Witness the development of H.265

The First JCT-VC Meeting, Dresden, DE

2010-06-19 H.265/HEVC View Comments Views(9,316)

During Apr. 15-23, 2010, the first meeting of JCT-VC was held in Dresden, DE. In the meeting, 27 responses to the Call for Proposal, issued in Jan. 2010, were reviewed and the associated video material was evaluated in extensive subjective tests.

Tentative Conclusions from CfP Responses Reviews

Coding tools used in respective proposals are listed in JCTVC-A203. High-level noteworthy aspects from the review of the proposals and the outcome of the subjective tests are outlined:

  • Substantial progress in coding efficiency has clearly been demonstrated, compared with H.264/AVC
  • There is no indication of a need to change the fundamental architecture of “conventional” hybrid video coding designs to achieve a substantial improvement
  • Inclusion of support of larger block sizes in a highly variable (typically tree-structured) block segmentation approach is a major common theme, although large block sizes were not found in all proposals that did well subjectively (and objectively)
  • Modified motion interpolation filtering is likely to be an element of the new design
  • Modified in-loop picture filtering is likely to be an element of the new design
  • Parallel processing has emerged as being increasingly important
  • New entropy coding concepts seem to be appearing that could potentially be substantially more parallel-friendly than CABAC without sacrificing coding efficiency
  • Techniques for reducing picture storage memory were in multiple proposals, providing potential memory bandwidth reduction (and other) benefits
  • There are many individual tool variants that appear in proposals as good candidates for contributing tool features to an overall design
  • The feasibility of a low complexity technology with enhanced subjective (and objective) quality appears to have been demonstrated
  • Higher degrees of sophisticated inference processing on the decoder side (e.g., displacement estimation within the decoding process) are evident in various proposals relative to the designs in prior standards, and understanding the complexity/performance tradeoffs for such features seems particularly important

Test Model under Consideration (TMuC)

After reviewing the proposals and test results, JCT-VC tried to transmit from the “competitive phase” of individual proposal development to the “collaborative phase” of working together on a single design.  A main issue was how to select a software testing platform, on which good core experiments can be conducted during the next period of meetings.

JCTVC-A033, received during the meeting, proposed an initial Test Model design having some features from the following 7 proposals.

  • JCTVC-A114 (from France Telecom, NTT, NTT DOCOMO, Panasonic and Technicolor)
  • JCTVC-A116 (from HHI)
  • JCTVC-A119 (“TENTM” from Tandberg, Ericsson, and Nokia)
  • JCTVC-A120 (from RIM)
  • JCTVC-A121 (from Qualcomm)
  • JCTVC-A124 (from Samsung, with BBC)
  • JCTVC-A125 (from BBC, with Samsung)

The suggested Test Model that was asserted to be able to provide a coding efficiency capability close to that of the best performing proposals in the CfP subjective testing and also a complexity point close to that of the lowest complexity submissions with good subjective testing results.

After further review of JCTVC-A033 by JCT-VC as a whole, a “Test model under consideration” (TMuC) design was selected by consensus, with the common understanding that this is not yet an initial version of the draft standard, as no thorough testing has been performed for such a possible combination of tools. Although it is not a formally-adopted complete test model, the TMuC is a coherent collection of design components that appear promising among those of the various proposals. The coding tools in the TMuC should be further tested to confirm their effectiveness when used together in a unified coding architecture.

An initial draft of TMuC was produced during the meeting as JCTVC-A205 draft 0. An ad-hoc group (AHG) was established to further improve the editorial quality in the interim period until the next meeting.

Tool Experiments (TEs) and Ad Hoc Groups (AHGs)

To create an optimized unified design, it is important to understand the performances and complexities of individual tool components and how they interact with each other. For the organization and planning of its future work, JCT-VC established four “Tool Experiments” (TEs) to evaluate specific proposed coding tools and nine “Ad Hoc Groups” (AHGs) to work on more general subject areas.

Four TEs:

Nine AHGs:

  • JCT-VC project management, chaired by G. J. Sullivan and J.-R. Ohm
  • Test Model under Consideration editing, chaired by K. McCann
  • Software development and TMuC software technical evaluation, chaired by F. Bossen
  • Intra prediction, chaired by C. Auyeung
  • Alternative transforms, chaired by R. Cohen and R. Joshi
  • MV precision, chaired by  B. Jeon
  • In-loop filtering, chaired by T. Yamakage
  • Large block structures, chaired by K. Panusopone
  • Parallel entropy coding, chaired by M. Budagavi and A. Segall

Informal Naming

JCT-VC agreed to use “High Efficiency Video Coding” (HEVC) as the informal name of the new standardization initiative. No conclusion was reached on the topic whether HEVC should be a new standard or an extension of H.264/ AVC.

Other Information

JCT-VC reflector: jct-vc@lists.rwth-aachen.de
Subscription to this mailing list: http://mailman.rwth-aachen.de/mailman/listinfo/jct-vc
Download JCT-VC documents: http://ftp3.itu.ch/av-arch/jctvc-site/

Reference

“Meeting report of the first meeting of the Joint Collaborative Team on Video Coding (JCT-VC), Dresden, DE, 15-23 April, 2010,” JCT-VC document JCTVC-A200, Apr. 2010.

Permanent Link: The First JCT-VC Meeting, Dresden, DE

Post Comment

*
To prove you're a person (not a spam script), type the security word shown in the picture.
Anti-spam image