JVET-T0007 JVET AHG report: Coding of HDR/WCG material (AHG7) [A. Segall, E. François, W. Husak, S. Iwamura, D. Rusanovskyy]
This document summarizes the activity of AHG7: Coding of HDR/WCG Material between the 19th meeting by teleconference (22 June – 1 July 2020) and the 20th meeting by teleconference (7–16 October 2020).
The primary activity of the AhG was related to the mandates of studying and evaluating available HDR/WCG test content, comparing the performance of the VTM for HDR/WCG content, and coordinating with AHG4 in preparation for verification testing for HDR video content.
The AhG generated CTC anchors for the VTM according to JVET-P2011. A summary of the performance is provided below, and more detailed information may be found in Excel spreadsheets attached with the AHG report.
NOTE: It is noted that the Class H2 sequences were updated at the 19th JVET meeting. This update consisted of changing the chroma sampling location for the sequences. In the tables below, the performance of the VTM10 and VTM9 are reported using the configurations provided in the VTM repository to provide a reference for participants tracking the performance. Similarly, the HM16.18 results use the VTM9 sequences. Please note that the difference in Class H2 is largely due to the sequence change.
All Intra | ||||||||||
Over VTM-9.0 | ||||||||||
| wPSNR | PSNR |
| |||||||
DE100 | PSNR-L100 | Y | U | V | Y | U | V | EncT | DecT | |
Class H1 | −0.03% | −0.02% | −0.02% | −0.01% | −0.09% | −0.02% | −0.04% | −0.11% | 100% | 101% |
Class H2 |
|
|
|
|
| 1.78% | −12.52% | 2.87% | 96% | 99% |
Overall | −0.03% | −0.02% | −0.02% | −0.01% | −0.09% | 0.64% | −4.58% | 0.98% | 98% | 100% |
Random Access | ||||||||||
Over VTM-9.0 | ||||||||||
| wPSNR | PSNR |
| |||||||
DE100 | PSNR-L100 | Y | U | V | Y | U | V | EncT | DecT | |
Class H1 | −0.08% | −0.07% | −0.07% | −0.07% | −0.08% | −0.07% | −0.07% | −0.08% | 102% | 100% |
Class H2 |
|
|
|
|
| 2.58% | −16.82% | 14.54% | 99% | 100% |
Overall | −0.08% | −0.07% | −0.07% | −0.07% | −0.08% | 0.89% | −6.16% | 5.23% | 101% | 100% |
VTM 10.0 versus HM 16.18
All Intra | ||||||||||
Over HM-16.18 | ||||||||||
| wPSNR | PSNR |
| |||||||
DE100 | PSNR-L100 | Y | U | V | Y | U | V | EncT | DecT | |
Class H1 | −41.23% | −26.74% | −26.21% | −56.67% | −52.06% | −23.54% | −52.57% | −45.24% |
|
|
Class H2 |
|
|
|
|
| −19.90% | −53.76% | −48.39% |
| |
Overall | −41.23% | −26.74% | −26.21% | −56.67% | −52.06% | −22.22% | −53.00% | −46.39% |
|
|
Random Access | ||||||||||
Over HM-16.18 | ||||||||||
| wPSNR | PSNR |
| |||||||
DE100 | PSNR-L100 | Y | U | V | Y | U | V | EncT | DecT | |
Class H1 | −31.63% | −31.52% | −31.23% | −46.19% | −38.96% | −28.19% | −41.12% | −31.28% |
|
|
Class H2 |
|
|
|
|
| −26.75% | −63.02% | −56.00% |
| |
Overall | −31.63% | −31.52% | −31.23% | −46.19% | −38.96% | −27.67% | −49.08% | −40.27% |
|
|
The chroma position issue was discussed. The test sequences had changed in preparation for the verification tests, which is the reason for the large changes in chroma performance VTM10 vs. VTM9 in the table above. (See the notes of the meeting report of the previous meeting.)
The AHG had significant coordination with AHG4, including sharing results on the new HDR content and expert viewing procedures.
Two contributions related to HDR video coding were noted to have been registered for the meeting.
- JVET-T0082 New HLG sequences for high bit-depth video coding [K. Kondo, M. Ikeda, and A. Browne]
- JVET-T0106 AHG4/AHG7: Report on Display Settings for HDR/WCG Verification Tests [A. Segall, M. Wien and V. Baroncini]
The AHG recommended to review the relevant input contributions.