Back to Search Document details
40th Meeting: Geneva, CH, October 2025 2025-10-03 14:12
JVET AHG report: Test model software development (AHG3)
Abstract
This report summarizes the activities of the AhG3 on Test model software development that has taken place between the 39th and 40th JVET meetings.
JVET-AN0003 JVET AHG report: Test model software development (AHG3) [F. Bossen, X. Li, K. Sühring (co-chairs), E. François, Y. He, K. Sharman, V. Seregin, A. Tourapis (vice chairs)]

The software model versions prior to the start of the meeting were:

Software for MFC and MFCD is only available as published by ITU-T and ISO/IEC. It is planned to create repositories with the latest versions available in ITU-T H.264.2 (02/2016). All development history is lost.

  1. Software development

Development was continued on the GitLab server, which allows participants to register accounts and use a distributed development workflow based on git.

The server is located at:

https://vcgit.hhi.fraunhofer.de

The registration and development workflow are documented at:

https://vcgit.hhi.fraunhofer.de/jvet/VVCSoftware_VTM/wikis/VVC-Software-Development-Workflow

Although the development process is described in the context of the VTM software, it can be applied to all other software projects hosted on the GitLab server as well.

Between the 31st and 32nd meetings, the GitLab server was affected by a huge number of spam issues, which were generated using newly registered accounts. Administrator user account confirmation was enabled to protect the server from spam postings. Proponents should plan for some time for server admins to confirm new accounts.

Between the 37th and 38th JVET meeting extensive renewal work was performed on the build server infrastructure. To finish these works, changes to the build configurations are required for all software projects using build pipelines. The necessary information will be communicated to the software coordinators when available.

  1. VTM related activities

The VTM software can be found at

https://vcgit.hhi.fraunhofer.de/jvet/VVCSoftware_VTM/

The software development continued on the GitLab server. VTM version 23.11 was tagged on July. 4, 2025. VTM 23.12 is expected during the 40th JVET meeting. VTM 23.11 was not followed up with a macro cleanup version. Macros will be cleaned up after VTM 23.12.

VTM 23.11 was tagged on July 4, 2025. Changes include:

  • JVET-AJ0183: Implementation of syntax element eoi_src_pic_flag in Encoder Optimization Information SEI message
  • JVET-AK0287: Signal flags in DSCI SEI to indicate substream dependency
  • JVET-AL0123/AL0310: Signalling changes to the EOI SEI message
  • JVET-AK0140: Packed Regions Info SEI
  • JVET-AL0301: Specify constraints on syntax elements related to matrix_coeffs
  • JVET-AJ0060: multilayer adaptation of Packed regions information SEI message
  • Introduce cfg parameter to remove dependency on GOP size when determining `pps_init_qp_minus26`
  • JVET-AL0324, AL0070: updates of PRI SEI message
  • JVET-AL0339: Spatial Resolution for the FGC SEI message
  • JVET-AL0075: Add nnpfa_selected_input_flag to NNPFA SEI
  • JVET-AK0326 and JVET-AI0070: Updated NNPF SEI messages
  • JVET-AL0117: On association of NAL units to DSC verification substreams
  • Faster temporal prefilter
  • Faster CRC32C computation
  • JVET-AL0056: Constraint for quant threshold delta in EOI SEI message
  • JVET-AL0067: OMI SEI constraints
  • Improved implementation of hash search
  • Fix incorrect constant in ISP search
  • WG03 N01464: Implementation of attenuation maps as new green metadata type
  • JVET-AK0114: Implementation of AI-Usage Restrictions SEI message
  • JVET-AJ0105: Complexity Info in SPO SEI message
  • JVET-AM0280: Decoder memory printout
  • Update for high performance cfg files
  • add self-contained cfg for CfE
  • JVET-AM0200: reduced run time configurations
  • Update version number to 23.11

VTM 23.12 was expected to be tagged during the 40th JVET meeting. Changes were expected to include:

