[FFmpeg-user] mjpeg "quality"

Carl Eugen Hoyos cehoyos at ag.or.at
Mon Mar 4 12:45:13 CET 2013


Erik Slagter <erik <at> slagter.name> writes:

> artemis erik:~LD_LIBRARY_PATH $ ./ffmpeg -y -i
> /var/foto/Todo/video/orig/00000.MTS -an -sn -codec:v mjpeg 
> -q:v 32 -f matroska test.mkv

Running "ls -l test.mkv" will allow you to see the size of 
the output file, now please test with "qscale 2" and 
investigate the size of the second output file.

For future requests:
Please always post your failing command line together with 
the complete, uncut console output whenever asking for 
help here. Please do not post configure output etc. 
except if asked to.

> ffmpeg version N-50460-g393dcbf Copyright (c) 2000-2013 
> the FFmpeg developers
>   built on Mar  4 2013 11:50:51 with gcc 4.7.2 (GCC) 
> 20120921 (Red Hat 4.7.2-2)

>   configuration: --prefix=/usr --libdir=/usr/lib64 --shlibdir=/usr/lib64
> --enable-libopencore-amrnb --enable-libopencore-amrwb
> --enable-libmp3lame --disable-libfaac --enable-libx264 --disable-vaapi
> --disable-zlib --enable-postproc --enable-static --enable-shared
> --enable-pthreads --enable-gpl --enable-version3 --enable-nonfree
> --disable-debug --cpu=nocona --enable-avfilter --enable-frei0r
> --enable-libfreetype --enable-runtime-cpudetect

Unrelated:
Did you encounter any problems without --disable-vaapi?
If yes, please report them!

The following look useless to me (they are default):
--disable-libfaac --enable-postproc --enable-static 
--enable-pthreads --enable-avfilter

I don't think you should use "--enable-nonfree" - the 
only difference it makes in your case is that you are 
not allowed to distribute the resulting binaries.

Only use --disable-zlib if you know what you are doing, 
it can lead to surprising error messages on run-time.
(zlib is needed for some things that you may not expect.)

In old MPlayer times, --enable-runtime-cpudetect was 
strongly discouraged. I still wonder what good it does 
with today's FFmpeg (if any).

Carl Eugen



More information about the ffmpeg-user mailing list