co-chair: Marta Karczewicz. Mandate: clarify issues and quickly resolve the problems with the tentatively adopted content of VCEG-N31 and investigate other aspects of motion compensation filtering.
I-Frame Compression Improvement
Via New Prediction Modes  HYPERLINK "VCEG-N54.doc" VCEG-N54 New prediction modes inserted. Also affects old modes: gains between 5-10%. Without affecting the old modes: gains are reduced. Group encourages further work. Software to be made available, IPR 2.2
Interlaced Video Coding
Handling interlaced video in H.26L  HYPERLINK "VCEG-N57.doc" VCEG-N57r2 Incorporating the ides of Annexes U and W as well as MPEG-2 for interlaced coding in H.26L. Idea: use fields as frames. 5 reference frames are used. Subjective comparison of H.26L with original showed impressive results. Preliminary results with interlaced show very good performance of H.26L for interlaced pictures. Syntax element for picture header is proposed to utilize interlaced scanned pictures.
VGEG-N76 Macroblock based field/frame mode. Create an AhG with the mandate to investigate interlaced coding within H.26L. Chair: P. Borgwardt. Adopt change of picture header following VCEG-N57r2.
Using H.263 Annex U techniques with Multiple-Picture Backward Prediction VCEG-N84 Document received after the closing of the meeting, made available for information purposes only for consideration in future work.
Non-Normative Encoding Methods HYPERLINK "VCEG-N10.doc"VCEG-N10 (test model document)
Fast Motion Estimation  HYPERLINK "VCEG-N25.doc" VCEG-N25 Asserted 4X-10X Speed-up, asserted no signif loss in video quality., not yet tested in H.26L context, but asserted should work properly (not proposing to remove the higher-complexity method, only to add this as an alternative). Adaptation to H.26L context is not straightforward.
Dynamic Search Range Decision  HYPERLINK "VCEG-N33.doc" VCEG-N33 Motion vectors highly correlated to neighbors. Sometimes large range generally helps and sometimes it doesn’t. Basic concept is to determine whether large search range is generally helping or not and to reduce the search range when it is not. Comment that this concept can be used to increase performance beyond that of “full search” in cases where the search range of “full search” turns out not to be adequate (unless the full search means a search of the full picture size). Work also assigned to AHG for analysis and merging as appropriate. Will provide software. Form AHG (chair: Pankaj Topiwala) on reduced-complexity motion estimation. Support of rapid adoption of VCEG-N25 techniques, provided no harmful impact on structure of code – final decision on reflector.
Loop Filtering  HYPERLINK "VCEG-N08.doc" VCEG-N08 (ad-hoc report) A particular solution was worked out in the ad-hoc activity and reported in the ad-hoc report. Significant complexity savings. No essential difference in visual quality. Already incorporated into TML 8.4 software. Content for TML document provided as annex to the report. Adopted.
Complexity Reduction HYPERLINK "VCEG-N17.doc"VCEG-N17 Use of skip mode and use of boundary strength parameter. Very similar design to that of the ad-hoc report VCEG-N08, with further complexity reduction. May hold potential but much of the reported improvement is from the ad-hoc output design and this is not directly compared to that.
TML 8.4 Loop Filter Analysis  HYPERLINK "VCEG-N29.doc" VCEG-N29 Loop filter with ad-hoc output design filter (N09) as loop and post filter. Analysis of complexity and quality of the result. Having the filter outside the loop requires memory copy and repetition of some filtering operations and requires and extra frame store in some implementations. Complexity analysis shows more that 2X complexity increase for the filtering when out of loop. Results reported with BD-PSNR (Bjøntegaard delta PSNR) method endorsed by the group show consistent improvement between 0.0 and 0.2 dB for luminance (averaging 0.1 dB) in favor of the loop method, with more improvement for chrominance. Several visual demonstrations shown – consensus that loop method has significantly higher visual quality.
Quality Enhancement  HYPERLINK "VCEG-N30.doc" VCEG-N30 TML design (prior to ad-hoc output N08) asserted to sometimes not detect/reduce blocking artifacts at higher bit rates. Alterations of both the blocking detection method and the subsequent filtering to address this problem without causing excessive blur. Visual quality shown for individual areas of individual frames of a few sequences. Group consensus that some visual improvement shown in these examples. Essentially no BD-PSNR impact (slight improvement). Promising contribution. Consensus to form AHG (chair: Peter List) to further investigate both quality and complexity issues in loop filtering.
Supplemental Enhancement Information  HYPERLINK "VCEG-N60.doc" VCEG-N60 Good idea. Adopted as presented. One byte SEI type (plus escape code) and one byte SEI data size (plus escape code), followed by SEI payload. Each SEI header with its associated payload to be handled as specified by NAL.
Entropy Coding
Ordered Coefficient Coding  HYPERLINK "VCEG-N34.doc" VCEG-N34 Coding gain up to 3% for intra coding (best for large quantizer sizes). Order transform coeffs by size, with ordering data to substitute for RUN data. Ordering and run information is combined… then CABAC processing. Does not seem sufficiently mature and effective for immediate adoption. Further work on intra coding performance in general is encouraged.
CABAC Improvement  HYPERLINK "VCEG-N41.doc" VCEG-N41 Proposed context model for levels based on prior level, proposed context model for runs. Up to 4.5% intra improvement (avg 2%) at low QP values. Adaptive scanning – adding scan orders – providing up to 5.5% improvement (avg. 3.5%) when used in conjunction with new context models. Adaptive scanning alone provides about 1.5% avg. Not proposed for baseline. Comment: question re quantity of data before reset of CABAC context models due to slice resync. Comment: interesting that it works at medium step sizes. Further work on intra coding performance in general is encouraged.
MB Prediction Coding for B Pictures  HYPERLINK "VCEG-N32.doc" VCEG-N32 Reducing prediction modes in B pictures for computational savings. Currently 17 types, plus the variations due to multiple intra 16x16 prediction types. 4x4, 4x8 and 8x4 forward and backward prediction modes not used much and their presence appears to have no influence on coding performance. Propose using only the block types 16x16, 16x8, 8x16, and 8x8 and not the other three. R-D optimization not used. Common conditions for coding efficiency, as written, do not specify using R-D optimization of mode decisions. Let’s change that. Agreed. Comment: When evaluating the utility of prediction mode methods, it seems fundamental to use R-D optimization to ensure most effective mode selection is the basis for that judgment of utility. Comment: Why not just move things to the end of the entropy table list, rather than deleting them? Comment: Suggest running a test, with use of R-D optimization, to determine the proper ordering of the table for B pictures. Comment: Description of direct skip mode for B pictures is missing from TML document and CBP list is wrong for B pictures (also should be loop in Fig A.2 back to just after PTYPE). SP picture description is also missing. Asked the editor to correct these problems. Significant restructuring of the TML document is authorized for clarity.
Modified UVLC Codewords  HYPERLINK "VCEG-N36.doc" VCEG-N36 UVLC design has interleaving of “prefix” bits (000…1) with “info payload” bits. Maximum prefix 16 bits with 15-bit payload bits. Proposal to not interleave these bits – instead simply send the prefix first. Comment: Current UVLC has self-aligning property and reverse-decoding property that would be removed by this. Currently we don’t really know how useful these properties of the current UVLC design are. Comment: Does the decoding process really cost more due to interleaving? (Yes, it appears that the other method would be less complex.) Isn’t there some other way to decode other than separating prefix from payload first? Currently we don’t really know for certain to what degree the complexity is impacted. Comment: Remember the symbol association problem pointed out by Kerofsky in the current design, and also remember the alternative structured VLC designs etc. by Kerofsky and others. Comment: VLD complexity matters because that part of the process is often done off-accelerator. Defer to future consideration.
Profiles & Applications  HYPERLINK "VCEG-N07.doc" VCEG-N07 (ad-hoc report) Comment: Baseline likely to contain everything currently in TML but 1/8-pel, CABAC, SP frames, B frames. Agreed. Comment: Let’s create an “important documents” directory/archive for ftp site. Agreed.
File Storage Format  HYPERLINK "VCEG-N65.doc" VCEG-N65 Look for requirements, work on design, … Comment: Is this to be part of the standard? (Proponent: Yes) Normative? (Proponent: Yes) Is this within our area of core competency, or should we contract this out? Perhaps we should define a packetization for use in some referenceable standard format or a method of easily altering some transport format to make it a file format (e.g. put an MPEG-2 transport stream into a file with some H.26L in it using some PID identifying H.26L, and name the file with a “.26L” extension). Defer to later consideration. Proposed Required: 1) Sufficient for interchange between H.26L-aware processing elements – Sounds useful. 2) Every H.26L encoding or decoding system capable of file input or output shall support the file format – Not agreed. 3) Provide means for identifying headers, etc. – Sounds useful. 4), … 5), Support multiple encodings of the same content, support SP transitions, random access, fast forward, fast backward, etc. 6), 7) … 12) Streamability Not required to be a substitute for highly-capable multimedia file formats – Agreed. We don’t want to take on such a task. Agree that having a file format is useful to us. Further study is encouraged. Decided to add a goal to the project goal list to define or find a file format with definition of how to encapsulate H.26L streams within that format for purposes such as simple stream interchange, http streaming service, random access, support of multiple streams with transitions (e.g. using SP pictures between streams), etc.
Non-Normative Decoding Methods
Post-processing – No contributions or actions. Work encouraged.
Error Concealment  HYPERLINK "VCEG-N62.doc" VCEG-N62 Closely related to N63. Provide basic error resilience, contribute software and test model content. Consider only lost slices. Spatial or motion concealment from neighbor MBs. Process by column to maintain more constant motion in center of picture and avoid spreading of error to the edges of the picture. Intra pictures: use weighted averaging from border samples based on distance from border. Inter pictures: If local motion activity small, use simple copy; otherwise choose one of neighboring MVs or zero based on luminance match with border sample values. Demonstration of effectiveness shown. More sophisticated method than used in other proposals. Make normative? See N61 notes.
Reference Software
Coding Style  HYPERLINK "VCEG-N46.doc" VCEG-N46 Propose common coding style for use of brackets, indentation, style of comment delineation, non-use of tab characters, prevention of multiple inclusion of header files, and removal of unused code. Rule 6: Mandating “old-style” C code comments is not adopted. All other rules are enthusiastically adopted as proposed. See below with regard to Doxygen documentation method. Require those contributing to software development to conform to these quidelines.
Documentation  HYPERLINK "VCEG-N47.doc" VCEG-N47 Use of Doxygen documenting style for all TML software development. Adopted.
Maintenance of H.120, H.261, H.263 – no contributions, no action taken.
Closing Review and Results
Presentation and Review of Results The following is a short description of the actions taken at the meeting:
Changes to TML draft standard document [VCEG-N83]
Correct description of picture sync for consistency
Change picture header content to support indication of interlaced picture types
Add syntax mechanism for carrying supplemental enhancement information [VCEG-N60]
Correct description of direct skip mode handling for B pictures, CBP list for B pictures, syntax diagram
Authorize significant restructuring of document as necessary for clarity
Specify that the draft “Baseline” profile contains all technical features except 1/8-pel motion, CABAC, SP frames, and B pictures
Alter network adaptation layer normative text [VCEG-N72]
Alter method for handling the edges of pictures for motion compensation to first apply picture area extension and then apply interpolation filter processing [VCEG-N19]
Alter method of performing motion compensation for sub-pixel position interpolation [VCEG-N31] (adopted conditioned on provision of software and no problems surfacing soon after the meeting)
Simplify loop filter design [VCEG-N08]
Add multiple leaky bucket VBV/HRD model [VCEG-N58r1]
Incorporate capabilities of H.263 Annex U for multiple reference picture buffer management (previous plan reemphasized)
Add SP picture description (previous plan reemphasized)
Changes to Experiment Common Conditions and Non-Normative Encoding and Decoding Techniques for Experiments
Add the use of rate-distortion optimization to common conditions for experiments in the area of coding efficiency capability
Change common conditions for testing performance for RTP/IP on 3GPP [VCEG-N37]
Tentatively adopted incorporation of multiple parallel decoding technique for error resilience testing capability [VCEG-N38, VCEG-N39, VCEG-N50] (adopted conditioned on checking and reporting on operation with new error concealment method and slice size based on number of bytes threshold)
Adopted new error concealment technique [VCEG-N62, VCEG-N63]
Tentatively adopted reduced-complexity motion estimation [VCEG-N25] (final decision on reflector)
Specify common conditions for testing for
Wire-line low-delay IP/UDP/RTP packet loss resilience [VCEG-N79]
Mobile IP/UDP/RTP on 3GPP/3GPP2 [VCEG-N80]
Coding efficiency [VCEG-N81]
Core Experiments Planned
Global Motion Vector Coding [VCEG-N77]
Reduced-Complexity Transform and Quantization [VCEG-N78]
Interlaced Video Coding [VCEG-N85]
Output Documents
Liaison Statements and Collaborative Letters To send a liaison statement to MPEG Reviewed draft of liaison statement. To add description of technical state of H.26L project. Add statement to LS that technical state of the project in VCEG when J-VCEG is formed should be the basis for the future work, in order to avoid the need to revisit and duplicate completed work and discussions. Add statement that JVCEG will have an email reflector for group communication. Add file format goal.
Other Output Documents IPR reporting form reviewed and agreed. Header design N72 reviewed See doc list VCEG-N00 for a list of all documents marked as output documents.
Plans for Future Work
General Plans Final plan in regard to MPEG joint work should be provided to the experts prior to start of joint work by the rapp with his recommendation. Final agreement authority rests with SG16 chairman.
Meeting Plans [Not the week of Dec 10, but sometime in that neighborhood, perhaps joint with MPEG]
Ad-Hoc Committees Established Transform and Quantization (Chair: Louis Kerofsky) Charter: To study reduced-complexity transform, inverse transform, quantization, and inverse quantization toward minimization of associated complexity for various architectures, esp. in regard to implementation on 16-bit and 32-bit processors. Should assess complexity using method outlined in VCEG-N69. Will include study of R-D performance, complexity, varying bit-depth input, lossless and near-lossless coding quality, as well as coarser quantization in the transform domain. Generalized Annex U Incorporation (Chair: Thomas Wiegand) Charter: To follow through on the intent to incorporate the capabilities of H.263 Annex U into the H.26L design along with further generalization to decouple the decoding process from the display process (to enable P picture coding with temporally-subsequent reference pictures and B pictures with similar agnostic treatment of temporal order, including in particular with respect to enabling multiple-picture “forward”, “backward”, and “bi-directional” prediction). Motion Interpolation (Chair: Thomas Wedi, Co-Chair: Marta Karczewicz) Charter: To clarify and quickly resolve the quality and complexity aspects of sub-pixel motion compensation interpolation methods, including consideration of the impact on both encoder and decoder; to verify the appropriateness of use for the techniques of VCEG-N31 to minimize complexity; to incorporate the VCEG-N31 techniques into the test model if the group’s tentative adoption is verified to be appropriate (software availability announcement and final decision on reflector); and to investigate other aspects of motion compensation filtering. Motion Estimation (Chair: Pankaj Topiwala) Charter: To further the work on reduced-complexity motion estimation techniques for H.26L, including rapid adoption of VCEG-N25 method provided no harmful impact results for the structure of the source code (software availability announcement and final decision on reflector), potentially in combination with other fast motion estimation techniques such as found for example in VCEG-N33. Interlaced Video Coding (Chair: Peter Borgwardt) Charter: To further the work on support of interlaced video coding for H.26L, including investigation of field coding with multiple reference fields as supported by H.263 Annexes U and W and other such techniques as may be useful for the representation of interlaced-scan video, including investigation of macroblock-level frame/field adaptivity – while retaining maximal design simplicity and minimal encoder/decoder computational complexity. Also to define common conditions and test sequences for experiments and to design and conduct core experiments for interlaced-scan video coding. Loop Filter (Chair: Peter List) Charter: To investigate quality and complexity issues for loop filter design in H.26L and to assess the potential for improved visual quality, reduced decoder computational complexity, and enhanced design simplicity. Test Model and Software (Chair: Karsten Sühring) Charter: To further the work on the documentation and software implementation of the H.26L test model design, including incorporation of modifications as approved by the group, and to rapidly provide improved software for group use in future experiments and for eventual approval as standardized reference software. H.26L Development (Chair: Gary Sullivan) Charter: To further the work on the H.26L project as a whole, including project planning, work coordination, and status review. Complexity Minimization Efforts (Chair: Michael Horowitz) Charter: To study the implementation complexity of the H.26L design and to recommend methods of minimizing that complexity in terms of encoder and decoder computational and implementation complexity and design simplicity. Profiles and applications (Chair: Thomas Wiegand) Charter: To study the applications of H.26L and the appropriate methods of addressing these applications with profiles and levels of the H.26L design, including particular emphasis on the design of a baseline profile. NOTE: Members working in the area covered by an ad-hoc group are strongly requested to inform the AHG chair so that work can be conducted in a maximally productive and coordinated fashion. Particular care should be taken to coordinate efforts for activities that impact the work of multiple ad-hoc groups.
All Other Business
Closing of the Meeting
The group thanked Microsoft, the meeting host, and Preferred Meetings, the meeting organizing agency
The group thanked those providing equipment and help, in particular the University of California, Santa Barbara and Prof. Allen Gersho, and the Santa Barbara Technology Group and Drew Hawn for their help in providing a computer server and hub for use at the meeting and Prof. Stephan Wenger for his usual aid in providing networking equipment and help to get the network operating.
Closing of the meeting at 7:45 pm on Thursday, Sept. 27, 2001.