  • Fix memory leak when RPR is used
  • Fix a link error of parcat when JVET_J0090_MEMORY_BANDWITH_MEASURE is on
  • Initialize pointer to VPS of a pilot Slice object
  • Correctly set input chroma format when reading a y4m file
  • Avoid recomputing LMCS parameters when using debugBitstream by copying APS info (same as AFL APSs)
  • Fix trace output of coded block flag (CBF)
  • JVET-AM0227: Fix CABAC statistics collection
  • JVET-AJ0258: Image Format Metadata SEI
  • New enum for splits - removes magic numbers from CtxSplit
  • Increase minimum required CMake version
  • DSC: fix detection of last picture
  • JVET-AK0206: Digitally signed content ID
  • remove dead code in EncModeCtrl.h
  • DSC: Align DSCI syntax with spec
  • JVET-AL0222: Digitally signed content intended start and end flags
  • JVET-AL0066: Add sample tolerance value to OMI SEI
  • JVET-AM0164: move dsci_key_source_uri and increase length of dscv_signature_length_in_octets_minus1
  • JVET-AL0058: Change aur_context to fixed length
  • JVET-AM0334: Update to GFV SEI chroma key syntax
  • JVET-AK0281: AUR SEI message in SPO SEI message
  • JVET-AM0117: aur_sei_exclusion_flag

CTC Performance

The VTM 23.11 shows some improvements compared to VTM 23.9/23.10 due to MR2840, which enables DeblockingFilterOffsetInPPS, and faster code with reduced memory usage due to improvements in temporal prefilter and hash ME.

All Intra Main 10

Over VTM-23.9

Y

U

V

EncT

DecT

Class A1

0.00%

0.00%

0.00%

97%

100%

Class A2

0.00%

0.00%

0.00%

99%

99%

Class B

0.00%

0.00%

0.00%

99%

98%

Class C

0.00%

0.00%

0.00%

99%

98%

Class E

0.00%

0.00%

0.00%

97%

97%

Overall

0.00%

0.00%

0.00%

99%

98%

Class D

0.00%

0.00%

0.00%

98%

90%

Class F

0.00%

0.00%

0.00%

100%

100%

Class TGM

0.00%

0.00%

0.00%

98%

97%

Random Access Main 10

Over VTM-23.9

Y

U

V

EncT

DecT

Class A1

0.00%

0.00%

0.00%

98%

102%

Class A2

-0.01%

-0.01%

-0.01%

98%

103%

Class B

-0.01%

-0.01%

-0.01%

97%

99%

Class C

-0.02%

-0.02%

-0.02%

96%

95%

Class E

Overall

-0.01%

-0.01%

-0.01%

97%

99%

Class D

-0.06%

-0.07%

-0.07%

97%

95%

Class F

-0.03%

-0.03%

-0.03%

96%

95%

Class TGM

-0.01%

-0.01%

-0.01%

99%

99%

Low delay B Main 10

Over VTM-23.9

Y

U

V

EncT

DecT

Class A1

Class A2

Class B

-0.01%

-0.01%

-0.01%

98%

99%

Class C

-0.02%

-0.02%

-0.02%

98%

100%

Class E

-0.11%

-0.11%

-0.10%

95%

104%

Overall

-0.04%

-0.04%

-0.04%

97%

100%

Class D

-0.06%

-0.05%

-0.05%

97%

100%

Class F

-0.04%

-0.04%

-0.04%

96%

101%

Class TGM

-0.01%

-0.01%

-0.01%

97%

98%

Low delay P Main 10

Over VTM-23.9

Y

U

V

EncT

DecT

Class A1

Class A2

Class B

-0.01%

-0.01%

-0.01%

96%

97%

Class C

-0.01%

-0.01%

-0.01%

98%

103%

Class E

-0.09%

-0.09%

-0.08%

94%

101%

Overall

-0.03%

-0.03%

-0.03%

96%

100%

Class D

-0.04%

-0.03%

-0.03%

97%

95%

Class F

-0.02%

-0.03%

-0.03%

96%

98%

Class TGM

-0.01%

-0.01%

-0.01%

95%

94%

For the HDR CTCs, coding performance of VTM 23.11 compared to VTM 23.10 are reported in the table below. Slight variations are observed in RA. Encoding run time seems to be slightly reduced. Decoding run time is almost not changed.

