JVET-AM0334 AHG9: Common specification text for GFV SEI message [J. Boyce, M. M. Hannuksela (Nokia), S. McCarthy, P. Yin, G. J. Sullivan (Dolby), J. Chen, Y. Ye, B. Chen (Alibaba)] [late]
General discussion
It is noted that there is a ballot comment in AVC that SEI message names should be uniquely searchable. GFV and GFVE are not unique. It is suggested to change the name of the GFVE SEI message and its syntax elements. This can be left to the editors or contributions to the next meeting can make suggestions.
Discussed on 1 July 2025 chaired by S. Deshpande during 11:55-12:10
This contribution proposes common specification text for the generative face video SEI message. The proposed modifications are intended for VSEI v4.
The attached specification text incorporates editorial bug fixes and improvements for clarity, persistence, and constraints. The text also specifies syntax element, gfv_chroma_key_purpose_idc.
It was suggested that note 3 conditions on chroma key purpose idc should be delegated to the editors.
Base picture, driving picture and fusion pictures are mutually exclusive (i.e. a picture can be only of one of those types).
The syntax and semantics were reviewed. Some further editorial changes may be necessary.
Decision: Adopt to VSEI v4.
Other (1)
Contributions in this area were discussed during 0900–1040 on Saturday 28 June 2025 (chaired by J. Boyce).
JVETAM0126 AHG9: Temporal quality indication for EOI SEI message [J. Chen, Y. Ye, B. Chen (Alibaba)]
It is proposed to add an indicator in the encoder optimization information (EOI) SEI message when the value of eoi_type indicates temporal quality optimization has been applied. The proposed indicator can be used to indicate how the quality of a picture associated with the SEI message has fluctuated due to temporal quality optimization. It is asserted that such information could be useful at the decoder side for some applications to improve temporal quality consistency.
In version 3, an additional option to simplify the proposed semantics is provided.
JVET-AL0101 was suggested to be related.It was noted that the quality metric SEI in TuC can signal per picture quality. The contribution is describing a purpose.
It was questioned why this contribution is being proposed for VSEI v4.
The proposed option 1 semantics relate to human viewing, for which there is an existing eoi_for_human_viewing_idc syntax element. Option 2 does not mention human viewing.
A GFV use case was mentioned, but it might apply to other use cases.
It was noted that the current EOI SEI message is aimed more at CLVS properties, although has a persistence flag. This contribution uses the persistence, so only needs to signalled when the picture quality changes.
It was suggested that a flag may be sufficient rather than indicator.
Contribution AM0113 is related but specifically addresses NNPF.
Decision: Add option 2 to TuC, but move the syntax into the extension. A revision to the document with the syntax in the extension to be uploaded.
AHG9: Extensions of SEI messages in VSEI (3)
Contributions in this area were discussed during 1100–1145 on Saturday 28 June 2025 (chaired by J. Boyce).
JVET-AM0334 AHG9: Common specification text for GFV SEI message
It is noted that the list above may not be complete; if some adoption is missing that is recorded somewhere else in the meeting notes it shall also be considered included.
Primary editor: J. Boyce.
Editors are requested to make preparations for integration into a new edition until the next meeting.
Remains valid – not updated: JVET-AJ2007 Guidelines for NNVC software development [F. Galpin, S. Eadie, L. Wang, Z. Xie, Y. Li]
Remains valid – not updated: JVET-X2008 Conformance testing for versatile video coding (Draft 7) [J. Boyce, F. Bossen, K. Kawamura, I. Moccagatta, W. Wan]
The document number was planned to be re-used for a 3rd edition in ITU , once that is submitted to ITU-T (could be in October 2025).
Remains valid – not updated: JVET-AJ2009 Reference software for versatile video coding 2nd edition (Draft 2) [F. Bossen, K. Sühring, X. Li] [WG 5 DIS N 322)]
Software relating to H.266.2 and ISO/IEC 23090-16 can be found at https://vcgit.hhi.fraunhofer.de/jvet/VVCSoftware_VTM/-/tree/2nd-edition.
A DoC WG 5 N 366 on the DIS was reviewed and approved on Thursday 3 July at 1640. The text was requested to be published as IS with editorial improvements as requested by ISO editor in the context of the ballot. Public availability was also requested.
Primary editor: F. Bossen.
Remains valid – not updated: JVET-AL2010 VTM and HM common test conditions and software reference configurations for SDR 4:2:0 10 bit video [F. Bossen, X. Li, V. Seregin, K. Sharman, K. Sühring] (2025-04-18)
Add alternative configuration from JVET-AL0055 and clarification about intra picture period (see discussion under JVET-AL0114). Also links to test sequences are updated due to the change of the content server.