Annex A: List of Documents
ITU-T Q.6/16 Video Coding Experts Group (VCEG) Document VCEG-N00 rev. 3 (filename VCEG-N00r3.doc)
14th meeting, Santa Barbara, California, USA, September 24-27, 2001 last date revised: 27 Sept., 2001
LIST OF DOCUMENTS FOR THE FOURTEENTH MEETING
ITU-T Q.6/SG16
VIDEO CODING EXPERTS GROUP (VCEG)



DOCUMENT
NO. 
SOURCE



COUNTRY EXPERT/
ORGANIZATIONP
U
R
P
O
S
E
TITLEA
D
M
I
N
H.
2
6
L
H.
2
6
3P
R
I
O
R
O
T
H
E
RMeeting Invitation and Meeting PlanHYPERLINK "VCEG-NT0r1.doc"VCEG-N-TD-0 (r1)HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurIInvitation to the Fourteenth Meeting of the ITU-T Video Coding Experts Group (VCEG)XHYPERLINK "VCEG-NT1r4.doc"VCEG-N-TD-1 (r4)HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurIPlan of the Fourteenth Meeting of the ITU-T Video Coding Experts Group (VCEG)XHYPERLINK "VCEG-N00r2.doc"VCEG-N00 (r2)HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurIList of Documents of the Fourteenth Meeting of the ITU-T Video Coding Experts Group (VCEG)XReport of Previous MeetingHYPERLINK "VCEG-N01.doc"VCEG-N01HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurIMeeting report of the Video Work at the ITU-T Study Group 16 meeting in Porto Seguro, BrazilXList of Meeting Participants and Experts GroupHYPERLINK "VCEG-N02.doc"VCEG-N02*HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurOList of Participants of the 14th Meeting (Meeting N) of the ITU-T Q.6/SG16 Video Coding Experts GroupXHYPERLINK "VCEG-N03.doc"VCEG-N03*HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurOList of Experts of the ITU-T Q.6/SG16 Video Coding Experts Group (VCEG)XAd Hoc ReportsHYPERLINK "VCEG-N04.doc"VCEG-N04 HYPERLINK "mailto:stewe@cs.tu-berlin.de" Stephan Wenger, AHC ChairpersonIReport of the Ad Hoc Committee on Test Model Enhancement and Software DevelopmentXXHYPERLINK "VCEG-N05.doc"VCEG-N05HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, AHC ChairpersonIReport of the Ad Hoc Committee on H.26L DevelopmentXXHYPERLINK "VCEG-N06.doc"VCEG-N06 HYPERLINK "mailto:mhorowitz@austin.polycom.com" Michael J. Horowitz, AHC Chairperson; with  HYPERLINK "mailto:tij@tandberg.no" Tom-Ivar JohansenIReport of the Ad Hoc Committee on H.26L Complexity MinimizationXXHYPERLINK "VCEG-N07r1.doc"VCEG-N07 (r1) HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand, AHC ChairpersonIReport of the Ad Hoc Committee on H.26L Profiles and ApplicationsXXHYPERLINK "VCEG-N08r1.doc"VCEG-N08 (r1) HYPERLINK "mailto:Peter.List@t-systems.de" Peter List, AHC ChairpersonIReport of the Ad Hoc Committee on H.26L Loop Filter Complexity ReductionXXHYPERLINK "VCEG-N09.doc"VCEG-N09HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, AHC ChairpersonIReport of the Ad Hoc Committee on H.26L Motion RenditionXXMeeting ContributionsHYPERLINK "VCEG-N10.doc"VCEG-N10 HYPERLINK "mailto:Gisle.Bjontegaard@international.telenor.no" Gisle Bjøntegaard, EditorIH.26L TML 8 Reflecting Decisions of Porto SeguroXHYPERLINK "VCEG-N11.doc"VCEG-N11 (WG11 M7511)HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, HYPERLINK "http://www.itu.int/ITU-T/com16/rapporteurs.html"Q6/16 RapporteurISummary Information for ITU-T VCEG Draft H.26L Algorithm Provided in Response to MPEG Video and DCinema CfPsXXHYPERLINK "VCEG-N12.doc"VCEG-N12 (WG11 N4307) HYPERLINK "http://drogo.cselt.stet.it/mpeg/" ISO/IEC JTC1 SC29 WG11 (MPEG)PLiaison Statement from MPEG to ITU-T SG 16 regarding the Joint Video TeamXXHYPERLINK "VCEG-N13.doc"VCEG-N13 (WG11 N4240) HYPERLINK "http://drogo.cselt.stet.it/mpeg/" ISO/IEC JTC1 SC29 WG11 (MPEG)IPreliminary Results of MPEG Subjective Assessment of Responses to Video Call for New Tools to Further Improve Coding EfficiencyXXHYPERLINK "VCEG-N14.doc"VCEG-N14 (WG11 N4341) HYPERLINK "http://drogo.cselt.stet.it/mpeg/" ISO/IEC JTC1 SC29 WG11 (MPEG)PProposed Work Plan for the Joint Video Project between ITU-T SG16 and ISO/IEC JTC 1/SC 29/WG 11XXHYPERLINK "VCEG-N15.doc"VCEG-N15 (WG11 N4298) HYPERLINK "http://drogo.cselt.stet.it/mpeg/" ISO/IEC JTC1 SC29 WG11 (MPEG)ILiaison Statement from MPEG to ITU-T SG 16 regarding the MPEG Call for Proposals Regarding Image Coding Technology Suitable for Digital Cinema ApplicationsXXHYPERLINK "VCEG-N16.doc"VCEG-N16 (and  HYPERLINK "VCEG-N16.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:ssun@sharplabs.com" Shijun Sun,  HYPERLINK "mailto:shawmin@sharplabs.com" Shawmin Lei,  HYPERLINK "http://www.sharplabs.com/" Sharp Labs of AmericaPMotion Vector Coding with Global Motion ParametersXHYPERLINK "VCEG-N17.doc"VCEG-N17 (and  HYPERLINK "VCEG-N17.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:ssun@sharplabs.com" Shijun Sun,  HYPERLINK "mailto:shawmin@sharplabs.com" Shawmin Lei,  HYPERLINK "http://www.sharplabs.com/" Sharp Labs of AmericaPImproved TML Loop Filter with Lower ComplexityXHYPERLINK "VCEG-N18.doc"VCEG-N18 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA &  HYPERLINK "http://www.lonelyplanet.com/destinations/north_america/canada/" Canada HYPERLINK "mailto:pnt@fastvdo.com" Pankaj Topiwala,  HYPERLINK "http://www.fastvdo.com/" FastVDO; HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan,  HYPERLINK "http://www.microsoft.com" Microsoft Corp.;  HYPERLINK "mailto:anthonyj@ece.ubc.ca" Anthony Joch,  HYPERLINK "mailto:faouzi@ece.ubc.ca" Faouzi Kossentini,  HYPERLINK "http://www.ubc.ca" Univ. British ColumbiaIH.26L TML 8 Performance Evaluation vs. H.263++ High Latency Profile & MPEG-4 Advanced Simple ProfileXHYPERLINK "VCEG-N19r1.doc"VCEG-N19 (r1) (and  HYPERLINK "VCEG-N19.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:tchu@sharplabs.com" Tianli Chu,  HYPERLINK "mailto:ssun@sharplabs.com" Shijun Sun,  HYPERLINK "mailto:lkerofsky@sharplabs.com" Louis Kerofsky,  HYPERLINK "mailto:shawmin@sharplabs.com" Shawmin Lei,  HYPERLINK "http://www.sharplabs.com/" Sharp Labs of AmericaPEfficient 16-bit Implementation for Motion Compensation InterpolationXHYPERLINK "VCEG-N20.doc"VCEG-N20 (and  HYPERLINK "VCEG-N20.ppt" presentation and  HYPERLINK "VCEG-N20.zip" zipped data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:lkerofsky@sharplabs.com" Louis Kerofsky,  HYPERLINK "mailto:shawmin@sharplabs.com" Shawmin Lei,  HYPERLINK "http://www.sharplabs.com/" Sharp Labs of AmericaPReduced Bit-depth QuantizationXHYPERLINK "VCEG-N21r1.doc"VCEG-N21 (r1) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:lkerofsky@sharplabs.com" Louis Kerofsky,  HYPERLINK "http://www.sharplabs.com/" Sharp Labs of AmericaPEnhanced Picture SyncXHYPERLINK "VCEG-N22.doc"VCEG-N22 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:zhou@ti.com" Minhua Zhou,  HYPERLINK "http://www.ti.com" Texas InstrumentsP16-bit Integer Transform and QuantizationXHYPERLINK "VCEG-N23.doc"VCEG-N23 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:zhou@ti.com" Minhua Zhou,  HYPERLINK "http://www.ti.com" Texas InstrumentsIBenchmark Analysis of H.26L Decoding Functional BlocksXHYPERLINK "VCEG-N24r1.doc"VCEG-N24 (r1) (and  HYPERLINK "VCEG-N24.ppt" presentation and  HYPERLINK "VCEG-N24.xls" data and  HYPERLINK "VCEG-N24-qtable.zip" zipped tables) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:jliang@fastvdo.com" Jie Liang,  HYPERLINK "mailto:tran@fastvdo.com" Trac Tran,  HYPERLINK "mailto:weidai@fastvdo.com" Wei Dai,  HYPERLINK "mailto:pnt@fastvdo.com" Pankaj Topiwala,  HYPERLINK "http://www.fastvdo.com/" FastVDOPFast Integer Transform and 16-Bit ArchitectureXHYPERLINK "VCEG-N25.doc"VCEG-N25 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:cjtu@fastvdo.com" Chengjie Tu,  HYPERLINK "mailto:pnt@fastvdo.com" Pankaj Topiwala,  HYPERLINK "http://www.fastvdo.com/" FastVDOPA hybrid feature/image block motion-vector estimation approach (HyFIVE)XVCEG-N26–––– Document Registration Withdrawn –VCEG-N27–––– Document Registration Withdrawn –HYPERLINK "VCEG-N28.doc"VCEG-N28 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:wedi@tnt.uni-hannover.de" Thomas Wedi,  HYPERLINK "http://www.uni-hannover.de" Univ. HannoverPAdaptive Motion Compensation Interpolation FiltersXHYPERLINK "VCEG-N29.doc"VCEG-N29 (and  HYPERLINK "VCEG-N29.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:jani.lainema@nokia.com" Jani Lainema,  HYPERLINK "mailto:Marta.Karczewicz@nokia.com" Marta Karczewicz,  HYPERLINK "http://www.nokia.com/main.html" NokiaITML 8.4 Loop Filter AnalysisXHYPERLINK "VCEG-N30r1.doc"VCEG-N30 (r1) (and HYPERLINK "VCEG-N30r1.xls"data (r1)) HYPERLINK "http://www.lonelyplanet.com/destinations/north_east_asia/south_korea/" Korea HYPERLINK "mailto:mhong@e.ssu.ac.kr" Min-Cheol Hong,  HYPERLINK "http://www.soongsil.ac.kr/english/" Soongsil University;  HYPERLINK "mailto:ympark@lge.co.kr" Young Man Park,  HYPERLINK "http://www.lge.com/" LG ElectronicsPModified Loop Filter for Improving Visual QualityXHYPERLINK "VCEG-N31r1.doc"VCEG-N31 (r1) (and  HYPERLINK "VCEG-N31.