Random Access

Over VTM23.10

wPSNR

PSNR

DE100

PSNR-L100

Y

U

V

Y

U

V

EncT

DecT

Class H1

-0.03%

-0.03%

-0.03%

-0.03%

-0.03%

-0.03%

-0.03%

-0.03%

97%

97%

Class H2

-0.01%

-0.01%

-0.01%

96%

101%

Overall

-0.03%

-0.03%

-0.03%

-0.03%

-0.03%

-0.02%

-0.02%

-0.02%

96%

99%

All Intra

Over VTM23.10

wPSNR

PSNR

DE100

PSNR-L100

Y

U

V

Y

U

V

EncT

DecT

Class H1

0.00%

0.00%

0.00%

0.00%

0.00%

0.00%

0.00%

0.00%

98%

101%

Class H2

0.00%

0.00%

0.00%

98%

102%

Overall

0.00%

0.00%

0.00%

0.00%

0.00%

0.00%

0.00%

0.00%

98%

101%

Coding performance of VTM 23.11 compared to HM 18.0 are reported in the table below.

Random Access

Over HM18.0

wPSNR

PSNR

DE100

PSNR-L100

Y

U

V

Y

U

V

EncT

DecT

Class H1

-39.49%

-37.66%

-36.75%

-54.96%

-48.52%

-33.76%

-49.45%

-40.94%

255%

79%

Class H2

-31.99%

-57.73%

-63.35%

220%

79%

Overall

-39.49%

-37.66%

-36.75%

-54.96%

-48.52%

-33.12%

-52.46%

-49.09%

241%

79%

All Intra

Over HM18.0

wPSNR

PSNR

DE100

PSNR-L100

Y

U

V

Y

U

V

EncT

DecT

Class H1

-41.44%

-27.31%

-26.79%

-57.73%

-52.77%

-24.00%

-52.92%

-45.07%

1511%

103%

Class H2

-21.76%

-47.23%

-50.60%

1225%

96%

Overall

-41.44%

-27.31%

-26.79%

-57.73%

-52.77%

-23.18%

-50.85%

-47.08%

1400%

100%

Issues in VTM affecting conformance

The following issues in the VTM master branch may affect conformance:

  • Missing HLS features (see sections below)

Several merge requests for proposals from the 36th or 37th JVET meeting, mostly related to SEI messages in VSEI are still pending review and/or waiting for resolution of comments and rebasing.

Status of implementation of proposals of previous JVET meetings

The following list contains all adoptions of the Q and R meetings that were not marked as merged (or submitted) or specification only change in the software coordinator tracking sheet:

SEI TuC software

Per the decision during the 32nd JVET meeting, an SEI TuC repository was created based on VTM-22.2. The repository is located at:

https://vcgit.hhi.fraunhofer.de/jvet-tuc/VVCSoftware_VTM

It should be noted that some of the technologies for which merge requests were submitted to TuC, got moved into the working draft. The merge requests were closed, when appropriate merge requests for VTM were available.

VTM-22.2-TuC-5.0 is expected to be tagged during the 40th JVET meeting.

The following changes were merged into the TuC branch:

The following merge requests were submitted and are pending:

  1. HM related activities

There was no new HM version tagged during this meeting cycle.

The following MRs were merged:

  • JVET-AE0101: implement phase indication SEI message
  • Update auto-build from VTM build definition
  • Update copyright headers to include 2025
  • JVET-AK0107: Modality Information SEI
  • JVET-AK0194: Digitally Signed Content SEI messages
  • Fix building on ARM
  • Enable build on macOS/arm
  • Modify semantics of IntraPeriod to enable autosetting based on frame rate
  • Fix yuv output when bitdepth is changed between sequence
  • Add missing initialization in TComPPS and TComScalingList (fix asan/msan errors)

The following MRs are pending [with status indicated]:

  • JVET-AL0059: Packed regions information SEI (pending reviewing)
  • JVET-AL0148: implementation of GFV and GFVE SEI messages (pending review)
  • Avoid reading SEI content when payloadSize = 0 (one issue remains)
  • JVET-AL0062: AI usage restrictions SEI message (pending review)
  • JVET-AK2006: Add support for SPTI SEI message (pending review/rebasing)
  • JVET-AL0061: Import encoder optimization information SEI Message form VTM (pending review)
  • JVET-AL0148: implementation of GFV and GFVE SEI messages (pending review)
  • Port the Y4M support [one issue remains]
  • Mark the current picture as short-term ref (for SCM) [need SCC expert reviewer]

A new version of HM is expected to be tagged during or after the 40th JVET meeting.

