[FFmpeg-devel] Encoding less than a whole frame

[James Darnley]

I have a few vague questions I want to ask but keep in mind I have no
knowledge of FFmpeg's APIs.  It is my impression that encoding with
avcodec requires a whole picture to be given to the encoder at any time.

Has anyone ever started working on allowing avcodec to encode anything
less than a whole picture?  Does anyone know of anyone else ever
attempting similar work?

I think x264 allows this but it was long, long ago when I hung around
that project and I might be mis-remembering its features.  As an extreme
example the Dirac/VC2 encoder in avcodec could encode as little as 2
lines without needing the rest of the picture.