A Brief Theora and Vorbis Encoding Guide
The Theora video format, Vorbis audio format, and Ogg container formats were developed by Xiph.org as free and open-source media formats. ffmpeg can create these formats by using the external encoding libraries libtheora and libvorbis.
To use these encoders make sure your ffmpeg build has been compiled with --enable-libtheora and --enable-libvorbis, or refer to the output of ffmpeg -codecs. If you want to compile ffmpeg to support these encoders see the various FFmpeg Compilation Guides for detailed instructions.
libvorbis (-codec:a libvorbis) is recommended over the very experimental, native FFmpeg Vorbis audio encoder (-codec:a vorbis -strict experimental) since it does not provide comparable quality to libvorbis.
Note: More modern alternatives like VP9 can often provide better video quality at a lower bitrate. See the VP9 Encoding Guide.
Variable Bitrate (VBR)
ffmpeg -i input.mkv -codec:v libtheora -qscale:v 7 -codec:a libvorbis -qscale:a 5 output.ogv
-qscale:v– video quality. Range is 0–10, where 10 is highest quality. 5–7 is a good range to try. If you omit-qscale:v(or the alias-q:v) then ffmpeg will use the default-b:v 200kwhich will most likely provide a poor quality output, and libtheora may drop/skip frames if the bitrate is too low.
-qscale:a– audio quality. Range is -1.0 to 10.0, where 10.0 is highest quality. Default is-q:a 3with a target of 112kbps. The formula 16×(q+4) is used below 4, 32×q is used below 8, and 64×(q-4) otherwise. Examples: 112=16×(3+4), 160=32×5, 200=32×6.25, 384=64×(10-4).
