JVET-R0016 JVET AHG report: Implementation studies (AHG16) [M. Zhou, J. An, E. Chai, K. Choi, S. Sethuraman, T. Hsieh, X. Xiu]

This document summarizes the activity of AHG16: Implementation studies, between the 17th JVET meeting in Brussels, BE (7–17 January 2020) and the 18th meeting by teleconference (15–24 April 2020).

The following three issues were suggested to need to be resolved:

• The newly adopted CCALF was identified to have a line buffer issue in 4:2:2/4:4:4 coding. For 4:2:2/4:4:4 chroma format, the luma deblocking is two lines behind the chroma deblocking, but the CCALF uses the collocated luma lines (after de-blocking and SAO) for processing, resulting in the need to buffer two chroma lines for each chroma component. JVET-R0233, JVET-R0291 and JVET-R0322 were submitted to address this cost issue.
• After the adoption of slice_ts_residual_coding_disabled_flag at the last meeting, the transform skip can now be combined with the regular transform coefficients coding, but its interaction with the dependent quantization and sign data hiding was suggested to be a potential issue. Several proposals had been submitted that discuss this.
• The availability check of the upper-left neighbouring sample used in CCLM is discussed in some contributions. There is asserted to be a problem when the raster scan slice is in use, such that the upper-left neighbouring sample may still be unavailable even if both the left and top neighbours are available. JVET-R0314 and JVET-R0375 discuss this. In the discussion of the AHG report, there was discussion of whether this is really a possibility or not in the current design. Further discussion was needed about this during the meeting.

The following contributions were identified for the AHG.

• In-loop filters (9)
• JVET-R0233, “AHG16: Line buffer problem of CC-ALF for 4:2:2 and 4:4:4 sequences”, N. Hu, V. Seregin, M. Karczewicz (Qualcomm)
• JVET-R0291, “AHG16: On ALF attenuation near virtual boundaries”, F. Bossen (Sharp)
• JVET-R0312, “AHG2/AHG16: A fix on chroma ALF virtual boundary position”, Y. Wang, L. Zhang, H. Liu, K. Zhang, Z. Deng (Bytedance)
• JVET-R0322, “CCALF virtual boundary issue for 4:4:4 and 4:2:2 format”, X.W. Meng (PKU), X. Zheng (DJI), S.S. Wang, S.W. Ma (PKU)
• JVET-R0208, “AHG16: Rounding correction for ALF virtual boundary processing”, A. M. Kotra, S. Esenlik, B. Wang, H. Gao, E. Alshina (Huawei)
• JVET-R0313, “AHG2/AHG16: Cleanups of chroma ALF and CC-ALF on/off control”, Y. Wang, L. Zhang, H. Liu, K. Zhang (Bytedance)
• JVET-R0128, “AHG16: On CCALF clipping” M. G. Sarwer, Y. Ye, J. Luo (Alibaba)
• JVET-R0133, “AHG16: On Clipping values for Non-linear ALF “, T. Tsukuba, M. Ikeda, Y. Yagasaki, T. Suzuki (Sony)
• JVET-R0289, “AHG16: On deblocking filter process”, N. Park, J. Nam, H. Jang, J. Lim, S. Kim(LGE)
• TSRC disabling/dependent quantization/sign data hiding (8)
• JVET-R0050, “AHG9: HLS on dependent quantization and sign data hiding”, S.-T. Hsiang, T.-D. Chuang, Y.-W. Huang, S.-M. Lei (MediaTek)
• JVET-R0083, “AHG14: Residual coding constraints for transform skip blocks”, A. Nalci, H.E. Egilmez, M. Coban, V. Seregin, M. Karczewicz (Qualcomm), M. G. Sarwer, Y. Ye, J. Luo (Alibaba)
• JVET-R0116, “AHG11/AHG14: On sign data hiding of transform skip block”, M. G. Sarwer, Y. Ye, J. Luo (Alibaba), A. Nalci, H. E. Egilmez, M. Coban, V. Seregin, M. Karczewicz (Qualcomm)
• JVET-R0141, “Disabling Dependent Quantization and Sign Data Hiding in Transform Skip blocks”, T. Hashimoto, E. Sasaki, T. Aono, T. Ikai (Sharp)
• JVET-R0153, “AHG9/AHG16: On slice_ts_residual_coding_disabled_flag”, J. Choi, S. Yoo, J. Heo, J. Choi, J. Lim, S. Kim (LGE)
• JVET-R0154, “AHG9/16: On sign data hiding for BDPCM blocks”, S. Yoo, J. Choi, J. Lim, S. Kim (LGE)
• JVET-R0271, “AHG9/AHG14: High-level constraints of dependent quantization and sign data hiding”, A. Nalci, M. Coban, M. Karczewicz (Qualcomm)
• JVET-R0325, “AHG14: Disabling dependent quantization and sign bit hiding for transform skip mode”, T.-C. Ma, X. Xiu, Y.-W. Chen, H.-J. Jhu, X. Wang (Kwai Inc.)
• Intra prediction/CCLM bug fixes/LMCS (7)
• JVET-R0280, “AHG16: Cleanup of intra reference sample filter selection”, J. Heo, H. Jang, J. Choi, J. Nam, M. Koo, J. Lim, S. Kim (LGE)
• JVET-R0281, “AHG16: Cleanup MIP flag signalling”, J. Heo, H. Jang, J. Choi, J. Lim, S. Kim (LGE)
• JVET-R0288, “AHG16: Reference samples for ISP”, F. Bossen (Sharp)
• JVET-R0314, “AHG2/AHG16: Fixes on CCLM “, Y. Wang, K. Zhang, L. Zhang, H. Liu (Bytedance)
• JVET-R0375, “AHG2/AHG16: CCLM bug fix in luma reference down-sampling “, L. Pham Van, G. Van Der Auwera, J. Chen, V. Seregin, M. Karczewicz (Qualcomm)
• JVET-R0290, “AHG16: LMCS constraint cleanup”, F. Bossen (Sharp)
• JVET-R0330, “AHG16: On clipping average luma value for chroma residual scaling factor derivation”, X. Xiu, Y.-W. Chen, T.-C. Ma, H.-J. Jhu, X. Wang (Kwai)
• Palette (2)
• JVET-R0309, “AHG16: Clean-up on palette predictor update for local dual tree.”, H. Jang, J. Nam, S. Yoo, N. Park, S. Kim, J. Lim (LGE)
• JVET-R0310, “AHG16: Clean-up by removing parsing dependency for palette”, H. Jang, J. Nam, S. Yoo, N. Park, S. Kim, J. Lim (LGE)
• Implementation (3)
• JVET-R0316, “AHG16: Normative constraints on BT and TT split under MER “, Y. Wang, K. Zhang, L. Zhang, H. Liu, Z. Deng (Bytedance)
• JVET-R0224, “AHG16: Realization of RPR based real-time VVC decode and playback on ARM based mobile devices”, J. Shingala, A. Natesan, A. Chelawat (Ittiam)
• JVET-R0390: “AHG16: VVC multi-thread decoder and performance analysis”, S. Gudumasu, T. Poirier, F. Urban, F. Hiron, P. de Lagrange (interdigital)
• Other AHG16-related contributions (1)
• JVET-R0223, “AHG16: On DMVR and wraparound motion compensation”, J. Luo, J. Chen, Y. Ye (Alibaba)

The AHG recommended to review the input contributions (see JVET-R0340).