The HM SCC (SCM) branch (HM-16.21+SCM-8.8) has not been updated for the recent HM versions. Updating SCM to, for example, HM-18.0+SCM-8.8 should be considered. It may though be helpful to move SCC related functionality into separate source files. Volunteer work towards merging the branches would be appreciated.

As reported in the previous reports, further information on lambda optimization in HM would be appreciated, including comparison of allocation of bits within the GOP structures between HM and VTM.

Otherwise the HEVC bug tracker lists:

  • 44 tickets for “HM”, (most of which were created before 2019; 1 created in last year),
  • 1 ticket for “HM RExt” (created in 2020),
  • 9 tickets for “HM SCC” (most of which were created before 2018),
  • 1 ticket for “RExt Text” (created in 2015),
  • 1 ticket for “SCC Text” (created in 2016),
  • 8 tickets for text (6 created before 2021; 2 created in the last year),
  • 2 tickets for encoder description (created in 2020 and 2014)

Help to address these tickets would be appreciated.

  1. 360Lib related activities

The latest 360Lib software can be found at https://vcgit.hhi.fraunhofer.de/jvet/360lib/-/tags/360Lib-13.6

The following table is for the projection formats comparison using VTM-23.0 according to 360-degree video CTC (JVET-U2012) compared to that using VTM-22.0 (VTM-22.0 as anchor).

End-to-end WS-PSNR

End-to-end S-PSNR-NN

Y

U

V

Y

U

V

Class S1

0.00%

-0.01%

0.05%

0.00%

0.00%

0.05%

Class S2

0.08%

0.16%

0.34%

0.08%

0.15%

0.34%

Overall

0.03%

0.06%

0.17%

0.03%

0.06%

0.16%

The following table compares generalized cubemap (GCMP) coding and padded equi-rectangular projection (PERP) coding using VTM-23.0 (PERP as anchor).

End-to-end WS-PSNR

End-to-end S-PSNR-NN

Y

U

V

Y

U

V

Class S1

-11.52%

-5.47%

-6.09%

-11.50%

-5.41%

-6.03%

Class S2

-3.65%

0.98%

1.48%

-3.62%

1.08%

1.56%

Overall

-8.37%

-2.89%

-3.06%

-8.35%

-2.82%

-3.00%

The following tables are for PERP and GCMP coding comparison between VTM-23.0 and HM-16.22 (HM as anchor), respectively.

End-to-end WS-PSNR

End-to-end S-PSNR-NN

Y

U

V

Y

U

V

Class S1

-30.91%

-38.82%

-41.24%

-30.91%

-38.88%

-41.24%

Class S2

-36.84%

-37.23%

-39.43%

-36.82%

-37.27%

-39.49%

Overall

-33.28%

-38.19%

-40.51%

-33.28%

-38.23%

-40.54%

End-to-end WS-PSNR

End-to-end S-PSNR-NN

Y

U

V

Y

U

V

Class S1

-34.64%

-40.04%

-42.14%

-34.59%

-40.02%

-42.09%

Class S2

-38.34%

-38.18%

-40.06%

-38.32%

-38.19%

-40.11%

Overall

-36.12%

-39.30%

-41.31%

-36.08%

-39.29%

-41.30%

  1. SCM related activities

There had not been any further developments to SCC’s SCM during this meeting cycle.

  1. SHM related activities

There had not been any further developments to SHVC’s SHM during this meeting cycle.

  1. HTM related activities

There had not been any releases of HTM of MV-HEVC and 3D-HEVC.

One merge request is was merged:

  • JVET-AE0295: MV Main 10 profile support (waiting for proponent response)

The next release will also include the following changes:

  • JVET-Z0209: Early termination during calculating RDcost of depth
  1. HDRTools related activities

Two new versions (0.25 and 0.26) of HDRTools have been tagged.

Version 0.25 includes the following changes:

  • Update HDRMontage to support temporal interleaving
  • Fixes to TIFF files with Alpha
  • support for absolute crop
  • Compute metrics with files of different bitdepth

Version 0.26 includes the following changes:

  • fixes computation of WPSNR metric to match VTM software when EnableJVETPSNR=1
  • fixes build on recent gcc versions (13 and above)
  1. JM, JSVM, JMVM related activities

There had not been any further developments to JSVM or JMVM during this meeting cycle.

