Joint Collaborative Team on Video Coding (jct-vc) Contribution

3Project development, status, and guidance (29)

PDAM ballot comments review and response.

3.1Communication to and by parent bodies (0)

See section 7.1.

3.2Conformance test set development (2)

JCTVC-P0085 Conformance testing [C. Fogg (Harmonic), A. Wells, D. Le Gall (Ambarella)]

Conformance bitstreams testing some high-level aspects including HRD, are largely absent from the current test suite. It is proposed to add key combinations that will be relied upon in systems. The contributor asserted that bitstreams demonstrating correlation among the display timing elements (pic_struct, vui_time_scale, etc.) and bitstream splice users (time code, HRD) could assist implementers in the use of these elements.

For repeated pictures, the contribution suggested that the reconstructed .yuv file output by the reference decoder would contain the repeated pictures.

It was remarked that the "repeated picture" concept is more a display issue than a decoder output conformance issue, so it may not be desirable for the reference software to modify its output file content based on that, except perhaps in some special mode of decoder operation. Repetition is not a required output behaviour, and thus should not be tested for conformance.

It was agreed that having additional conformance bitstreams testing high-layer syntax (incl. SEI, VUI, parameter sets, HRD, etc.) would be highly desirable (and that the HM reference software may also need work in these areas).

It was asked whether the HM software parses all SEI messages that have been specified.

JCTVC-P0276 Editor's proposed draft text of HEVC conformance testing [T. Suzuki, W. Wan, G. Sullivan] [late]

Latest update from the editors. Detailed review was not requested.

3.3Version 1 bug reports and cleanup (2)

See also notes for P0187 (no_output_of_prior_pics_flag) and P0043 (AU boundary detection) and P0044 (MinCR) and P0130 (end of bitstream).

LS input from ITU-R to the parent body is relevant to the VUI. Decision: Correct the equations accordingly.

JCTVC-P0064 Defect correction for HEVC version 1 (esp. modulo CPB removal delay wrapping) [G. J. Sullivan, L. Zhu (Microsoft), Y.-K. Wang, A. K. Ramasubramonian (Qualcomm)]

Discussion chaired by M. Hannuksela 01-16.

This contribution proposes an alternative to the approach previously proposed in JCTVC-N0094 to handling modulo wrapping for interpretation of the value of the syntax element au_cpb_removal_delay_minus1 for pictures that begin new buffering periods.

The prior planned fix for the reported problem included introducing a syntax change to version 1 bitstreams. This contribution suggests that the syntax change is unnecessary and that an alternative interpretation of how to compute CPB removal delays is actually what was intended when the version 1 syntax (and the similar syntax in AVC) was originally designed. An alternative approach to the reported problem is proposed that does not include this syntax change.

The contribution also suggests follow-up action to initiate a clarification of the same aspect of the AVC specification.

A couple of other editorial corrections for HEVC version 1 are also reported. These are asserted to be obvious corrections and merely typographical in nature.

It was requested the update HM according to the proposal.

Two other very small errata reports were included in the contribution.

Decision: Adopt.

Parent bodies may wish to make note of the AVC action request.

JCTVC-P0223 JPNB comment on HEVC defect report [Japan National Body]

A defect report of HEVC (23008-2) was produced at Geneva meeting (WG 11 document N 13931). It includes the syntax change of HRD parameters. Although the number of fixes is small, it was asserted to be desirable to produce such correction in formal way to avoid market confusion. JPNB requested WG11 to issue a corrigendum of 23008-2 as soon as possible.

After discussion, it was agreed that finalizing the approval of the corrections ASAP is desirable, and that rather than issuing a separate corrigendum, it was planned to integrate the corrections into near-term amendments.

3.4HEVC coding performance, implementation demonstrations and design analysis (10)

3.4.1General

JCTVC-P0302 BoG report on performance analysis and demos [J. Samuelsson]

A break out group discussion was held on Sunday the 12th of January 2014 to review the documents assigned to categories 3.4.2–3.4.5 in version 2 of the draft JCT-VC meeting report. P0261 was noted to be a technical change proposal, so it was not reviewed in the BoG.

Four contributions were reviewed in the BoG.

• 
  P0169 reported on performance benefits obtained using a parallel-processing implementation for SHVC decoding.
  
• 
  P0082 was tutorial information about using the pic_struct syntax coding of interlaced
  
• 
  P0083 contained an analysis comparison of different techniques for coding interlaced material with HEVC (version 1). It contained a comparison against picture-adaptive frame/field coding (which is something not supported in version 1) and sequence-adaptive frame/field coding (which is supported in version 1).
  
• 
  P0158 reported informal subjective video quality comparison results for HEVC versus AVC implemenation.
  

3.4.2Version 1 verification test (2)

See also P0042 (on source content availability).

JCTVC-P0054 Updates to HEVC verification test plan draft 2 [TK Tan, Vittorio Baroncini, Marta Mrak]

Discussion 01-16.

This document contains updates to the JCTVC-O1011 document. The schedule and test sites have been updated and additional test sequences and proposed encoding bitrates for use in the verification test were added. The copyright statements for the sequences used have been clarified and listed in Annex B.

Some aspects incorporated into P1011 after discussion included the following:

• 
  Subjective testing was expected to be completed by 3 March
  
• 
  Test sites were identified as follows (further detail in P1011):
  
  • 
    BBC
    
  • 
    FUB
    
  • 
    EPFL
    
• 
  HM 12.1 was planned to be used for HM.
  
• 
  5 sequences were recommended for each of the 4 resolutions to be tested
  

Adjustments as necessary may be needed for 4K content due to limitations of 4K playout.

3.4.3RExt performance and design aspects (3)

JCTVC-P0200AhG8: Performance and complexity analysis of Range Extensions tools [J. Sole, T. Hsieh, C. Pang, M. Karczewicz (Qualcomm)]

See BoG report P0288 and related notes.

JCTVC-P0201 AhG8: Performance of encoder and parameter only changes for Screen Content Coding [J. Sole, C. Pang, L. Zhang, M. Karczewicz (Qualcomm)]

See BoG report P0288.

JCTVC-P0213 Comparison of Compression Performance of HEVC Range Extensions Draft 5 with AVC High 4:4:4 Predictive profile [B. Li, J. Xu, G. J. Sullivan (Microsoft)]

See BoG report P0288 and related notes.

3.4.4SHVC performance and design aspects (1)

JCTVC-P0169 Update of JCTVC-O0115 on a pipeline and parallel architecture of an SHVC decoder [W. Hamidouche, M. Raulet (IETR/INSA)]

See BoG report P0302 and related notes.

3.4.5Interlace (3)

JCTVC-P0082 How to use pic_struct [C. Fogg, O. Bar-Nir, P. Haskell (Harmonic), A. Wells, D. Le Gall (Ambarella)]

See BoG report P0302 and related notes.

JCTVC-P0083 pic_struct coding results [C. Fogg (Harmonic)]

See BoG report P0302 and related notes. Also see related contribution P0261.

3.4.6Implementation demonstrations (1)

JCTVC-P0158 Informal Subjective Video Quality Comparison Between the eBrisk-UHD HEVC and x264 AVC Encoders [M. Horowitz, F. Kossentini, H. Tmar (eBrisk)

See BoG report P0302 and related notes.