JVET-F0102 Industry recommendation for Future Video Coding Standard Development [E. Alshina (Samsung), J. Zhou (Huawei), S. Lei (MediaTek), C. Luo (China Telecom), A. Hinds (CableLabs)] [late]

This document made recommendations for Future Video Coding standard development and suggestions regarding standardization practices. This contribution was described as being intended to serve as a starting point for discussing improvements for an improved future video coding standard process to meet market requirements in an efficient and timely matter.

This was reviewed in the joint meeting.

In JVET-F0102, a late contribution commenting on process, planning, and requirements, was presented. Regarding HEVC development, it praised the test model methodology and advocated for “one tool, one functionality”, the principle that clear design and sufficient gain need to be shown to add something to the design, and the incorporation of strong complexity analysis. It advocated holding to a firm two-year timeline upon adoption of the first test model, and “freezing” the low-level design 6 months prior to completion. It advocated having a large and diverse CTC test set, perhaps changing the selected sequences frequently, and ensuring that testing bit rates are aligned to application requirements. It requested emphasis on simpler coding features with modest complexity needs, and evaluation of complexity on GPU and MPU platforms as well as reference software.

Some experts expressed concern that they did not have a sufficient chance for studying this document before it was presented. It was also mentioned that some of the aspects may be premature to discuss, while other aspects are probably valid and reflect the best practices of past standardization efforts rather than being new suggestions.

Break-out Groups