Remains valid – not updated: JVET-AC2011 VTM and HM common test conditions and evaluation procedures for HDR/WCG video [A. Segall, E. François, W. Husak, S. Iwamura, D. Rusanovskyy]
Links to test sequences need to be updated due to the change of the content server.
Remains valid – not updated: JVET-U2012 JVET common test conditions and evaluation procedures for 360° video [Y. He, J. Boyce, K. Choi, J.-L. Lin]
Links to test sequences need to be updated due to the change of the content server.
Remains valid – not updated: JVET-T2013 VTM common test conditions and software reference configurations for non-4:2:0 colour formats [Y.-H. Chao, Y.-C. Sun, J. Xu, X. Xu]
Links to test sequences need to be updated due to the change of the content server.
Remains valid – not updated: JVET-Q2014 JVET common test conditions and software reference configurations for lossless, near lossless, and mixed lossy/lossless coding [T.-C. Ma, A. Nalci, T. Nguyen]
Links to test sequences need to be updated due to the change of the content server.
Remains valid – not updated: JVET-Q2015 JVET functionality confirmation test conditions for reference picture resampling [J. Luo, V. Seregin]
Links to test sequences need to be updated due to the change of the content server.
Remains valid – not updated: JVET-AJ2016 Common test conditions and evaluation procedures for neural network-based video coding technology [E. Alshina, F. Galpin, R.-L. Liao, S. Liu, A. Segall]
Remains valid – not updated: JVET-AI2017 Common test conditions and evaluation procedures for enhanced compression tool testing [M. Karczewicz, Y. Ye]
Links to test sequences need to be updated due to the change of the content server.
It was also commented that it might be good to add conditions relating to HDR, even though currently no HDR tests are conducted in CTC.
Remains valid – not updated: JVET-AA2018 Common test conditions for high bit depth and high bit rate video coding [A. Browne, T. Ikai, D. Rusanovskyy, X. Xiu, Y. Yu]
Links to test sequences need to be updated due to the change of the content server.
JVET-AM2019 Description of algorithms version 12 and software version 14 in neural network-based video coding (NNVC) [F. Galpin, Yue Li, Yun Li, D. Rusanovskyy, T. Shao, J. Ström, L. Wang] (2025-09-30)
New elements in text and software from notes elsewhere in this report:
Decision: Adopt JVET-AM0135 EE1-2.1.
Decision: Adopt JVET-AM0175 EE1-3.2 (non CTC, also provide training script).
It is noted that the list above may not be complete; if some adoption is missing that is recorded somewhere else in the meeting notes it shall also be considered included.
Remains valid – not updated: JVET-AK2020 Film grain synthesis technology for video applications ed. 2 (Draft 2) [W. Husak, P. de Lagrange, A. Norkin, A. Tourapis] (2025-03-14)
Note that draft 3 (JVET-AL2020, planned as output of 38th meeting) had not yet become available by the time of releasing this report.
JVET-AM2021 Verification test plan for VVC multilayer coding (update 6) [O. Chubach, P. de Lagrange, M. Wien] (2025-09-15)
Developed from JVET-AM0223.
HHI (B. Bross) expresses interest for participating in bitstream generation.
To be discontinued if no increase of activity is observed over the next two meeting cycles.
Remains valid – not updated: JVET-AJ2022 Plan for subjective quality testing of the FGC SEI message (update 4) [P. de Lagrange, W. Husak, M. Radosavljević, M. Wien]
Number might be re-used (activity discontinued) as there has been no activity for 3 meeting cycles.
JVET-AM2023 Exploration experiment on neural network-based video coding (EE1) [E. Alshina, R. Chang, F. Galpin, Yue Li, Yun Li, M. Santamaria, T. Shao, J. Ström, Z. Xie (EE coordinators)] (2025-07-25)
An initial draft of this document was reviewed and approved at 1125-1140 on Friday 4 July.
This round of EE1 tests includes:
EE1-1: LOP in-loop filter
EE1-1.1 – LOP with Overlapped Feature Integration (JVET-AM0131)
EE1-1.2 –Backbone Block Enhancement of LOP In-Loop Filter with Over-Parameterized Training and Variable Channels (JVET-AM0122)
EE1-1.3 – Conditional LOP loop-filter (JVET-AM0053)
EE1-2: VLOP in-loop filter
EE1-2.1 – Conditional VLOP loop-filter (JVET-AM0053)
EE1-2.2 – Decomposed Content-Adaptive VLOP (JVET-AM0185)
EE1-3: NN-inter prediction
EE1-3.1 – Deep Reference Frame Generation for Inter Prediction Enhancement with Structural Re-parameterization (JVET-AM0177)
EE1-3.2 – RA/LDB Reference Frame Synthesis for VVC Inter Coding (JVET-AM0115)
EE1-3.3 – Network architecture from EE1-3.1 with training from EE1-3.2
EE1-3.4 – Network architecture from EE1-3.2 with training from EE1-3.3
EE1-4: NN-based super-resolution
EE1-4.1 NNSR with new backbone block based on Spatial-Channel Mixing (SCM) (JVET-AM0199)
EE1-4.2 Cross-component enhanced NNSR (JVET-AM0257)
JVET-AM2024 Exploration experiment on enhanced compression beyond VVC capability (EE2) [V. Seregin, D. Bugdayci Sansli, J. Chen, R. Chernyak, K. Naser, J. Ström, F. Wang, M. Winken, X. Xiu, K. Zhang (EE coordinators)] (2025-08-01)
An initial draft of this document was reviewed and approved at 0920-0925 on Friday 4 July.
