Only putting this here for documentation that I already had this implemented and while indeed the compression ratio was better than with e.g. QOI (170 mbps peek with WebP on a test feed where QOI peeks at 270 mbps) the CPU usage was so insanely high that I can't see the scenario where this would ever be useful. It's also visible in this benchmark.
Only putting this here for documentation that I already had this implemented and while indeed the compression ratio was better than with e.g. QOI (170 mbps peek with WebP on a test feed where QOI peeks at 270 mbps) the CPU usage was so insanely high that I can't see the scenario where this would ever be useful. It's also visible in this benchmark.