Closing that last few percentage gap in performance with dav1d is going to be tough because it's the de-facto implementation. However, positioning rav1d as a safer platform to build the reference AV2 implementation could give that lead to rav1d. Have you started that work or done any lobbying in this area?
Closing that last few percentage gap in performance with dav1d is going to be tough because it's the de-facto implementation. However, positioning rav1d as a safer platform to build the reference AV2 implementation could give that lead to rav1d. Have you started that work or done any lobbying in this area?