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:Marta.Karczewicz@nokia.com" Marta Karczewicz,  HYPERLINK "mailto:Antti.Hallapuro@nokia.com" Antti Hallapuro,  HYPERLINK "http://www.nokia.com/main.html" NokiaPInterpolation Solution with Low Encoder Memory Requirements and Low Decoder ComplexityXHYPERLINK "VCEG-N32.doc"VCEG-N32 (and  HYPERLINK "VCEG-N32.xls" data) HYPERLINK "http://www.lonelyplanet.com/destinations/north_east_asia/south_korea/" Korea HYPERLINK "mailto:jeonbm@lge.com" Byeong-Moon Jeon,  HYPERLINK "http://www.lge.com/" LG ElectronicsPReduced MB Prediction Types for Computational Saving in B PicturesXHYPERLINK "VCEG-N33r1.doc"VCEG-N33 (r1) (and HYPERLINK "VCEG-N33r1.xls"data (r1)) HYPERLINK "http://www.lonelyplanet.com/destinations/north_east_asia/south_korea/" Korea HYPERLINK "mailto:mhong@e.ssu.ac.kr" Min-Cheol Hong,  HYPERLINK "http://www.soongsil.ac.kr/english/" Soongsil University;  HYPERLINK "mailto:ympark@lge.co.kr" Young Man Park,  HYPERLINK "http://www.lge.com/" LG ElectronicsPDynamic Search Range Decision for Motion EstimationXHYPERLINK "VCEG-N34.doc"VCEG-N34 (and  HYPERLINK "VCEG-N34.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:Gero.X.Baese@mchp.siemens.de" Gero Bäse,  HYPERLINK "mailto:Norbert.Oertel@mchp.siemens.de" Norbert Oertel,  HYPERLINK "mailto:Juergen.Pandel@mchp.siemens.de" Jürgen Pandel, HYPERLINK "http://www.siemens.com/en2/flash/"Siemens, AGPOrdered Coefficient CodingXHYPERLINK "VCEG-N35.doc"VCEG-N35 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:Peter.Amon@mchp.siemens.de" Peter Amon,  HYPERLINK "mailto:Gero.X.Baese@mchp.siemens.de" Gero Bäse,  HYPERLINK "mailto:klaus.illgner@mchp.siemens.de" Klaus Illgner,  HYPERLINK "mailto:Juergen.Pandel@mchp.siemens.de" Jürgen Pandel, HYPERLINK "http://www.siemens.com/en2/flash/"Siemens, AGPEfficient Coding of Synchronized H.26L StreamsXHYPERLINK "VCEG-N36.doc"VCEG-N36 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:zhou@ti.com" Minhua Zhou,  HYPERLINK "http://www.ti.com" Texas InstrumentsPModified Universal Length CodingXHYPERLINK "VCEG-N37.doc"VCEG-N37 (and HYPERLINK "VCEG-N37_software.tgz"tar+gzip software) HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany,  HYPERLINK "http://www.lonelyplanet.com/destinations/europe/sweden/" Sweden,  HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:Viktor.Varsa@nokia.com" Viktor Varsa,  HYPERLINK "mailto:Marta.Karczewicz@nokia.com" Marta Karczewicz,  HYPERLINK "http://www.nokia.com/main.html" Nokia;  HYPERLINK "mailto:goran.roth@era-t.ericsson.se" Göran Roth,  HYPERLINK "mailto:Rickard.Sjoberg@era.ericsson.se" Rickard Sjöberg,  HYPERLINK "http://www.ericsson.com/" Ericsson;  HYPERLINK "mailto:liebl@ei.tum.de" Günther Liebl,  HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer,  HYPERLINK "http://www.tu-muenchen.de/index_e.html" Tech. Univ. Munich;PCommon Test Conditions for RTP/IP over 3GPP/3GPP2 – Software and AmendmentsXHYPERLINK "VCEG-N38.doc"VCEG-N38 (and  HYPERLINK "VCEG-N38.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany &  HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer,  HYPERLINK "mailto:liebl@ei.tum.de" Günther Liebl,  HYPERLINK "mailto:drehvial@gmx.net" Tobias Oelbaum,  HYPERLINK "http://www.tu-muenchen.de/index_e.html" Tech. Univ. Munich;  HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand,  HYPERLINK "mailto:marpe@hhi.de" Detlev Marpe,  HYPERLINK "http://www.hhi.de/" Heinrich Hertz Inst.ISimulation Results for Common Test Conditions for RTP/IP over 3GPP/3GPP2XHYPERLINK "VCEG-N39.doc"VCEG-N39 (and  HYPERLINK "VCEG-N39.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer,  HYPERLINK "mailto:drehvial@gmx.net" Tobias Oelbaum,  HYPERLINK "http://www.tu-muenchen.de/index_e.html" Tech. Univ. Munich;  HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand,  HYPERLINK "mailto:marpe@hhi.de" Detlev Marpe,  HYPERLINK "http://www.hhi.de/" Heinrich Hertz Inst.ISimulation Results for Common Conditions for Video Performance Evaluation in H.324/M Error-Prone SystemsXHYPERLINK "VCEG-N40.doc"VCEG-N40 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:mflierl@stanford.edu" Markus Flierl, HYPERLINK "mailto:girod@ee.stanford.edu"Bernd Girod,  HYPERLINK "http://www.lnt.de/" Univ. Erlangen-NurembergPMultihypothesis Prediction for B framesXHYPERLINK "VCEG-N41r1.doc"VCEG-N41 (r1) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USAHYPERLINK "mailto:Ragip.Kurceren@nokia.com"Ragip Kurcerin,  HYPERLINK "mailto:Marta.Karczewicz@nokia.com" Marta Karczewicz,  HYPERLINK "http://www.nokia.com/main.html" NokiaPImprovements on CABACXHYPERLINK "VCEG-N42.doc"VCEG-N42 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USAHYPERLINK "mailto:Ragip.Kurceren@nokia.com"Ragip Kurcerin,  HYPERLINK "mailto:Marta.Karczewicz@nokia.com" Marta Karczewicz,  HYPERLINK "http://www.nokia.com/main.html" NokiaISP-frame demonstrationsXHYPERLINK "VCEG-N43.doc"VCEG-N43 (and  HYPERLINK "VCEG-N43.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:Antti.Hallapuro@nokia.com" Antti Hallapuro,  HYPERLINK "mailto:Marta.Karczewicz@nokia.com" Marta Karczewicz,  HYPERLINK "http://www.nokia.com/main.html" NokiaPLow complexity (I)DCTXHYPERLINK "VCEG-N44r1.doc"VCEG-N44 (r1) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USAHYPERLINK "mailto:malvar@microsoft.com"Henrique S. Malvar,  HYPERLINK "http://www.microsoft.com" Microsoft Corp.PLow-Complexity Length-4 Transform and Quantization with 16-Bit ArithmeticXHYPERLINK "VCEG-N45r1.doc"VCEG-N45 (r1) (and HYPERLINK "VCEG-N45.xls"data (r1)) HYPERLINK "http://www.lonelyplanet.com/destinations/north_east_asia/japan/" Japan HYPERLINK "mailto:shun@mml.yrp.nttdocomo.co.jp" Shun-ichi Sekiguchi,  HYPERLINK "mailto:kato@mml.yrp.nttdocomo.co.jp" Sadaatsu Kato,  HYPERLINK "mailto:adachi@spg.yrp.nttdocomo.co.jp" Satoru Adachi,  HYPERLINK "mailto:etoh@mml.yrp.nttdocomo.co.jp" Minoru Etoh,  HYPERLINK "http://www.nttdocomo.com/company/index.html" NTT DoCoMo Inc.PLow-overhead INTER Prediction ModesX HYPERLINK "VCEG-N46.doc" VCEG-N46 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:suehring@hhi.de" Karsten Sühring,  HYPERLINK "mailto:marpe@hhi.de" Detlev Marpe,  HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand,  HYPERLINK "http://www.hhi.de/" Heinrich Hertz Inst.PUnified Coding Style for H.26L Reference SoftwareX HYPERLINK "VCEG-N47.doc" VCEG-N47 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:suehring@hhi.de" Karsten Sühring,  HYPERLINK "mailto:marpe@hhi.de" Detlev Marpe,  HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand,  HYPERLINK "http://www.hhi.de/" Heinrich Hertz Inst.PProposed Documentation System for Reference SoftwareX HYPERLINK "VCEG-N48.doc" VCEG-N48 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/france/" France HYPERLINK "mailto:yann.lemaguet@philips.com" Yann Le Maguet,  HYPERLINK "mailto:cecile.dufour@philips.com" Cécile Dufour,  HYPERLINK "http://www.research.philips.com" Philips Research FrancePLong Term Prediction SyntaxXHYPERLINK "VCEG-N49r1.doc"VCEG-N49 (r1) (and  HYPERLINK "VCEG-N49.ppt" presentation) HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:wien@ient.rwth-aachen.de" Mathias Wien,  HYPERLINK "http://www.rwth-aachen.de/zentral/aguid_english_default.htm" RWTH AachenPNew ABT Results using CABACX HYPERLINK "VCEG-N50.doc" VCEG-N50 (and  HYPERLINK "VCEG-N50.