Optimize x264 based remote desktop by dirty regions

[Useprxf]

I was using x264 to achieve remote desktop, but had some problems on handling P_SKIP detection.

Dirty regions indicate changed areas. I would like to encode those macroblocks which don't intersect any dirty region as P_SKIP types.

I inserted the following code into x264_macroblock_prob_skip_internal function:

if (! h->isdirty[h->mb.i_mb_x][h->mb.i_mb_y] && ! M32(h->mb.cache.pskip_mv))
    return 1;

but there is almost no speed-up. I think it may be the information preparation for the macroblock analysis that takes influence.

My question is how to speed up x264 by considering dirty regions?

[Balling]

There is now support for that in ffmpeg: https://github.com/FFmpeg/FFmpeg/com...29d40daee4284a