11HEVC
HEVC technical development is under the responsibility of JCT-VC. The video subgroup assessed, jointly with the requirements subgroup, input documents that were related to extensions, including the CfP on scalable coding for HEVC, as well as profiles/levels issues, as far as they were not purely technical. When documents or NB inputs were of rather technical nature, they were brought to the attention of the JCT for appropriate processing.
On HEVC scalability: The current version of the MPEG CfP was presented in a joint meeting with VCEG on Tuesday, whereas VCEG had sent their version in the context of a liaison letter. It is intended to proceed with a joint Call (provided negotiation can be made about such a cooperation between the two organizations). Technically about the contents of the Call, no major divergence of opinions was found.
Profile/level discussion (Thursday afternoon with requirements):
-
m24302: A still picture profile may be desirable, further investigation
-
m24718: In the current structure, currently no level for QHD30 (between level 3 and 3.1 or better modify level 3). Level 4.2 and 5.3: Lower max bit rate; a second suggestion would be to have more than one level table (i.e. change the definition of levels per profile), this would eventually give up the “onion” structure in terms of decoder complexity. For different application domains, that might make sense (e.g. consumer/professional). Define profiles beyond main: Simple (up to 16x16, transcoding from AVC simplified? But loss of performance), medium (similar to current main but without tiles), separate profile for field pictures, highest coding efficiency. Certain tendency in the room to not include tiles.
-
mM24774 is similar in terms of levels, at least up to levels 3.xx.
Discussions were held about general HEVC development: After finalization of phase 1 in January 2013, extensions towards full range (color sampling beyond 4:2:0, larger bit depths) are the consequential next steps which could be finalized one year later. As it becomes evident from the work in 3DV, an MVC-like extension would not incur major definition difficulties, nor are dedicated developments necessary, provided that the inclusion of specialized tools would be kept low, and changes at CTB and below are avoided. Such a standard (at least for stereo) could be available similarly early (one year after version 1) which could be seen as a strong argument not confusing markets by various stereo solutions (such as frame compatible). Though starting after the Call by October 2012, the development of scalable extensions would probably require more time, where it would also be advantageous to have the full-range and MVC functionalities in place before the finalization of scalability, for possible combined modes. Further developments on depth-map coding and dependent-view coding, as currently investigated in the 3DV activity could be ongoing in parallel and are likely to be less urgently needed by markets than a straightforward stereo solution.
Document(s) reviewed:
m24248
|
Summary of Voting on ISO/IEC CD 23008-2 (forwarded to JCT)
|
SC 29 Secretariat
|
m24499
|
Downsampling filter for anchor generation for scalable extensions of HEVC
Discussed in AHG
|
Jie Dong, Yong He, Yan Ye (InterDigital)
|
m24703
|
Sharp’s proposals for HEVC scalability Extension
Offer of two new 4K test sequences (shot by RED one camera); Gain for MPEG-2 base + HEVC spatial enhancement reported 40% (single layer HEVC rate vs. HEVC enhancement layer), or 20% (compared to simulcast).
Discussion:
One expert noted that here each 8th frame in MPEG-2 base is intra coded, and that gain would be lower in more realistic case of less frequent I. Most likely, the gain versus simulcast by interlace base layer would be marginal.
Would the case of MPEG-2 base layer bring additional information in the CfP? This seemed unlikely, but it appears straightforward to apply hybrid coding also to the MPEG-2 base case.
Put requirement on MPEG-2 BL in reqs. document
|
Hisao Kumai, Tomoyuki Yamamoto, Andrew Segall, Norio Ito
|
m24714
|
3D-HEVC HLS: Parameter sets modifications for extension hooks (forwarded to JCT)
|
Jill Boyce, Danny Hong, Wonkap Jang
|
m24715
|
3D-HEVC HLS: SEI message for sub-bitstream profile & level indicators (forwarded to JCT)
|
Jill Boyce, Danny Hong, Wonkap Jang
|
m25113
|
Low Complexity Scalable Extension of HEVC intra pictures based on content statistics
-
uses different transform base?
-
uses optimization of non-uniform quantizers
-
For information only – no current work on this.
|
Sébastien Lasserre, Fabrice Le Léannec, Eric Nassor
|
m24263
|
USNB contribution: Comments on HEVC
|
A. G. Tescher for USNB
|
m24470
|
UKNB Comment on HEVC Extensions
|
UK National Body
|
m24728
|
JNB comment on HEVC extension
Problem with mixing inter-view and inter-frame prediction related to scaling of MV in AMVP (forwarded to JCT)
|
Japan National Body
|
m24901
|
FRNB Comments on HEVC (forwarded to JCT)
|
Pierrick Philippe, on behalf of the FRNB
|
m25009
|
JNB comments on HEVC extensions to support non-4:2:0, n-bit video
|
Japan National Body
|
m25026
|
JNB comment on study of tools for high resolution video for HEVC (forwarded to JCT)
|
Japan National Body
|
m25093
|
CANNB comments on HEVC profiles (forwarded to JCT)
|
G.Martin-Cocher, L.Winger
|
m25103
|
CNNB comments on HEVC (forwarded to JCT)
|
China National Body
|
m25203
|
ITNB comment on 3DTV Frame Packing for HEVC (forwarded to JCT)
|
IT National Body
|
Document(s) approved:
Dostları ilə paylaş: |
