Differences
This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision Next revision | Previous revision | ||
|
vp8 [2020/06/09 04:11] beandog |
vp8 [2020/06/09 23:11] beandog |
||
|---|---|---|---|
| Line 6: | Line 6: | ||
| * [[https://trac.ffmpeg.org/wiki/Encode/VP8|ffmpeg wiki: Encode VP8]] | * [[https://trac.ffmpeg.org/wiki/Encode/VP8|ffmpeg wiki: Encode VP8]] | ||
| * [[http://ffmpeg.org/ffmpeg-all.html#libvpx|ffmpeg libvpx documentation]] | * [[http://ffmpeg.org/ffmpeg-all.html#libvpx|ffmpeg libvpx documentation]] | ||
| + | |||
| + | ==== Notes ==== | ||
| + | |||
| + | Both VP8 and [[VP9]] are a bit confusing on how to get good encodes related to specific levels of quality, when using the documentation and ''vpxenc'' help output as references. | ||
| + | |||
| + | In it's simplest approach, if using ''vpxenc'', use ''max-q'' as the target CRF -- the encoder won't drop quality below that level (starts at 0, ends at 63). | ||
| + | |||
| + | For ''ffmpeg'', use ''-crf'' with your number and ''-b:v 0'' is required. | ||
| + | |||
| + | For HandBrake, use ''-q'' with your CRF. I can't tell from docs or the code if it's passing anything about min-q and max-q to the encoder, but I don't think it is. I think it's using libavcodec to do crf=X and then b:v of 0. | ||
| ==== Encoding Quality ==== | ==== Encoding Quality ==== | ||
| Line 15: | Line 25: | ||
| <code> | <code> | ||
| ffmpeg -i video.y4m -vcodec libvpx -crf 6 -b:v 0 -y video.mkv | ffmpeg -i video.y4m -vcodec libvpx -crf 6 -b:v 0 -y video.mkv | ||
| + | </code> | ||
| + | |||
| + | HandBrake: | ||
| + | |||
| + | <code> | ||
| + | HandBrakeCLI -i video.y4m -e VP8 -q 6 -o video.mkv | ||
| </code> | </code> | ||
