I am trying to encode a stream using x264 (by submitting individual images), but the unusual thing is that I already have motion information for my frames. I know for sure which areas were changed in each frame, and I know where the movement occurred in the frame.
Is there a way to pass x264 my own motion information? I would like to give him motion vectors for given areas in the frame and somehow say that certain areas in the frame are guaranteed to have no movement in them.
I think that this can significantly improve the coding performance (because I allow the codec to completely skip the motion estimation phase), and should also improve the quality somewhat in cases where the encoder motion estimation aldego might have missed the motion that actually happened.
Do I need to modify the encoder for this, or is it supported in the existing API?
source share