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer,  HYPERLINK "mailto:drehvial@gmx.net" Tobias Oelbaum,  HYPERLINK "http://www.tu-muenchen.de/index_e.html" Tech. Univ. Munich;  HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand,  HYPERLINK "mailto:marpe@hhi.de" Detlev Marpe,  HYPERLINK "http://www.hhi.de/" Heinrich Hertz Inst.ISimulation Results for Common Conditions for H.323/Internet CaseX HYPERLINK "VCEG-N51.doc" VCEG-N51 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:szhong@broadcom.com" Sheng Zhong,  HYPERLINK "mailto:jalvarez@broadcom.com" José Roberto Alvarez,  HYPERLINK "mailto:macinnis@broadcom.com" Alexander (Sandy) MacInnis,  HYPERLINK "http://www.broadcom.com" Broadcom CorporationPImproved Sub-Pel Interpolation FiltersX HYPERLINK "VCEG-N52.doc" VCEG-N52 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:stewe@cs.tu-berlin.de" Stephan Wenger,  HYPERLINK "http://www.tu-berlin.de/" TU Berlin /  HYPERLINK "http://www.teles.de/holding/eng/index.html" Teles AG;  HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer,  HYPERLINK "http://www.tu-muenchen.de/index_e.html" Tech. Univ. MunichPH.26L over IP and H.324 FrameworkX HYPERLINK "VCEG-N53.doc" VCEG-N53 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:sehan@stanford.edu" Sang-Eun Han,  HYPERLINK "mailto:bgirod@stanford.edu" Bernd Girod,  HYPERLINK "http://www.stanford.edu/~sehan/" Stanford UniversityPScalable Coding with Leaky PredictionX HYPERLINK "VCEG-N54.doc" VCEG-N54 (and  HYPERLINK "VCEG-N54.xls" data) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:gregc@real.com" Greg Conklin,  HYPERLINK "http://www.real.com" Real NetworksPNew Intra Prediction ModesX HYPERLINK "VCEG-N55.doc" VCEG-N55 HYPERLINK "http://www.lonelyplanet.com/destinations/north_east_asia/japan/" Japan HYPERLINK "mailto:miyamoto@dsp.cl.nec.co.jp" Yoshihiro Miyamoto,  HYPERLINK "mailto:chono@ct.jp.nec.com" Kei-ichi Chono,  HYPERLINK "http://www.nec.com/" NEC Corp.IAnalysis of Fractional Pixel Interpolation for MC PredictionX HYPERLINK "VCEG-N56.doc" VCEG-N56 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/sweden/" Sweden HYPERLINK "mailto:Per.Frojdh@era.ericsson.se" Per Fröjdh,  HYPERLINK "mailto:rickard.sjoberg@era.ericsson.se" Rickard Sjöberg,  HYPERLINK "http://www.ericsson.com/" EricssonIConstrained Intra Prediction CodingXHYPERLINK "VCEG-N57r2.doc"VCEG-N57 (r2) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:peter.borgwardt@videotele.com" Peter Borgwardt,  HYPERLINK "http://www.videotele.com/" VideoTele.comPHandling Interlaced Video in H.26LXHYPERLINK "VCEG-N58r1.doc"VCEG-N58 (r1) HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USAHYPERLINK "mailto:jordir@microsoft.com"Jordi Ribas-Corbera,  HYPERLINK "mailto:pachou@microsoft.com" Philip A. Chou,  HYPERLINK "http://www.microsoft.com" Microsoft Corp.PA Generalized Hypothetical Reference Decoder for H.26LX HYPERLINK "VCEG-N59.doc" VCEG-N59 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/finland/" Finland HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "http://www.nokia.com" Nokia Mobile PhonesPPicture and Slice Header StructureX HYPERLINK "VCEG-N60.doc" VCEG-N60 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/finland/" Finland HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "http://www.nokia.com" Nokia Mobile PhonesPSyntax for Supplemental Enhancement InformationX HYPERLINK "VCEG-N61.doc" VCEG-N61 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/finland/" Finland HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "http://www.nokia.com" Nokia Mobile PhonesPHandling of Erroneous Bit-StreamsX HYPERLINK "VCEG-N62.doc" VCEG-N62 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:viktor.varsa@nokia.com" Viktor Varsa,  HYPERLINK "http://www.nokia.com" Nokia Research Center;  HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "mailto:ari.hourunranta@nokia.com" Ari Hourunranta,  HYPERLINK "http://www.nokia.com" Nokia Mobile Phones;  HYPERLINK "mailto:wangy@cs.tut.fi" Ye-Kui Wang,  HYPERLINK "http://www.tut.fi/public/" Tampere Univ. Tech.PNon-Normative Error Concealment AlgorithmsX HYPERLINK "VCEG-N63.doc" VCEG-N63 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/finland/" Finland HYPERLINK "mailto:wangy@cs.tut.fi" Ye-Kui Wang,  HYPERLINK "mailto:moncef.gabbouj@tut.fi" Moncef Gabbouj,  HYPERLINK "http://www.tut.fi/public/" Tampere Univ. Tech.;  HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "http://www.nokia.com" Nokia Mobile Phones;IInternet Simulations with Error ConcealmentXVCEG-N64–––– Document Registration Withdrawn – HYPERLINK "VCEG-N65.doc" VCEG-N65 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/finland/" Finland HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "http://www.nokia.com" Nokia Mobile PhonesPRequirements for H.26L File FormatX HYPERLINK "VCEG-N66.doc" VCEG-N66 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:Istvan.Sebestyen@icn.siemens.de" Istvan Sebestyen, HYPERLINK "http://www.siemens.com/en2/flash/"Siemens, AGPVCEG Patent/IPR Information Collection PoliciesXX HYPERLINK "VCEG-N67.doc" VCEG-N67 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/norway/" Norway HYPERLINK "mailto:halbach@tele.ntnu.no" Till Halbach,  HYPERLINK "http://www.ntnu.no/indexe.html" Nor. Univ. Sci. & Tech.ITML 8.4 Error Resilience PerformanceXLate-Registered Contributions and Output Contributions HYPERLINK "VCEG-N68.doc" VCEG-N68 HYPERLINK "http://www.lonelyplanet.com/destinations/europe/finland/" Finland HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. Hannuksela,  HYPERLINK "http://www.nokia.com" Nokia Mobile PhonesIComments Expressed Regarding Hypothetical Reference DecoderX HYPERLINK "VCEG-N69.doc" VCEG-N69 HYPERLINK "mailto:lkerofsky@sharplabs.com" Louis Kerofsky, Activity CoordinatorORequirements for Reduced Complexity Inverse Transform and DequantizationX HYPERLINK "VCEG-N70.doc" VCEG-N70* HYPERLINK "mailto:lindbergh@mediaone.net" David Lindbergh, Activity CoordinatorOIPR Reporting FormX HYPERLINK "VCEG-N71.doc" VCEG-N71 HYPERLINK "mailto:jalvarez@broadcom.com" José Roberto Alvarez, Activity CoordinatorODiscrepancies in Documentation and Implementation of Sub-pel Interpolation in TML-8 (Draft 0)X HYPERLINK "VCEG-N72.doc" VCEG-N72 HYPERLINK "mailto:stewe@cs.tu-berlin.de" Stephan Wenger, Activity Coordinator; with  HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer and  HYPERLINK "mailto:miska.hannuksela@nokia.com" Miska M. HannukselaOH.26L over IP Framework: Proposed H.26L NAL Normative TextX HYPERLINK "VCEG-N73.doc" VCEG-N73* HYPERLINK "mailto:stewe@cs.tu-berlin.de" Stephan Wenger, Activity CoordinatorOSkeleton Draft of RTP Payload Packetization for H.26LX HYPERLINK "VCEG-N74.txt" VCEG-N74 HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:wedi@tnt.uni-hannover.de" Thomas Wedi,  HYPERLINK "http://www.uni-hannover.de" Univ. HannoverIComments on Direct vs. Subsequent Motion InterpolatioonX HYPERLINK "VCEG-N75.doc" VCEG-N75* HYPERLINK "http://www.lonelyplanet.com/dest/eur/ger.htm" Germany HYPERLINK "mailto:Istvan.Sebestyen@icn.siemens.de" Istvan Sebestyen, HYPERLINK "http://www.siemens.com/en2/flash/"Siemens, AGPComments on ISO+ITU-T Joint Video Team FormationX HYPERLINK "VCEG-N76.doc" VCEG-N76 HYPERLINK "http://www.lonelyplanet.com/dest/nam/usa.htm" USA HYPERLINK "mailto:fogg_ce@yahoo.com" Chad Fogg,  HYPERLINK "http://www.pixonics.com/home.shtml" Pixonics; et alPAdaptive Field/Frame Block Coding Experiment ProposalX HYPERLINK "VCEG-N77.doc" VCEG-N77* HYPERLINK "mailto:ssun@sharplabs.com" Shijun Sun, Activity CoordinatorOCore Experiment on Global Motion Vector CodingX HYPERLINK "VCEG-N78.doc" VCEG-N78 HYPERLINK "mailto:lkerofsky@sharplabs.com" Louis Kerofsky, Activity CoordinatorOCore Experiment on Reduced-Complexity Transform and QuantizationXHYPERLINK "VCEG-N79r1.doc"VCEG-N79 (r1) HYPERLINK "mailto:stewe@cs.tu-berlin.de" Stephan Wenger, Activity CoordinatorOCommon Conditions for Wire-Line Low-Delay IP/UDP/RTP Packet Loss Resilience TestingX HYPERLINK "VCEG-N80.doc" VCEG-N80* HYPERLINK "mailto:stockhammer@ei.tum.de" Thomas Stockhammer, Activity CoordinatorOCommon Conditions for Mobile IP/UDP/RTP on 3GPP/3GPP2X HYPERLINK "VCEG-N81.doc" VCEG-N81*HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, Activity CoordinatorOCommon Conditions for Coding Efficiency TestsXHYPERLINK "VCEG-N82.doc"VCEG-N82*HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, RapporteurOLiaison Statement to MPEG on Joint Video WorkX HYPERLINK "VCEG-N83.doc" VCEG-N83* HYPERLINK "mailto:Gisle.Bjontegaard@international.telenor.no" Gisle Bjøntegaard and  HYPERLINK "mailto:wiegand@hhi.de" Thomas Wiegand, EditorsOH.26L Test Model Number 9X HYPERLINK "VCEG-N84.doc" VCEG-N84 HYPERLINK "http://www.lonelyplanet.com/destinations/north_america/canada/" Canada HYPERLINK "mailto:mgallant@videolocus.com" Michael Gallant,  HYPERLINK "mailto:gcote@videolocus.com" Guy Côté, HYPERLINK "http://www.videolocus.com/"VideoLocus, Inc.PHigh Rate, High Resolution Video Using H26LX HYPERLINK "VCEG-N85.doc" VCEG-N85* HYPERLINK "mailto:peter.borgwardt@videotele.com" Peter Borgwardt, Activity CoordinatorOCore Experiment on Interlaced Video CodingX HYPERLINK "VCEG-N86.doc" VCEG-N86HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, Activity CoordinatorOQ.6/16 Goals for the H.26L Project and Proposed Joint Standardization ScheduleXX HYPERLINK "VCEG-N87.doc" VCEG-N87HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, Activity CoordinatorOTerms of Reference Proposed for Joint Video ProjectXXMeeting Report DocumentHYPERLINK "VCEG-N88d0.doc"VCEG-N88*HYPERLINK "mailto:Gary.Sullivan@itu.int"Gary Sullivan, Q6/SG16 RapporteurRMeeting report of the fourteenth meeting (meeting N) of the ITU-T Q.6/16 Video Coding Experts Group (VCEG) – Santa Barbara, California, 24-27 Sept., 2001XXEnd of Document List
Key to purpose codes: I = Information and discussion, P = Proposal, O = Output of meeting or subject-specific activity, R = Report of the meeting