There was no new JM version tagged during this meeting cycle, although merge requests were submitted for SEI messages in VSEI.

The following MRs are pending [with status indicated]:

  • JVET-AK0107: Modality Information SEI (pending review)
  • JVET-AK2006: NNPFC and NNPFA SEI message (pending review)
  • JVET-AL0062: AI usage restrictions SEI message (pending review)
  • JVET-AK2006: SPTI SEI Message (pending review)
  • JVET-AL0061: Encoder Optimization Information SEI Message (pending review)
  • JVET-AE0101: implement phase indication SEI message (pending review)
  • JVET-AL0148: Add support for Generative Face Video (GFV) and Generative Face Video Enhancement (GFVE) SEI messages(pending review)

For the previous release of JM only basic testing was performed to confirm that the code still works, but performance testing is pending. Performance changes are unlikely but may occur due to bug fixes. A possible encoder/decoder mismatch was reported privately, but needs more investigation.

  1. Bug tracking

The bug tracker for VTM and specification text is located at:

https://jvet.hhi.fraunhofer.de/trac/vvc

The bug tracker uses the same accounts as the HM software bug tracker. Users may need to log in again due to the different sub-domain.

https://hevc.hhi.fraunhofer.de/trac/hevc

New account creation was disabled due to a large amount of spam registration. Extensive maintenance of the Trac bug tracker would be required to keep it secure. It should be considered to move bug tracking to GitLab, although no automatic conversion of existing bugs is available.

Bug tracking for HDRTools is located at:

https://gitlab.com/standards/HDRTools/-/issues

Contributors were requested to please file all issues related to the VVC reference software and HDRTools into the appropriate bug tracker. Try to provide all the details, which are necessary to reproduce the issue. Patches for solving issues and improving the software are always appreciated.

  1. CTC alignment and merging

There are currently 8 JVET CTC documents:

Further merging of HM RExt CTC into the appropriate VVC CTC was investigated (non 4:2:0 chroma formats), but proper comparable HM configuration files were not yet available by the beginning of this meeting.

The HM HDR test conditions require additional post-processing stages, due to WPSNR, deltaE100 and PSNRL100 values not being generated directly by the HM encoder. Merging of respective functions from VTM is recommended to streamline the process.

  1. Guidelines for reference software development

Current versions of the software guidelines are:

  • JVET-AC1001 Guidelines for HM-based software development [K. Sühring, F. Bossen, X. Li (software coordinators)]
  • JVET-AJ2003 Guidelines for VTM-based software development [F. Bossen, X. Li, K. Sühring]
  1. Related proposals

JVET-AN0292 proposes to add multi-view functionality to HM. Mutli-view HEVC is usually supported by HTM. It is asserted that range extension support in HTM is very limited but required for the new multi-view profiles. It is asserted that it was easier to port multi-view functionality to HM instead if range extension to HTM.

  1. Recommendations

The AHG recommends to:

  • Continue to develop reference software.
  • Improve documentation, especially the software manual.
  • Encourage people to test VTM and other reference software more extensively outside of common test conditions.
  • Encourage people to report all (potential) bugs that they are finding.
  • Encourage people to submit bitstreams/test cases that trigger bugs in VTM and other reference software.
  • Encourage people to submit non-normative changes that either reduce encoder run time without significantly sacrificing compression performance or improve compression performance without significantly increasing encoder run time.
  • Design and add configuration files to the VTM software for testing of HLS features.
  • Review VTM-related contributions and determine whether features should be added (or removed) from the software.
  • Continue to investigate the merging of branches.
  • Continue to investigate merging of CTC documents.
  • Verify correctness of CTC documents and issue updates as appropriate
  • Keep common test conditions aligned for the different standards.

Numerous merge requests for adding new SEI messages to JM and HM had not yet been integrated due to lack of time of software coordinators. This should be used later as basis for issuing new versions of AVC and HEVC reference software.

Further discussion on JVET-AN0292 about harmonization of HM and HTM – see section 4.3. It was mentioned that also document JVET-AN0048 reports activities on harmonizing HM and SHM

Decisions
There are currently 8 JVET CTC documents:
Citation