Can VP8 do CRF, similar to x264?

[orion44]

The documentation about VP8 on the internet is confusing. From what I've read, it has constant quality mode in one pass,
which is probably constant quantizer mode, without regard to visual perception.

VP9, however, does seem to have a mode similar to CRF, to achieve a certain perceptual quality level, but you have to do it in two passes:

ffmpeg -i input.mp4 -c:v libvpx-vp9 -b:v 0 -crf 30 -pass 1 -an -f null /dev/null && \
ffmpeg -i input.mp4 -c:v libvpx-vp9 -b:v 0 -crf 30 -pass 2 -c:a libopus output.webm

I was just curious, if anyone knows, does this also work with VP8?
Can FFmpeg's libvpx achieve a certain perceptual quality level in two passes, similar to x264's CRF?

[Selur]

afaik: no

[orion44]

I've been testing it, and it doesn't have a mode similar to x264's CRF. Never mind, I'll use the regular two-pass VBR mode.

Why hasn't VP8 codec been more popular, like x264 or x265?

IMO, it produces a more visually pleasing look than x264, VP9 or XviD, once the loop filter has been set correctly.

[Selur]

Why hasn't VP8 codec been more popular, like x264 or x265?

From my experience:
a. it is slow, was even slower when it was newer and folks really tested it
b. you couldn't get any useful documentation on it (got better with vp9)
c. I had to wait ages for any feedback when reporting bugs
d. personally, I prefer x264 and Xvid over it.

Cu Selur

[benwaggoner]

Selur is right. VP8 wasn't good enough to get enough attention to get it good. It also had some architectural limitations that made it suboptimal for HW and multithreaded HW implementations.

VP9 got a good bit more refinement, and AV1 was where the VPx lineage became viable for broader use cases instead of mainly being used for user generated content.