python scripts for my stupid encoding workflow, the repo name is just randomly suggested by github
read the code it's short
calculates ssim with ffmpeg. assuming you have one output file per vapoursynth script and the output name is (vpy name)+_fin.hevc and your vs script has a video node object named src which is your reference for ssim calculation but you can modify the code anyway.
drag n drop ffmpeg generated ssim log.
it uses and displays dB form of ssim because it's easier to compare for human eyes ;p
encodes audio from source file strings find in vpy, with concat support.
and you need to manually edit this script to choose stream index.
(video) segments/segmented (encoding) all in one. moved from gist.
uses mvtools or <insert your scd method here> to detect scene change then encode in fragments with multi-process parallel work, on the fly. (i know how to use pool, i use my stupid implementation just because i thought pool was too complicated)
aomenc not doing good in mt, svtav1 refuses to use scenecut like a(n) ■■■■■, so either should benefit from this script.
react to svtav1's refusing to use scenecut: they said yes, we say no (Railgun T NCOP SVT-AV1 8bit preset 6 tune 0/VQ crf 38)