This round of EE2 tests will include:
Tests
Tester
Cross-checker
1 Partitioning
1.1a
Restrictions for TT splitting
G.Wang (vivo)
1.1b
Test 1.1a as encoder-only change
G.Wang (vivo)
1.1c
Temporal partitioning prediction optimization
G.Wang (vivo)
1.1d
Test 1.1a + Test 1.1c
G.Wang (vivo)
1.1e
Test 1.1b + Test 1.1c
G.Wang (vivo)
2 Intra prediction
2.1
TMRL blend
S. Blasi
(Nokia)
V. Rufitskiy (TCL)
2.2a
Alignment of interpolation filters used in reference samples unwrapping and predictor generation
V. Rufitskiy
(TCL)
2.2b
Disabling smoothing in interpolation and reference sample processing operations for DIMD and TIMD
V. Rufitskiy
(TCL)
2.3a
Unification of unwrapping interpolation filters in TIMD template and block prediction
T. Dong
(TCL)
2.3b
8-tap interpolation filter for template generation in TIMD
T. Dong
(TCL)
2.4
Longer tap interpolation filtering
V. Rufitskiy
(TCL)
2.5
Combination of Test 2.2, Test 2.3, and Test 2.4
V. Rufitskiy
(TCL)
2.6a
Adaptive subsampling filter selection for CCLM/intra-CCCM/ALF-CCCM
Y. Kidani
(KDDI)
2.6b
Adaptive subsampling filter selection for inter-CCCM
Y. Kidani
(KDDI)
2.6c
Test 2.6a + Test 2.6b
Y. Kidani
(KDDI)
2.7
Reducing candidate modes in decoder derived CCP
S. Wan(NWPU)
S. Xie(ZTE)
2.8
Enhanced CCP Fusion mode
P.Bordes (InterDigital)
3 Inter prediction
3.1
Generated merge candidates
D. Bugdayci Sansli
(Nokia)
3.2a
Joint reordering of GPM with affine prediction
L. Zhang
(OPPO)
3.2b
Modifications to the affine merge list
L. Zhang
(OPPO)
3.2c
Test 3.2a + Test 3.2b
L. Zhang
(OPPO)
3.3
Joint reordering of GPM with intra prediction
Z. Sun
(OPPO)
3.4
Test 3.2c + Test 3.3
Z. Sun
(OPPO)
4 Transform and coefficients coding
4.1
Linked sign prediction
Y. Zhang
(OPPO)
4.2a
Shifting quantization center for TSRC under non-CTC (RDOQ on)
Y. Yu
(OPPO)
4.2b
Shifting quantization center for RRC under non-CTC (DQ disabled, RDOQ on)
Y. Yu
(OPPO)
4.2c
Test 4.2a + Test 4.2b
Y. Yu
(OPPO)
4.2d
Shifting quantization center for TSRC under CTC
Y. Yu
(OPPO)
4.2e
Shifting quantization center for RRC under CTC
Y. Yu
(OPPO)
4.2f
Test 4.2d + Test 4.2e
Y. Yu
(OPPO)
4.2g
Shifting quantization center for RRC under non-CTC (DQ disabled, RDOQ disabled)
M. Le Pendu (InterDigital)
4.2h
Shifting quantization center for TSRC under non-CTC (DQ disabled, RDOQ disabled)
M. Le Pendu (InterDigital)
4.2i
Test 4.2g + Test 3.2h
M. Le Pendu (InterDigital)
4.3a
Residual sign prediction restriction
C. Hollmann
(TCL)
4.3b
RSP context modelling
C. Hollmann
(TCL)
4.3c
Test 4.3a with retraining of existing contexts
C. Hollmann
(TCL)
4.3d
Test 4.3a + Test 4.3b
C. Hollmann (TCL)
5 In-loop filtering
5.1a
Regularization of ALF-CCCM
Z. Xie
(OPPO)
5.1b
Regularization for ALF-CCCM enabled in I slice
Z. Xie
(OPPO)
5.2a
Non-downsampled ALF-CCCM
L. Xu
(OPPO)
5.2b
Reuse of CU partition in ALF-CCCM
L. Xu
(OPPO)
5.2c
Test 5.2a + Test 5.2b
L. Xu
(OPPO)
5.3a
Remove single model of ALF-CCCM
N. Song
(OPPO)
5.3b
Separate “bad window” condition for Cb and Cr components
N. Song
(OPPO)
5.3c
Test 5.3a + Test 5.3b
N. Song
(OPPO)
5.4
Additional models for ALF-CCCM
F. Wang
(OPPO)
5.5a
Test 5.1 + Test 5.2
L. Xu
Z. Xie
(OPPO)
5.5b
Test 5.1 + Test 5.2 + Test 5.4
L. Xu
Z. Xie
F. Wang
(OPPO)
5.5c
Test 5.1 + Test 5.2 + Test 5.3 + Test 5.4
N. Song
L. Xu
Z. Xie
