⬢github C · 5.6K ★ +26 since we first saw it · pushed 4 d ago · NOASSERTION
Netflix/vmaf
Perceptual video quality assessment based on multi-method fusion.
VMAF (Video Multi-Method Assessment Fusion) is an Emmy-winning perceptual video quality metric developed by Netflix. It ships as a C library (libvmaf) with a Python wrapper, letting users score how good compressed video looks to humans and even train custom models. It also includes other metrics like PSNR, SSIM, MS-SSIM, PSNR-HVS, CIEDE2000, and the CAMBI banding detector.
Why now: Netflix just released a new generation of VMAF models (v1) in June 2026, describing it as 'good is not good enough', and is soliciting user feedback.
Who it is for: Video engineers and researchers who encode, stream, or evaluate video quality, especially those building or tuning codecs and transcoding pipelines.
Stars over our 367 snapshots: 5.5K to 5.6K, since 5 d ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/Netflix/vmaf