JVET-AN0010 JVET AHG report: Encoding algorithm optimization (AHG10) [K. Andersson, P. de Lagrange, A. Duenas (co-chairs), T. Ikai, T. Solovyev, A. Tourapis (vice chairs)]

Following contributions were identified relating to AHG10 and summarized in the following sections.

JVET-AN0052: [AHG17] [ Performance of GOP based RPR under CfE test conditions

This document presents results for enabling GOP based RPR encoder control for VTM (option ‘--GOPBasedRPR=1’) on CfE test conditions. VVC standard supports RPR functionality, which allows changing the scaling factor for video during coding. When GOP based RPR is enabled the first picture in each GOP is down-sampled and up-sampled before coding. Decision on scaling factor is based on PSNR computed between pristine and distorted by down-/up-sampling. Scaling factor decided this way is used for the whole GOP. The BD-rate compared to corresponding anchor configuration is on average is reported to be about -5% for random access and -2% for low-delay configurations. The document proposes to enable GOP based RPR for CfE/CfP anchor.

JVET-AN0189: AHG17/AHG10: On perceptual coding for next-generation video coding standard

This document suggests considering perceptually oriented coding tools in the development of next-generation video coding standard. In the development of previous standards (e.g., VVC, HEVC), the CTC are mostly based on PSNR metric. This paradigm of developing a standard may have two limitations that perceptual tuning capability of the standard may not be well tested and verified, and CTC which emphasizes on PSNR may not encourage the development of normative perceptual optimization tools. The document recommends investigating in objective metrics with better correlation with human vision during the preparation of CfP, in AhG17 and perhaps with the help from AG5. The document also suggest that a call for objective quality metrics for video coding quality assessment may be issued to solicit more reliable metrics beyond what we have been familiar with.

JVET-AN0190: AHG17/AHG10: On test model simulating hardware encoder

This document recommends developing the next-generation video coding standard of a native design for hardware-friendly encoding. During previous standardization, hardware decoder complexity was emphasized more than hardware encoder complexity, partly because the standards specify the decoding process. A test model developed with an encoding configuration that mimics some key characteristics of hardware encoder (besides a configuration for simulating software encoder) may facilitate more accurate evaluation on the performance of coding tools in the context of hardware encoder implementation and solicit hardware encoding friendly coding technologies. The document also encourages to have experts with hardware encoder design expertise participating in the new standard and improving its hardware friendliness with their valuable comments and suggestions.

Recommendation

The AHG recommended that the related input contributions are reviewed, and to further continue the study of encoding algorithm optimizations in JVET.

Review of the related documents in the context of CfE.