[Ffmpeg-cvslog] [ffmpeg]: r5412 - /trunk/doc/faq.texi

Robert Swain robert.swain
Sat May 27 19:40:35 CEST 2006

On Saturday 27 May 2006 10:44, michael at ffmpeg.org wrote:


> +ffmpeg -i input -acodec aac -ab 128 -vcodec h264 -b 1200 -ar 48000 -mbd 2
> -coder 1 -cmp 2 -subcmp 2 -s 368x192 -r 30000/1001 -title X -f psp -flags
> loop -trellis 2 -partitions parti4x4+parti8x8+partp4x4+partp8x8+partb8x8
> output.mp4

-mbd, -cmp, -subcmp ? As far as I am aware (and I quickly skimmed the source 
that I updated...) those three switches will do nothing, except for -cmp 
which may affect chroma depending on its argument. The me_method switch 
affects the ME search method/pattern (dia, hex, umh, esa are available) but 
the 'cmp' functions are fixed internally. With regard to -mbd 2, which 
enables RD MB decision, there's the -subq 6 and 7 which use RDO for ME and MB 
decisions and -flags2 +brdo to enable RDO for b-frames.

Also, this message might be of use (though all flag switches require arguments 
and I didn't know this at the time):



More information about the ffmpeg-cvslog mailing list