• Optimize --vpp-unsharp. ( #288 )
• Optimize --vpp-tweakby removing unnecessary pow() calls. ( #288 )
• Fix index calculation errors in --vpp-knn. ( #288 )
• Fix index calculation errors in --vpp-nlmeans. ( #288 )
• Add sigmoid options to --vpp-libplacebo-shader. ( #286 )
• Add input color space specification to --vpp-libplacebo-shader. ( #286 )
• Add warning when using --vpp-libplacebo-shader in res specification seems required. ( #286 )
• Automatically disable output thread when --lowlatency is specified.
• Optimize --lowlatency to reduce pipe latency.
• Add debug feature --vpp-mfx-insert-clcopy which will insert OpenCL copy if mfx vpp is th last of vpp pipeline.