Thread 107157111

Are VMAF scores interpreted as a percentage of quality? Like does a score of 95 correlate with 95% quality?

I'm asking because I'm curious if I'm wasting space ripping DVDs to x264 CRF 18. I can usually reduce the 4-8 GB DVD mpeg2 mux to around 2GB.

Mp4 related is an example I'll be using as a baseline to compare to other smaller rips. Has a bitrate of ~1,600 Kbps.

Using x264 CRF 41, I got ~50% VMAF with this one. Does it really look half as good as the CRF 18 one? Bitrate shrunk to a meager ~100 kbps but quality is potato of course.

Next is x264 CRF 33 which gave me ~80% VMAF. Does this look somewhat as good as the CRF 18 one?

Bitrate still shrunk to a puny ~200 kbps.

x264 CRF 29 = ~90% VMAF. Looks pretty close to the original.

~300 Kbps

>>107157111 (OP)
wmaf scores you say?

Finally x264 CRF 25 = ~95% VMAF. I can barely tell this one and the CRF 18 rip apart.

~500 kbps

>>107157275
Ye, they seem to be less confusing than SSIM scores. Those will usually be like 0.9XXX from shit quality to high quality so interpreting them is kind of a pain.

Ignoring the audio bitrate for now, if I redo my ~2GB ~2 Mbps DVD rips I'd roughly get the following savings:

95% VMAF = ~75% lower file size or ~500 MB per movie

90% VMAF = 85% lower file size or ~300 MB per movie

80% VMAF = 90% lower file size or ~200 MB per movie

50% VMAF = 95% lower file size or ~100 MB per movie

I don't think things are bad enough for me to consider the last 2. Kinda strange having a video bitrate lower than the AAC audio bitrate too...

>>107157111 (OP)
Of course Netflix would make a rating for Vantablack male and Asian female pairings.

>>107157408
Better than Sexually Stimulated Indian Men I suppose.

viciously malfunctioning autism finder

>>107157408
Mutt's law. I interpreted it as "vampire males".