Pleased to announce my second report on LCEVC Technology entitled LCEVC x264 Report: Live Sports & eGames, ABR Ladder. The first report assessed VOD performance at 1080p; this report tested live performance using a full encoding ladder and weighted average performance as well as BD-Rate comparisons. Here’s the Table of Contents. Note, in particular, the GB Tech report in the Appendix which documents the subjective testing performed for this report.
The weighted average measure compared two full encoding ladders, one encoded with LCEVC with x264 as a base layer, and the other encoded with x264. To compute bandwidth savings and quality uplift we factored in estimated rung viewing consumption for each ladder, computing bandwidth consumption, VMAF, and MOS performance for each encoding approach. As opposed to BD-Rate scores, this approach delivers a much closer to a real-world assessment of bandwidth savings and quality impact.
The results show that LCEVC with x264 as a base layer should enable a bandwidth reduction of 14% for eGames and 22% for sports while increasing both MOS and VMAF quality scores.
BD-Rate performance based on the full encoding ladders without rung consumption numbers was also impressive. As shown below, LCEVC with an x264 base layer delivered 54.2 savings in eGames and 44% savings in Sports clips as measured by subjective trials (MOS), for a total savings of 45.8%. VMAF results were 33.1% savings for eGames and 38.2% for Sports for an average savings of 32.5%.
All in all, it was another impressive performance by LCEVC in two very relevant use cases.