On 31.03.2026 22:48, Ferdi Scholten via ffmpeg-user wrote:
On 3/31/26 21:45, Richard Bartczak via ffmpeg-user wrote:breathe a sigh of relief, if yes. I did use AV01 encoding, with libaom. Not usable on my system, speedx leads to <= 0.002 (hints to improve it ?).Therefore libsvtav1 I use. The first results lead me to do further tests, I compare it with libx265 encoding.Different parameter, crf, tune (further parameter ?)-Further hints to make the comparison complete (tune 1-3, crf < 13, ..., 780p, 50fps, an example I can upload).My intention, compression without visible less. Regards Richardaverage crf equality I found out testing by encoding the same sources with both libx265 and libsvtav1 using standard encoder settings for the given preset settings and calculating vmaf score of the encoded video.This is with libx265 on preset slow and libsvtav1 on preset 3 (The presets that are a good tradeoff beween speed/size on my system)libx265 crf 30 == libsvtav1 crf 42 vmaf score > 75 libx265 crf 24 == libsvtav1 crf 36 vmaf score > 80 libx265 crf 20 == libsvtav1 crf 30 vmaf score > 95 libx265 crf 16 == libsvtav1 crf 24 vmaf score > 99
It's impossible to compare or equal crf values like this. They're arbitrary and depend on all other encoder settings. crf 30 at one preset can be a vastly different outcome than in another. And the relation between that can and will also differ between encoders.
OpenPGP_0xD85AAA6874B7507E.asc
Description: OpenPGP public key
OpenPGP_signature.asc
Description: OpenPGP digital signature
_______________________________________________ ffmpeg-user mailing list -- [email protected] To unsubscribe send an email to [email protected]
