> I never wrote a H.263+ codec from scratch, I just hacked the TMN stuff > (years ago). Thus it is basically all TMN ... > > TMN is (was?) the reference implementation. Purpose of that encoder was > to check out the how good H.263 compression works, it isn't designed for > speed. > > Not sure what the openh323 / gnomemeeting people use these days. H.261 and now QV30 - H.263 is a patent minefield