F. Wang
(OPPO)
5.6
Reuse of TALF control information
Y. Bai
(ZTE)
JVET-AM2025 Algorithm description of Enhanced Compression Model 18 (ECM 18) [M. Coban, R.-L. Liao, K. Naser, J. Ström, L. Zhang] (2025-09-30)
New elements from notes elsewhere in this report:
Decision: Adopt JVET-AM0307 test 1.12.
Decision: Adopt JVET-AM0163 test 1.7a.
Decision: Adopt JVET-AM0157 test 1.8.
Decision: Adopt JVET-AM0106 test 2.1b.
Decision: Adopt JVET-AM0215 test 2.2a.
Decision: Adopt JVET-AM0056 Test 3.1.
Decision: Adopt JVET-AM0209 test 4.2.
Decision: Adopt JVET-AM0063 test 4.3c.
Decision: Adopt JVET-AM0231 (non-CTC) test 4.5c (non-CTC).
Decision: Adopt JVET-AM0104 on top of EE2-1.7a.
Decision (SW): add the missing implementation of JVET-AE0102 and JVET-AG0100 encoder operations to the next version of ECM.
It is noted that the list above may not be complete; if some adoption is missing that is recorded somewhere else in the meeting notes it shall also be considered included.
JVET-AM2026 Joint Call for Evidence on video compression with capability beyond VVC [J.-R. Ohm, M. Wien, F. Bossen] [WG 5 N 367] (2025-07-04)
Final version was reviewed in a BoG meeting Fri 4 July 0800-0915, where the substantial changes made were adjustments of bit rates and adding an annex for registration of submissions. The content of the document was approved by JVET at 0955 on Friday 4 July. The latest draft with all changes was uploaded as an attachment to the BoG report JVET-AM0336. Experts were asked to carefully review and inform the editors about possible need of changes until the end of the meeting.
Remains valid – not updated: JVET-AJ2027 Common test conditions for gaming applications [J. Sauer, R. Chernyak, S. Puri, S. Thiebaud]
Remains valid: JVET-AL2028 Additions and corrections for VVC conformance (draft 1) [S. Iwamura, P. de Lagrange, I. Moccagatta] (2025-05-02)
This is a delta change document, some aspects related to bitstream corrections and editorial notes to be removed. For DIS, WG 5 N 365 (with editing period 2025-7-25) was issued integrating additional corrected bitstreams (as per JVET-AM0005) together with the valid previous bitstreams as an attachment.
A DoC WG 5 N 364 on the CD was reviewed and approved on Thursday 3 July at 1700.
Editors are requested to make preparations for integration into a new edition of ITU-T H.266.1 until the next meeting.
Remains valid – not updated: JVET-AH2029 Visual quality comparison of ECM/VTM encoding [V. Baroncini, J.-R. Ohm, M. Wien] [AG 5 N 118]
Remains valid – not updated: JVET-AL2030 Optimization of encoders and receiving systems for machine analysis of coded video content (Draft 9) [S. Liu, J. Chen, J. Ström] [WG 5 DTR N 354] (2025-05-02)
Primary editor: S. Liu.
Remains valid – not updated: JVET-AI2031 Common test conditions for optimization of encoders and receiving systems for machine analysis of coded video content [S. Liu, C. Hollmann]
JVET-AM2032 Technologies under consideration for future extensions of VSEI (version 9) [S. McCarthy, J. Boyce, J. Chen, S. Deshpande, M. M. Hannuksela, H. Tan, Y.-K. Wang] (2025-08-31)
New elements from notes elsewhere in this report:
Encoder optimization info (EOI)
JVETAM0126 AHG9: Temporal quality indication for EOI SEI message
Scalability dimension info (SDI) SEI