Annex B: List of Meeting Participants
ITU-T Q.6/16 Video Coding Experts Group (VCEG) VCEG-N02 rev. 1 (filename VCEG-N02r1.doc)
14th meeting, Santa Barbara, California, USA, September 24-27, 2001 last date revised: September 27, 2001
LIST OF PARTICIPANTS FOR THE 14TH MEETING OF ITU-T Q.6/16 VIDEO CODING EXPERTS GROUP (VCEG)

NAMEORGANIZATIONCOUNTRYTELEPHONEFACSIMILEE-MAILATTENDANCEDr. Demin WangCommunications Research CenterCanada+1 613 991 5621+1 613 990 6488demin@m31.dgbt.crc.calmn*Mr. Anthony JochUniversity of British ColumbiaCanada+1 604 737 2426anthonyj@ece.ubc.caknMr. Patrick RaultVixsCanada+1 416 646 2000+1 416 646 1042prault@zixs.comn*Mr. Miska HannukselaNokiaFinland+358 40 5212845miska.hannuksela@nokia.comdhijklmnMr. Marc LegrandPhilips Digital NetworksFrance+33 1 47 28 37 33+33 1 1 47 28 37 25marc.legrand@philips.comnMs. Cécile DufourPhilips Research FranceFrance+33 1 47 28 35 44+33 1 47 28 35 05cecile.dufour@philips.comlnDr. Peter ListDeutsche TelekomGermany+49 6151 83 3878+49 6151 83 4842peter.list@t-systems.deabcdeghijklmnMr. Karsten SuehringHHIGermany+49 3031002266suehring@hhi.denDr. Thomas WiegandHHIGermany+49 30 31 002 617+49 30 392 7200wiegand@hhi.decdefghijlmnMr. Thomas StockhammerMunich University of TechnologyGermany+49 89 28923474+49 89 28923490stockhammer@ei.tum.deiklmn*Mr. Mathias WienRWTH AachenGermany+49 241 8027681+49 241 80 22196wien@ient.rwth-aachen.deh*j*klmn*Dr. Stephan WengerTeles AGGermany+49 172 3000813+49 30 36282387stewe@cs.tu-berlin.dea*bcdefghijklmnMr. Markus FlierlUniversity of Erlangen-NurembergGermany+1 650 723 3476+1 650 723 8473mflierl@stanford.edul*m*nMr. Satoru AdachiNTT DoCoMo, Inc.Japan+81 468 40 2519+81 468 40 3788adachi@spg.yrp.nttdocomo.co.jpjnMr. Shun-ichi SekiguchiNTT DoCoMo, Inc.Japan+81 468403517shun@mml.yrp.uttdocomo.co.jpnDr. Byeong-Moon JeonLG ElectronicsKorea+82 2 526 4415+82 2 3461 1347jeonbm@lge.comjklmnDr. Chul-Woo KimMcubeWorksKorea+82 31 601 5875+82 31 601 5878charlie@mcubeworks.comfghjkmn*Mr. Angelo KimOnTimetekKorea+82 2 3472 4529+82 2 3472 4593drgfly@ontimetek.comnMr. Sang-Hee LeeSK TelecomKorea+82 2 2121 3466raimism@sktelecom.comn*Mr. Sang-Woo RhieSK TelecomKorea+82 2 2121 2897shleens@sktelecom.commn*Dr. Min-Cheol HongSoongsil UniversityKorea+82 2 820 0716+82 2 821 7653mhong@e.ssu.ac.krfghijklmnDr. Till HalbachNTNU Norway+47 4152 963halbach@tele.ntnu.non*Mr. Tom-Ivar JohansenTandbergNorway+47 67 125 125+47 67 125 234tij@tandberg.nobcdfhjklmnDr. Richard Fryer Essential Viewing Systems Ltd.Scotland+44 1 41 552 4773+44 1 41 552 5330richard.fryer@essential-viewing.comacdf*ghi*kn*Mr. Rickard SjöbergEricssonSweden+46 8 7572373+46 8 7575550rickard.sjoberg@era.ericsson.seagijklmnMr. Sherman ChenBroadcomUSA+1 949 293 9684schen@broadcom.comnMr. José AlvarezBroadcomUSA+1 408 501 8347+1 408 501 7836jalvarez@broadcom.comnMr. Huipin ZhangDoCoMo Comm. Labs. USAUSA zhang@docomolabs-usa.comnDr. Pankaj TopiwalaFastVDO Inc.USA+1 410 730 9191+1 410 730 2957pnt@fastvdo.comklmnMr. Dzung HoangGlobespan USA+1 408 855 2572dhoang@globespan.netnDr. Eric ViscitoGlobespan SemiconductorsUSA+1 408 919 6514+1 408 727 5078viscito@globespan.netklnMr. John ApostolopoulosHewlett-Packard LaboratoriesUSA+1 650 857 4416+1 650 852 3791japos@hpl.hp.comnMr. John JohlIntrinsixUSA+1 949 250 4395 ext102jjohl@intrinsix.comnMr. John NelsonIntrinsixUSA+1 805 558 8055jnelson@intrinix.com.co.jpnMr. Jordan IsailovicJRI TechnologyUSA+1 714 921 2286isailovicj@yahoo.comn*Ms. Iole MoccagattaLSI LogicUSA+1 408 954 3101+1 408 954 3599iole.moccagatta@lsil.comnDr. Henrique S. MalvarMicrosoftUSA+1 425 936 2978+1 425 706 7329malvar@microsoft.comnDr. Gary J. SullivanMicrosoftUSA+1 425 703-5308+1 425 706-7329garys@ieee.orgnMr. Rajeev GandhiMotorolaUSA+1 858 404 2388rgandhi@si.comnDr.Faisal IshtiaqMotorolaUSA+1 847 538 3782+1 847 576 8378faisal@labs.mot.comimnMr. Ajay LuthraMotorolaUSA+1 858 404 3470+1 858 404 2501aluthra@gi.comnDr. Limin WangMotorolaUSA+1 858 404 3893+1 858 404 2501liwang@gi.comlnDr. Marta KarczewiczNokiaUSA+1 972 894 6907+1 972 894 4589Marta.Karczewicz@nokia.comacdfghmnMr. Ragip KurcerenNokiaUSA+1 214 460 5367+1 972 894 4589ragip.kerceren@nokia.comlmnMr. Jani LainemaNokiaUSA+1 214 803 0039+1 972 894 4589jani.lainema@nokia.comdfghijklmnDr. Jiangtao WenPacketVideo Corp.USA+1 858 731 5381+1 858 455 2532gwen@pv.combdhinMr. David LindberghPictureTelUSA+1 978 292 4351+1 978 292 3318lindbergh@pictel.comacklmnMr. Chad FoggPixonicsUSA+1 408 829 5506chadfogg@yahoo.comi*jklmnDr. Michael HorowitzPolycomUSA+1 512 372 7091+1 512 372 7001mhorowitz@austin.polycom.comhijklmnMr. Karl LillevoldRealNetworksUSA+1 206 892 6286+1 206 674 3580karll@real.comabgi*kmn*Mr. Brian HogstromSensormaticUSA+1 858 642 2400 ext 1237+1 858 642 2445bhogstrom@sensormatic.comn*Dr. Louis KerofskySharp Labs of AmericaUSA+1 560 817 7644+1 360 817 8436lkerofsky@sharplabs.comjklmnDr. Shawmin LeiSharp Labs of AmericaUSA+1 360 817 8446+1 360 817 8436shawmin@sharplabs.coma*i*nDr. Shijun SunSharp Labs of AmericaUSA+1 360 817 7531+1 360 817 8436ssun@sharplabs.comkmnMs. Sang-Eun Han Stanford UniversityUSA+1 650 723 9564+1 650 724 3648sehan@stanford.edunMr. Minhua ZhouTexas InstrumentsUSA+1 214 480 3816+1 972 761 6969zhou@ti.comnMs. Jill BoyceThomson MultimediaUSA+1 609 734 9859+1 909 734 9870boycej@tce.comnMr. Adam LiUCLAUSA+1 310 825 5178+1 310 825 7928adamli@ucla.edugjknDr. Peter BorgwardtVideoTele.comUSA+1 503 594 1303+1 503 594 1456peter.arthur.borgwardt@videotele.comn*
Annex C: Liaison Letter to ISO/IEC JTC1/SC29/WG11 (MPEG)
Document: VCEG-N82 (VCEG-N82.doc)

ITU-T STUDY GROUP 16 Question 6 Video Coding Experts Group (VCEG)
Santa Barbara, 24-27 September 2001Questions:6/16Source:ITU-T SG 16 Q.6/16Title:Liaison Statement to ISO/IEC JTC1/SC29/WG11 Re: Future Joint Video Coding Work

LIAISON STATEMENT
From:ITU-T SG 16 Q.6/16To:ISO/IEC JTC1/SC29/WG11Approval:Approved by SG 16 Question 6 VCEG (Santa Barbara, 27 September 2001)For:Response and ActionDeadline:November 12, 2001 if possible (for consideration at Meetings of ITU-T SG16 WP3 in Geneva)Contact:Gary SullivanMicrosoftTel: +1 425 703 5308One Microsoft WayFax: +1 425 936 7329Redmond / WA / 98052Email: Gary.Sullivan@itu.ch USA

ITU-T Q.6/16 (“VCEG”) thanks ISO/IEC JTC1/SC29/WG11 (“MPEG”) for its Sydney liaison regarding the formation of a joint video team. VCEG continues to be interested in working together with MPEG toward a single unified video codec standard.

At VCEG’s Santa Barbara meeting, we reviewed the goals and scheduling plans for the H.26L project in view of MPEG’s liaisons and goals. We believe there are no barriers to the proposed joint project in this regard. The refined goals and schedule approved by VCEG are attached [VCEG-N86].

Careful consideration was given to the Terms of Reference that will be appropriate for our joint work, with particular attention to policies and procedures that will ensure the most harmonious, highest-quality, and most productive technical work.

As a result of these discussions, VCEG proposes the attached revised draft Terms of Reference [VCEG-N87, with VCEG-N86, VCEG-L33r5, and VCEG-N70r3] as the basis for our joint work. If MPEG agrees with these proposals, and after consultation with SG16 management, the joint work can commence as soon as each respective organization gives formal approval.

Actions Taken at Santa Barbara Meeting

The technical work on the draft H.26L standard progressed well at the Santa Barbara meeting. In order to enable coordination of technical status, we provide the following short description of the actions taken at the meeting.

