International organisation for standardisation organisation internationale de normalisation


Particular additional syntax features that would be beneficial to have



Yüklə 8,2 Mb.
səhifə109/277
tarix02.01.2022
ölçüsü8,2 Mb.
#13030
1   ...   105   106   107   108   109   110   111   112   ...   277
Particular additional syntax features that would be beneficial to have

Syntax element

Case to be tested

Volunteers

sps_video_parameter_set_id

Value other than 0

J. Boyce

conf_win_bottom_offset

conf_win_top_offset

conf_win_left_offset

conf_win_right_offset



Values other than 0, preferably not multiples of 4

G. Sullivan ?

num_short_term_ref_pic_sets

Maximum value 64

R. Sjöberg

num_long_term_ref_pic_sps

Value other than 0

Y.K. Wang

sub_layer_profile_present_flag

sub_layer_level_present_flag



Values other than 0

J. Boyce

use_delta_flag

Value other than 0

T. K. Tan

init_qp_minus26

Value less than 0

T. Suzuki

nal_unit_type

Values 38 (filler data)

G. Sullivan ?


Additional items noted in review

pic_output_flag

Value other than 1

Y.-K. Wang

There was a question on the tier of each conformance bitstreams. The tier should be clarified in the conformance spec.

P0085 and P0276 were noted to be related.

The AHG recommended discussing the plan to collect missing bitstreams and to continue to collect more bitstreams, especially corner case bitstreams.


14.1.97.1.1.1.1.1.75JCTVC-P0005 JCT-VC AHG report: HEVC range extensions development (AHG5) [C. Rosewarne, M. Naccari]

This report summarizes the activities of the HEVC range extensions development Ad-Hoc Group (AHG5) between the 15th JCT-VC meeting in Geneva, CH (October-November 2013) and the 15th JCT-VC meeting in San José, CA, USA (January 2014).

After a kick-off message issued on the 18th of November, no further emails were sent to the reflector regarding AHG5.

The ‘HEVC Range extensions test model 1 encoder description’ (JCTVC-O1013) was prepared and uploaded to the document repository.

Contribution JCTVC-O1006 was uploaded, defining the test conditions as agreed at the 15th JCT-VC meeting for Range extensions development. As discussed at the 15th JCT-VC meeting, the number of encoded frames was reduced for several of the test sequences. This contribution also includes the spread sheet templates to be used in RExt Core Experiments (RCEs) at this meeting.

Relevant contributions were listed according to subject areas:



  • RCE contributions (see RCE summary reports)

  • Profiles & levels (6)

  • High-level syntax and auxiliary pictures (8)

  • High dynamic range (3)

  • Cross-component decorrelation (4)

  • Miscellaneous (5)

14.1.97.1.1.1.1.1.76JCTVC-P0006 JCT-VC AHG report: Range extensions draft text (AHG6) [J. Sole, D. Flynn, C. Rosewarne, G. Sullivan, T. Suzuki]

This document reports on the work of the JCT-VC ad hoc group on Range extensions draft text (AHG6) between the 15th JCT-VC meeting in Geneva, CH (October-November 2013) and the 16th JCT-VC meeting in San Jose, US (January 2014).

The High Efficiency Video Coding (HEVC) Range Extensions test model was developed following the decisions taken at the 15h JCT-VC meeting in Geneva, CH (October-November 2013).

Four versions of JCTVC-O1005 were published by the Editing AhG following the 15th JCT-VC meeting in Geneva. Versions were based upon JCTVC-L1003_v34. The text of JCTVC-O1005 (revision 4) was submitted as WG11 N13933 Study text of ISO/IEC.

Changes in JCTVC-O1005 relative to the previous version were listed in the report.

JCTVC-P0007 JCT-VC AHG report: Range extensions software (AHG7) [K. Sharman, D. Flynn]

This report summarizes the activities of Ad Hoc Group 7 on support for range extensions between the 15th and 16th JCT-VC meetings.

The HM12.0_RExt4.1 software was upgraded to HM12.0_RExt4.2 (macro removal and code tidy) and then to HM12.1_RExt4.2 (reflecting the changes made to the main HM branch), with both revisions released on Tuesday November 19th. All were tested according to the HM and RExt4.0 test conditions; the latter was also tested using the latest RExt5 test conditions.

On Tuesday November 19th, the development code for RExt5.0 (RExt5.0_rc1) with changes for all tools was announced, and a request was made for proponents to verify their tools.

During the integration process, it was noticed that one of the tools changed the encoder search algorithm. This would mean that the RExt model would no longer be backwards compatible with the general HM model. The coordinators requested a scheme from the proponents, with the provided scheme being a duplication of one of the longer inter-encoder-search functions. To try to avoid code replication, the coordinators quickly made changes so that only one search function was used and that it was compatible with HM. The version in rc1 was backwards compatible with HM, however, there was a mistake when cross-component-decorrelation was used in the inter search.

To expedite the process of release of RExt5.0, the coordinators fixed the problem, but also used the duplicated search algorithm provided by the proponents (controlled by a macro). It is this second search algorithm that is being used in standard RExt test conditions; the original search is used in HM test conditions.

RExt5.0 was released on Tuesday December 3rd, and version RExt5.1 (with many of the new tool macros removed and a code tidy) was released on Thursday December 5th.

It was noted that there is an input contribution P0059 to address the search situation.

The changes adopted to this RExt5 were due to proposals listed in the report.

The naming of inter-component decorrelation / cross-component prediction was discussed, and it was agreed to suggest the term "cross-component prediction" to the editors.



Some items were identified as still to be integrated:

  • O0044 – CU-adaptive chroma QP offsets (highest priority – this item is normative)

  • O0043 – Best-effort decoding

  • N0383 – Motion constrained tile sets SEI

  • O0079 – Chroma sampling filter hint SEI (some HM patch software was in the contribution )

  • O0099 – Time code SEI (patch against HM5.1 provided by proponents)

Encoding performance differences were reported


Yüklə 8,2 Mb.

Dostları ilə paylaş:
1   ...   105   106   107   108   109   110   111   112   ...   277




Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©muhaz.org 2024
rəhbərliyinə müraciət

gir | qeydiyyatdan keç
    Ana səhifə


yükləyin