"yes, vp9 would be helpful with its larger transform sizes. Plus the improved coefficient coding scheme, assymmetric transforms, new intra prediction modes, etc.
There's also some investigation going on around genuine yuv444 support,
or even lossless. However, keep in mind that most improvements of vp9 compared to vp8
are related to video coding (inter-frame prediction). And there's the cost inherent to introducing another bitstream format, which is a long-haul effort. Takeway: need some experiments and data point about extra-gain vs extra-complexity (including project complexity)"
Comments
Found this:
"yes, vp9 would be helpful with its larger transform sizes. Plus the improved coefficient coding scheme, assymmetric transforms, new intra prediction modes, etc. There's also some investigation going on around genuine yuv444 support, or even lossless. However, keep in mind that most improvements of vp9 compared to vp8 are related to video coding (inter-frame prediction). And there's the cost inherent to introducing another bitstream format, which is a long-haul effort. Takeway: need some experiments and data point about extra-gain vs extra-complexity (including project complexity)"
https://groups.google.com/a/webmproject.org/d/msg/webp-discu...