Changes to TML draft standard document [VCEG-N83]:
Correct description of picture sync for consistency
Change picture header content to support indication of interlaced picture types
Add syntax mechanism for carrying supplemental enhancement information [VCEG-N60]
Correct description of direct skip mode handling for B pictures, CBP list for B pictures, syntax diagram
Authorize significant restructuring of document as necessary for clarity
Specify that the draft “Baseline” profile contains all technical features except 1/8-pel motion, CABAC, SP frames, and B pictures
Alter network adaptation layer normative text [VCEG-N72]
Alter method for handling the edges of pictures for motion compensation to first apply picture area extension and then apply interpolation filter processing [VCEG-N19]
Alter method of performing motion compensation for sub-pixel position interpolation [VCEG-N31] (adopted conditioned on provision of software and no problems surfacing soon after the meeting)
Simplify loop filter design [VCEG-N08]
Add multiple leaky bucket VBV/HRD model [VCEG-N58r1]
Incorporate capabilities of H.263 Annex U for multiple reference picture buffer management (previous plan reemphasized)
Add SP picture description (previous plan reemphasized)

Changes to Experiment Common Conditions and Non-Normative Encoding and Decoding Techniques for Experiments:
Add the use of rate-distortion optimization to common conditions for experiments in the area of coding efficiency capability
Change common conditions for testing performance for RTP/IP on 3GPP [VCEG-N37]
Tentatively adopted incorporation of multiple parallel decoding technique for error resilience testing capability [VCEG-N38, VCEG-N39, VCEG-N50] (adopted conditioned on checking and reporting on operation with new error concealment method and slice size based on number of bytes threshold)
Adopted new error concealment technique [VCEG-N62, VCEG-N63]
Tentatively adopted reduced-complexity motion estimation [VCEG-N25] (final decision on reflector)
Specify common conditions for testing for
wire-line low-delay IP/UDP/RTP packet loss resilience [VCEG-N79]
Mobile IP/UDP/RTP on 3GPP/3GPP2 [VCEG-N80]
Coding efficiency [VCEG-N81]

Ad-Hoc Groups Established:
H.26L Project Development (Chair: Gary Sullivan)
Profiles and applications (Chair: Thomas Wiegand)
Complexity Minimization (Chair: Michael Horowitz)
Transform and Quantization Complexity Reduction (Chair: Louis Kerofsky)
Generalized Annex U Incorporation (Chair: Thomas Wiegand)
Test Model and Software Development (Chair: Karsten Sühring)
Motion Interpolation Processing (Chair: Thomas Wedi, Co-Chair: Marta Karczewicz)
Motion Estimation Methods (Chair: Pankaj Topiwala)
Interlaced Video Coding (Chair: Peter Borgwardt)
Loop Filter Design (Chair: Peter List)

Core Experiments Planned:
Global Motion Vector Coding [VCEG-N77]
Reduced-Complexity Transform and Quantization [VCEG-N78]
Interlaced Video Coding [VCEG-N85]

Attachments:
VCEG-N86, Q.6/16 Goals for the H.26L Project and Proposed Joint Schedule
VCEG-N87, Terms of Reference Proposed for Joint Video Project
VCEG-L33r5, VCEG Rules for New Proposals to TMLx and H.26L, Eibsee, January 2001
VCEG-N70r3, IPR Strategy and Guidelines for H.26L, Santa Barbara, September 2001

END ANNEX C MAIN TEXT [KEEP THIS LINE TO HELP THE TSB]

____________________

Annex C Attachment 1: Q.6/16 Goals for the H.26L Project and Proposed Joint Schedule
ITU - Telecommunications Standardization Sector
STUDY GROUP 16 Question 6
Video Coding Experts Group (VCEG)
_________________
14th Meeting: Santa Barbara, CA, USA, 24-27 Sep., 2001Document VCEG-N86
Filename: VCEG-N86.doc
Generated: 27 September ’01
Question:Q.6/SG16 (VCEG)Source:Gary Sullivan Microsoft Corp. One Microsoft Way Redmond, WA 98052 USA Tel: Fax: Email: +1 (425) 703-5308 +1 (425) 936-7329 Gary.Sullivan@itu.intTitle:Q.6/16 Goals for the H.26L Project and Proposed Joint Development SchedulePurpose:OutputStatus:Approved by VCEG_____________________________
Q.6/16 Goals for the H.26L project
Video Coding Project Goals:
Simplification “back to basics” approach
adoption of a generally simple, straightforward design using well-known building blocks
for example, use of a minimal number of VLC tables (e.g. one) for all parameters to be coded
minimum number of conformance points
High compression performance
capable of 50% or greater bit rate savings from ’98 H.263v2 (with Annexes DFIJ&T) at all bit rates
Flexible application to delay constraints appropriate to a variety of services
low delay (e.g., no B pictures) for real-time conversational services
higher delay usage appropriate for storage or sever-based streaming application
Network friendliness
ease of packetization
information priority control
application to video streaming services
Error resilience
packet loss resilience
mobile channel corruption resilience
Complexity scalability in encoder and decoder
asymmetry of encoder and decoder processing complexity
scalability between amount of encoder processing and achievable quality
Full specification of decoding (no mismatch)
resolve mismatch problem (e.g., integer transform, VQ,…)
High quality application
performance improvement in higher bitrate
applicability to entertainment-quality applications
File storage support
simple stream exchange
http streaming service
random access
support of multiple streams with transitions (e.g., via SP pictures)
Royalty-free baseline codec

Proposed Joint Work Schedule (ITU-T H.264 & MPEG-5 Video)

Approx DateAuspicesPossible LocationProject MilestoneDec, 2001JTC1PattayaJWD 1Feb, 2002ITU-TGenevaJWD 2Apr, 2002UnknownNorth AmericaCDJul, 2002JTC1KlagenfurtFCDOct, 2002ITU-TGenevaAAP ConsentDec, 2002JTC1UnknownFDIS & Rec.Feb, 2003N/AIS
Annex C Attachment 2: Terms of Reference Proposed for Joint Video Project
ITU - Telecommunications Standardization Sector
STUDY GROUP 16 Question 6
Video Coding Experts Group (VCEG)
_________________
14th Meeting: Santa Barbara, CA, USA, 24-27 Sep., 2001Document VCEG-N87
Filename: VCEG-N87.doc
Generated: 27 September ’01
Question:Q.6/SG16 (VCEG)Source:Gary Sullivan Microsoft Corp. One Microsoft Way Redmond, WA 98052 USA Tel: Fax: Email: +1 (425) 703-5308 +1 (425) 936-7329 Gary.Sullivan@itu.intTitle:Terms of Reference Proposed for Joint Video ProjectPurpose:OutputStatus:Approved by VCEG_____________________________
[DRAFT PROPOSED]
Terms of Reference for a Joint Project between ITU-T Q.6/SG16 and ISO/IEC JTC 1/SC 29/WG11 for the Development of new Video Coding Recommendation and International Standard
1.0 Goals, Intent and Scope of the Joint Project
These Terms of Reference (ToR) apply to the joint project of ITU-T Q.6/SG16 (“VCEG”) and ISO/IEC JTC1/SC29/WG11 (“MPEG”).

The ultimate goal of the joint project is to complete and extend draft ITU-T H.26L as a single joint VCEG/MPEG video coder and decoder (“codec”) standard for general use, including in both ITU and ISO/IEC based systems. The goals and schedule are as described in VCEG-N86.

The intent is that the resulting ITU-T Recommendation and ISO/IEC International Standard will be technically identical, fully interoperable with each other, and offer the best possible technical performance, under the practical constraints of available implementation platforms for various applications.

The scope of the joint project will be the development of a new ITU-T Recommendation and a new ISO/IEC International Standard for a digital video codec.

The ITU-T Recommendation and ISO/IEC International Standard will be developed and drafted by a single jointly-chartered group. The formal common text process will not be used (in the interest of minimising co-ordination overhead) but alternative mechanisms will be employed to ensure technically identical results.
2.0 Joint Group
The work will be conducted by a jointly-constituted experts group which will be known as the Joint Video Coding Experts Group (“J-VCEG”). This name is chosen to clearly describe the nature of the group, and for its similarity to the well-known VCEG name, under which this work has been conducted for some years. The intent is to avoid confusion and any perception of instability among industry and press.

J-VCEG will operate as a joint group under the ordinary policies and procedures of both organisations. In the event of differences between policies of ISO/IEC and ITU-T not covered by these ToR, the J-VCEG Rapporteur will decide the issue, based on the consensus of the experts, in the best interests of standardization.

The group will develop a Recommendation for the ITU-T Alternative Approval Process (“AAP”) and a corresponding (technically identical, but not formally common text) International Standard for the ISO/IEC approval process.
3.0 Deliverables of the Joint Project
The deliverables are a new video codec to be called “J-VCEG-1”, to be approved by ITU as ITU-T Recommendation H.264 and by ISO/IEC as International Standard “MPEG-5 Video” (in order to indicate the dramatic performance improvement compared to previous standards). These deliverables will be developed with goals and requirements as described.

4.0 Dissolution
The joint group dissolves when the approval process in both organisations is completed for the new Recommendation and International Standard, or when in the independent opinion of the formal members of either ISO/IEC JTC 1/SC 29/WG11 or ITU-T Q.6/SG16, the agreed arrangements have not been honored.

Any potential joint work beyond the duration of this project (e.g., extensions, amendments, etc.) requires further mutual agreement.
5.0 Meetings
J-VCEG meeting venue and dates will be proposed by the J-VCEG Rapporteur, and authorized and announced as Rapporteur working-level meetings under the customary practices of both organisations. Meetings will operate under J-VCEG’s rules as described in this ToR.

J-VCEG meetings will be held as a separate entity, and will operate under rules as agreed in this ToR. Meetings should preferably be co-located with an ITU-T SG16 or ISO/IEC JTC/SC29 meeting.

The meeting dates and locations should be co-ordinated with those of meetings of the ITU-T SG16 and ISO/IEC JTC 1/SC 29/WG 11 (e.g. on an alternating basis if feasible for the progress of the project) in order to reduce the amount of travelling for participants. All technical work on the project will be conducted in J-VCEG meetings. J-VCEG meetings will be preferably held immediately before, during, or after the corresponding SG16 or WG 11 meeting dates.
6.0 Management
The management of the J-VCEG will consist of a Rapporteur/Chair (Gary Sullivan) and two Associate Rapporteurs/Co-Chairs (one each as appointed from SC 29/WG 11 and SG16 with joint consent), reporting to ITU-T SG16 WP3 and ISO/IEC JTC 1/SC 29/WG 11. Changes in the management team must be agreed by both organisations.
7.0 Participation in J-VCEG meetings
Meetings will be open to, and contributions accepted from, all parties qualified for participation either in a ITU-T Q.6/SG16 Rapporteur’s meeting or an ISO/IEC JTC 1/SC 29/WG 11 meeting, including experts invited by the Rapporteur, as permitted by ITU-T policy.

All contributions and all participants will have equal status in the technical work and will be considered on the same terms.
8.0 Documents and Contributions
J-VCEG will maintain a document registry and electronic distribution archive. The registry and archive will be linked to both the ITU-T Q.6/SG16 and ISO/IEC JTC 1/SC 29/WG 11 ftp sites.

Any document from a participant in the meeting should be available to all the participants before the meeting through the use of electronic document handling. A registration and uploading deadline several days in advance of the start of the meeting will be announced for each meeting. A “late, unannounced” document hand-carried to the meeting should be accepted only with the consensus of the meeting participants. This policy will be stated in the invitation letter that is provided for every meeting to both organisations.

All documents and contributions will be in electronic form (preferably MS Word).

In order to facilitate cross-organisational communication, all documents will be public unless a contributor indicates otherwise. In that exceptional circumstance, the document will be accessible only through a private, password-protected site accessible only to ITU/ISO/IEC members and invited experts.
9.0 Working Methods

9.1 General Policies and Procedures
All group decisions will be made by the consensus of the J-VCEG experts as determined by the J-VCEG Chairman/Rapporteur. All contributions related to the joint project will be addressed to J-VCEG for the duration of the joint project, rather than to the individual sponsoring organizations.

The general rules for handling new proposals, and general polices described in VCEG-L33r5 will remain in force. These rules may be modified by a consensus of the J-VCEG experts as determined by the J-VCEG Chairman/Rapporteur

9.2 Requirements
All requirements for the project, including baseline codec, profiles, levels, and versions will be determined by a consensus of the J-VCEG experts, based on contributions from all interested parties.

The initial set of requirements are described in VCEG-N86.

9.3 Document Control
J-VCEG will maintain a single master draft document and a single reference software codebase for the developing Recommendation/Standard, each under the control of a single editor, appointed by the J-VCEG Rapporteur with the consensus of the experts. The document and codebase will contain the exact text to be submitted to ITU-T and ISO/IEC for approval, so that the resulting ITU-T Recommendation and ISO/IEC International Standard will be word-for-word identical, and technically fully interoperable.

Amendments, corrections, or corrigenda will be processed for approval to maintain technically identical Recommendations/Standards as necessary, based on the judgment of the J-VCEG editors and Rapporteur.
10.0 Patent and Copyright Issues
The project and joint group will progress the project work in compliance with the Intellectual Property Rights (“IPR”) policies and IPR reporting requirements and procedures of both organisations ( HYPERLINK "http://www.itu.int/ITU-Databases/TSBPatent/" http://www.itu.int/ITU-Databases/TSBPatent/ and HYPERLINK "http://www.itscj.ipsj.or.jp/sc29/"http://www.itscj.ipsj.or.jp/sc29/29w7ipr.htm).

VCEG’s rules for the implementation of the ITU-T IPR policy are contained in documents VCEG-L33r5 and VCEG-N70r3 and will continue in force. (All references to ITU-T or VCEG in these pre-existing documents will be considered as referring to ITU-T/ISO/IEC and J-VCEG.) These rules are consistent with both the ITU-T and SC29 policies referenced above, and the rules may be modified in a manner consistent with those policies by a consensus of the J-VCEG.
11.0 Liaison Statements
The J-VCEG will conduct liaison communications. J-VCEG may approve such communications by a consensus of the J-VCEG experts. All liaison statements will indicate the level of approval and copies of all liaison communications will be provided to the secretariat of both organisations, posted on the J-VCEG FTP sites, and provided to the experts.
12.0 Meeting Reports
A meeting report will be provided by the J-VCEG management shortly after the conclusion of each meeting and will be submitted to ITU TSB and ISO/IEC JTC 1/SC 29/WG 11, posted on the group’s FTP sites, and distributed to the experts.

The report should include:
Dates and venue
Chairpersons/Rapporteurs of the meeting
Attendance list with affiliation
Agenda of the meeting
List of documents considered with source
Summary of results and an outline of any outstanding issues or resolutions
Any outgoing liaison statements/communications sent to other organisations
Future activities
13.0 Promotion and Public Relations Activities
Any public relations or promotional activities regarding the joint group, its project, and its results and deliverables will be approved by the J-VCEG Chairman/Rapporteur with the consensus of the experts, and jointly conducted or conducted in an otherwise mutually agreed fashion.
Annex C Attachment 3: VCEG Rules for New Proposals to TMLx and H.26L
ITU – Telecommunications Standardization Sector
STUDY GROUP 16
Video Coding Experts Group (Question 6)
_________________
12th Meeting: Eibsee Germany, 9-12 January 2001Document VCEG-L-33r5
File: VCEG-L-33r5.doc
Generated: 11 Jan 2001
Question:Q.6/16Source:Dave Lindbergh PictureTel Corp. 100 Minuteman Road Andover, MA 01810 USA
Stephan Wenger TU Berlin, Sekr. FR 6-3 Franklinstr. 28-29 D-10587 Berlin GERMANYTel: Fax: Email:

Tel: Fax: Email:+1 978 292 4351 +1 781 944 1267  HYPERLINK "mailto:lindbergh@pictel.com" lindbergh@pictel.com

+49-172-3000813 +49-30-31425156  HYPERLINK "mailto:stewe@cs.tu-berlin.de" stewe@cs.tu-berlin.deTitle:VCEG Rules for New Proposals to TMLx and H.26LPurpose:Information_____________________________
The rules in this document were formally approved by VCEG (ITU-T Q6/16) at their 12th meeting (Eibsee Germany, January 2001).



VCEG Rules for New Proposals to TMLx and H.26L

1. Abstract

This document defines rules VCEG will use for the future standardization process of H.26L.

VCEG reminds the reader that the ultimate purpose of standardization is not to develop an official Recommendation providing the highest possible theoretical performance, but to encourage actual interoperation between implementations in the field, at the highest practical level of performance. “Optional” modes which are unlikely to be commonly implemented do not enhance performance in the field, despite any theoretical advantages. This observation drives many of these rules.

The rules help to ensure maximal interoperability and simple capability negotiation of systems employing H.26L. A Profile-Version-Level framework is introduced, that, in spirit, corresponds to the one previously used for Annex X/H.263 (although hopefully doing a better job by preventing divergence at the outset).

Note that all these rules apply only to the described types of proposals. These rules do not apply to informative or discussion contributions.


2. Definitions

The Baseline mode is the common mode supported by all H.26L implementations, without exception. The Baseline supports video communication between all ITU-T H.32x series systems Recommendations (H.320, H.323, H.324).

A Profile is defined by a particular tradeoff between:

Encoder RAM
Encoder ROM
Encoder MIPS
Decoder RAM
Decoder ROM
Decoder MIPS
Latency
Error resilience
within the encoder or decoder implementation. Applications which require similar tradeoffs between these parameters should use the same Profile.

New Versions are defined by a change in syntax that will allow an improvement in performance for a given Profile and Level. Within each Profile, successive Versions may be defined by VCEG over time. Each Version represents the “best” available technology for the application identified by the Profile at the time of approval of the Version.

A Level describes performance parameters within each Profile, which describe upper limits on decoder capability. Performance parameters may include maximum picture size, macroblocks per second, bitrate, and similar parameters. See the Test Model document for a list of Levels. See Figure 1.

Approval
Date Profile X Profile Y Profile Z

2007 Version 4 L1 L2 L3
2005 Version 3 L1 Version 3 L1 L2
2003 Version 2 L1 L2 Version 2 L1 Version 2 L1 L2 L3
2001 Version 1 L1 L2 Version 1 L1 L2 Version 1 L1
2001 -----------------------BASELINE------------------------

Figure 1 - Profile-Version-Level scheme (example only)


3. Rules for VCEG itself

VCEG shall ordinarily define each Level and Version for a given Profile to require support of all lower-numbered Levels (L) and Versions (V) for the same Profile (P). Therefore, when an endpoint signals Pa Vb Lc, this indicates the ability to operate in any mode Profile a, Version (1..b), Level (1..c).

For example, in Figure 1, PZ V3 L2 requires support for PZ V3 L1, PZ V2 L1, PZ V2 L2, PZ V1 L1, and the baseline. It does not require support for PZ V2 L3. However, VCEG may exceptionally drop backward compatibility, primarily in cases where serious problems have been found with one or more previous versions.

The minimum time between new Versions is normally 2 years.

VCEG shall not begin the ITU-T approval process for Recommendation H.26L or any Annex(es) thereof, until the complete C-code for that version has been shown to work and test bitstreams have been generated and tested on the C-code model.

Contributed proposals which do not comply with the following rules will be considered by VCEG only with the consensus of the group, and only if time is available after consideration of normal proposals. VCEG may reach a tentative conclusion regarding such non-compliant proposals, but no proposal shall be formally accepted until the contributor supplies the required materials, and these are approved by VCEG.


4. Rules for Proposals of New Profiles

Every proposal of a new Profile or for a change in the requirements of an existing draft Profile shall include:

A description of the intended application from a user’s point-of-view.

Examples:
A Profile including “high encoder resource consumption, low decoder resource consumption, not latency-sensitive, highly error resilient”, might be appropriate for wireless streaming video applications.
A Profile including “moderate encoder resource consumption, moderate decoder resource consumption, latency-sensitive, not error resilient”, might be appropriate for ISDN video conferencing.

This description should include quantitative target values for the above parameters.

A list of Non-Required Attributes for the Profile. For example, “Encoder computational complexity may be very high, Algorithmic Delay is unimportant”.

A justification why proposed applications are not well-served by existing H.26L profiles.

At least one technical proposal for a solution of the posed problem, see below, OR an informative description of a possible solution. In the latter case it is expected that a detailed proposal will be presented at the next meeting, otherwise the profile, if accepted, will be deleted then.

A text draft of Common Simulation Conditions appropriate for the new Profile.


5. Rules for Proposals of New Versions

All proposals to start work on a new Version of a profile shall include a proposed requirements document for the version, stating quantitative target values for codec performance metrics such as computational complexity, delay, memory, etc.

The proposal shall also include evidence that a significant performance improvement can be achieved compared to the existing version.


6. Rules for Proposals of New Levels

Every proposal of a new Level for an existing Profile shall include:

A description of the intended performance parameters for the new Level. See the definition of ‘Level’.

A justification why the current Levels are inadequate.

A text draft of Common Simulation Conditions appropriate for the new Level.


7. Rules for Technical (algorithmic) Proposals

All technical proposals for algorithms to be included in one or more Profile-Version-Levels shall include:

The results of a reference implementation relative to the official TML design.

In the case of enhancements or evolutionary developments of algorithms already present in TML-n, the results must be based on the most recent TML-n. In the case of algorithms that address fundamentally new issues, it is sufficient that the software is relative to the earlier TML-n-1. This requirement is distinct from the requirement of timely inclusion of the software implementation into TML-n+1 in case of adoption (see below).

A textual description that fits into the current structure of the Test Model, and a change-bar version of TML-n including the new proposal, as a hint to the editor how the resulting TML-x+1 may look like.

A written transfer of copyright in the contributed C-code to the ITU. As the H.26L C-code model is a collaborative work by the VCEG experts, copyright on the jointly-developed software code shall reside in the ITU, and contributors must agree to this condition.

Contributors of software code retain the right to do whatever they wish with software they have contributed. Contributors also retain full patent rights in their contributed algorithms, according to the ITU-T Patent Policy. The ITU will grant, free of charge, the right to use this software or variations on this software to anyone for purposes of implementation of the standard. VCEG intends to include the final C-code version of encoder and decoder into the standard in the form of some normative or informative part of the standard (at the discretion of the group).

This transfer of copyright shall be contained in the headers to all source C-code, in the exact following form:

/*
***********************************************************************
* COPYRIGHT AND WARRANTY INFORMATION
*
* Copyright , International Telecommunications Union, Geneva
*
* The contributor(s) hereby donate this source code to the ITU, with
* the following understanding:
* 1. Contributor(s) retain the right to do whatever they wish
* with the contributed source code, without limit.
* 2. Contributor(s) retain full patent rights (if any exist) in
* the technical content of techniques and algorithms herein.
* 3. The ITU shall make this code available to anyone, free of
* license or royalty fees.
*
* DISCLAIMER OF WARRANTY
*
* These software programs are available to the user without any
* license fee or royalty on an "as is" basis. The ITU disclaims
* any and all warranties, whether express, implied, or
* statutory, including any implied warranties of merchantability
* or of fitness for a particular purpose. In no event shall the
* contributor or the ITU be liable for any incidental, punitive, or
* consequential damages of any kind whatsoever arising from the
* use of these programs.
*
* This disclaimer of warranty extends to the user of these programs
* and user's customers, employees, agents, transferees, successors,
* and assigns.
*
* The ITU does not represent or warrant that the programs furnished
* hereunder are free of infringement of any third-party patents.
* Commercial implementations of ITU-T Recommendations, including
* shareware, may be subject to royalty fees to patent holders.
* Information regarding the ITU-T patent policy is available from
* the ITU Web site at http://www.itu.int.
*
* THIS IS NOT A GRANT OF PATENT RIGHTS – SEE THE ITU-T PATENT POLICY.
************************************************************************
*/

IMPLEMENTATION OF RULES (a), (b), (c)

Only for rules (a), (b) and (c) above, the contributor does not need to include these items in an initial contribution to VCEG. In the case that VCEG tentatively accepts the proposal, this will be noted in the VCEG Meeting Report, and the contributor is then required to supply items (a), (b) and (c) before formal inclusion of the proposal. The VCEG Document Editor and Source Code Coordinator have the discretion to accept items (a), (b), and (c) between VCEG meetings. Such acceptance shall be notified to and reviewed by VCEG at their next meeting.

An intellectual property (patent) statement. This shall be in the format described in VCEG-L-37r1.

Performance and complexity figures as discussed below in the attached “Form for Resource Consumption Metrics”.

VCEG will choose proposals that in the opinion of the experts, best meet the Requirements for the Profile-Version-Level. As the metrics describing the performance offered by contributions will not be directly comparable to the metrics used in the requirements document (due to differences in development platform, optimization, etc.), the experts will have to make judgments about appropriate correction factors to compensate for measurement differences.

It is the policy of VCEG that all technical results shall be confirmed by independent implementations tested under common conditions. At the discretion of VCEG, technical results may be confirmed by submission of source code software to be tested under common conditions.

Note: in a later stage of the standardization process it may be desirable to have less strict rules for inclusion into TML provided that syntax and semantics of the bitstream/packetstream remain unchanged. (Similar to past rules regarding adopting proposals into TMN versus adopting proposals into the H.263++ draft).

8. Procedure to add code of accepted proposals into TML

Immediately after the closing of each meeting, all proponents of accepted proposals and the software coordinator meet (in person, or by telecommunication) in order to synchronize the implementation efforts. During this meeting the software coordinator assigns, solely on technical parameters, timeslots, during which proposals will have to be included into the new software release. Length, order, and eventual overlapping of those timeslots depend on the breakout group’s judgment on the best possible way to integrate the various proposals into the common code base. All integration work should be finished after half of the period between two meetings (including Study Group meetings that contain a Rapporteur’s meeting).

The software coordinator maintains an RCS-based source code archive. Any VCEG member can request from the software coordinator any version of TML, which will be sent by Email as soon as possible.

In case of situations where a proponent does not finish the integration work during the assigned timeslot, or the software coordinator judges the integration work’s quality inappropriate for acceptance, the software coordinator will inform the VCEG experts through the Email reflector itu-adv-video. The Rapporteur then may grant a single time extension for that proponent. If the Rapporteur doesn’t grant an extension, or if no successful integration work took place during the extension period, the proposal is automatically deemed to be withdrawn. It may, however, be proposed again at the next meeting.

Following these procedures, in the middle of the period between two VCEG meetings, software of reasonable quality will exist, that include all features of the actual TML document. All integration work for new technical proposals, as discussed in section 3 above, has to take place relative to this software base.

Software shall be provided in C or C++. The software should compile at least on the most current version of MS-C++. The use of specific features of this compiler, however, should be avoided, and the software coordinator will reject code that, in his judgment, is not portable to other platforms.

There is no real-time requirement for the code, but, as computational complexity is an important factor, the submission of somewhat optimized software is encouraged. It is VCEG’s intent to keep at least the decoder close to real-time performance (e.g., for QCIF resolution at 15 fps).


Form for Algorithm Resource Consumption Metrics

When VCEG experts contribute proposals for H.26L algorithms, contributions shall be accompanied by Resource Consumption Metrics. Contributions without the form below fully populated will not be considered by VCEG.

----

Algorithm Name: [one line]
Algorithm Description: [brief summary, one to five lines]
Source:
Contact:

Encoding metrics

Computational complexity: [cycles/second] on [platform] under [conditions [scene, etc.]]
{ “platform” may be hardware, simulation, or a theoretical calculation. But it must be stated.}
RAM memory required: [octets] on [platform] under [conditions]
ROM (table) memory required: [octets] on [platform] under [conditions]
Algorithmic delay: [seconds | frames] under [conditions]
Program space: [lines of C] | [bytes of executable] on [platform]

Decoding metrics

Computational complexity: [cycles/second] on [platform] under [conditions [scene, etc.]]
{ “platform” may be hardware, simulation, or a theoretical calculation. But it must be stated.}
RAM memory required: [octets] on [platform] under [conditions]
ROM (table) memory required: [octets] on [platform] under [conditions]
Algorithmic delay: [seconds | frames] under [conditions]
Program space: [lines of C] | [bytes of executable] on [platform]

Efficiency metrics

Signal/noise ratio: [value] under [conditions]
Compression efficiency: [bitrate] at [SNR] under [conditions]
Robustness to errors: [SNR decrease] under [conditions] at [BER] compared to [reference]


[end Annex C attachment 3]
Annex C Attachment 4: IPR Strategy and Guidelines for H.26L

ITU - Telecommunications Standardization Sector
STUDY GROUP 16 Question 6
Video Coding Experts Group (VCEG)
_________________

Fourteenth Meeting: Santa Barbara, CA, 24-27 September 2001Document VCEG-N70r3
Filename: VCEG-N70r3.doc
Generated: Jan ’01
Updated: Sep ‘01
Question:Q.6/SG16 (VCEG)Source:Istvan Sebestyen Siemens, ICN M SR3 81359 München Germany Tel: Fax: Email: +49 89 722 47230 +49 89 722 47713  HYPERLINK "mailto:istvan.sebestyen@icn.siemens.de" istvan.sebestyen@icn.siemens.deTitle:IPR Strategy and Guidelines for H.26LPurpose:Information_____________________________
This document was approved in this modified form by VCEG at their 14th meeting (Santa Barbara, 24-27 September 2001). This replaces the earlier document VCEG-L37r2.
_____________________________
H.26L Basic IPR Principles
Regarding Intellectual Property Rights (IPR) for the future ITU-T H.26L, Q.6/SG16 (VCEG) has agreed to the following basic principles:
H.26L should have a simple royalty free “baseline mode” (both on the encoder and decoder) in order to promote the wide implementation and use of ITU-T H.26L. All implementations should have such a common “baseline mode” core, in order to allow minimal interoperability among all H.26L codecs. The above requirement means that all technology applied in the “baseline mode” shall have either no, or expired IPR, or valid but license-fee-free IPR (according to Box 2.1 or 2.2.1 of the VCEG Patent Disclosure form, attached below) .
Special, more advanced profiles of the H.26L standard may contain patents per Box 2.2 of the VCEG Patent Disclosure form (reasonable terms and conditions).
Collection of IPR information during the standardization process
According to the ITU-T IPR policy, ITU-T members/experts are encouraged to disclose as soon as possible IPR information (of there own or anybody else) associated with any standardization proposal (of their own or anybody else). Such information should be provided on a best effort basis.
For collecting such information, VCEG has decided to use it’s own Patent Declaration form – note that this is distinct from the Patent Statement and Licensing Declaration that is to be submitted to the ITU Director when the contributed technology becomes part of the final standard.

Therefore, VCEG requires all technical (algorithmic) proposals include one or more of:

Attached at the end of each technical contribution, a fully filled-out “VCEG Patent Disclosure form” (attached below). At the contribution stage, this form is for information only, and may be signed by an expert or left unsigned. The form must be included in the contribution to VCEG,

Note that the submission of the VCEG Patent Disclosure form at the proposal stage does not have the same formal status as the final IPR declaration to the TSB.

Such information provided to the Rapporteur will be tabulated in a “IPR status list” (e.g. a simple Word table) of the information received. Not relevant information (e.g. if a proposed method was not accepted) will be removed from the “IPR status list” as early as possible. The “IPR status list” is a living document of VCEG.
Formal submission of IPR statements to the ITU TSB Director
In the final stage of standardization formal IPR statements using the ITU-T Patent Statement and Licensing Declaration have to be forwarded to the ITU TSB Director. This should be done by a responsible member (e.g. IPR Department) of the IPR holder. To avoid formal IPR statements on proposals not included in the final Recommendation, it is suggested that this submission be done only after the H.26L standard has been finalized for the “consent” process in the Study Group. The last call of the standard can only be issued by the TSB when all formal IPR statements have arrived to the ITU.
Outlook
It is hoped that the above implementation would facilitate to achieve the goals outlined under paragraph 1.

International Telecommunication Union Telecommunication Standardization SectorVideo Coding Experts Group - Patent Disclosure Form
(Typically one per contribution and one per Recommendation)

Please send to:
Q.6/16 Rapporteur Gary Sullivan, Microsoft Corp., One Microsoft Way, Bldg. 9, Redmond WA 98052-6399, USA
Email (preferred): Gary.Sullivan@itu.int Fax: +1 425 706 7329 (+1 425 70MSFAX)

This form provides the ITU-T Q.6/SG16 Video Coding Experts Group (VCEG) with information about the patent status of techniques used in or proposed for incorporation in a Recommendation. VCEG requires that all technical contributions be accompanied with this form. Anyone with knowledge of any patent affecting the use of VCEG work, of their own or of any other entity (“third parties”), is strongly encouraged to submit this form as well.

This information will be maintained in a “living list” by VCEG during the progress of their work, on a best effort basis. If a given technical proposal is not incorporated in a Recommendation, the relevant patent information will be removed from the “living list”. The intent is that the VCEG experts should know in advance of any patent issues with particular proposals or techniques, so that these may be addressed well before final approval.

This is not a binding legal document; it is provided to VCEG for information only, on a best effort, good faith basis. Please submit corrected or updated forms if your knowledge or situation changes.

This form is not a substitute for the ITU-T Patent Statement and Licensing Declaration, which should be submitted by Patent Holders to the TSB Director before final approval of any Recommendation.

Submitting Organization or Person:Organization name

Mailing addressCountryContact personTelephoneFaxEmailPlace and date of submissionRelevant Recommendation and, if applicable, Contribution:Number (ex: “H.26L”)TitleContribution number(Form continues on next page)

Disclosure information – Submitting Organization/Person (choose one box)2.0 The submitter is not aware of having any granted, pending, or planned patents associated with the technical content of the Recommendation/Contribution.

or,
The submitter (Patent Holder) has granted, pending, or planned patents associated with the technical content of the Recommendation/Contribution. In which case,
The Patent Holder is prepared to grant – on the basis of reciprocity for the above ITU-T Recommendation – a free license to an unrestricted number of applicants on a worldwide, non-discriminatory basis to manufacture, use and/or sell implementations of the above ITU-T Recommendation.2.2 The Patent Holder is prepared to grant – on the basis of reciprocity for the above ITU-T Recommendation – a license to an unrestricted number of applicants on a worldwide, non-discriminatory basis and on reasonable terms and conditions to manufacture, use and/ or sell implementations of the above ITU-T Recommendation.

Such negotiations are left to the parties concerned and are performed outside the ITU-T.2.2.1 The same as box 2.2 above, but in addition the Patent Holder is prepared to grant a “royalty-free” license to anyone on condition that all other patent holders do the same.2.3 The Patent Holder is unwilling to grant licenses according to the provisions of either 2.1, 2.2, or 2.2.1 above. In this case, the following information must be provided as part of this declaration:
patent registration/application number;
an indication of which portions of the Recommendation are affected.
a description of the patent claims covering the Recommendation;
In the case of any box other than 2.0 above, please provide the following:


Patent number(s)/status



Inventor(s)/Assignee(s)



Relevance to VCEG



Any other remarks:

(please provide attachments if more space is needed)

(form continues on next page)
Third party patent information – fill in based on your best knowledge of relevant patents granted, pending, or planned by other people or by organizations other than your own.

Disclosure information – Third Party Patents (choose one box)3.1 The submitter is not aware of any granted, pending, or planned patents held by third parties associated with the technical content of the Recommendation/Contribution.
3.2 The submitter believes third parties may have granted, pending, or planned patents associated with the technical content of the Recommendation/Contribution.
For box 3.2, please provide as much information as is known (provide attachments if more space needed) - VCEG will attempt to contact third parties to obtain more information:
3rd party name(s)

Mailing addressCountryContact personTelephoneFaxEmailPatent number/statusInventor/AssigneeRelevance to VCEG
Any other comments or remarks:

















[END]

File: FILENAME \* MERGEFORMAT VCEG-N88d1.doc Page:  PAGE 33 Date Printed:  DATE \* MERGEFORMAT 12/23/2001

File:  FILENAME \* MERGEFORMAT VCEG-N88d1.doc Page  PAGE 53 of  NUMPAGES 53


 :