I’m trying to reduce the size of my video collection because I’m running out of storage.
My strategy is to look for smaller sized files to replace very large ones. When that fails, I am using handbrake to rencode. Its working, but very slowly due to my ancient computer. Redownloading is the much faster method.
Some titles, which were originally released as non-english, have existing small files, but the English audio is only available less commonly as a huge file.
Is it possible to redownload the non-english video and map the English audio onto it? It might be too much fooling around to get syncing correct.
Extra question, does anybody know how to set up sonarr/radarr to try to reduce filesize automatically? Ive been doing all the downloading by hand.
Yes, absolutely, but not in an automated way. Remuxers do this sometimes. They compare video, audio and subtitles from multiple releases and select the best one and frame sync them. However, the sync process is very manual and there are cases when different releases have different cuts, leading audio is in sync at the beginning but went out of sync later on, ruining the whole experience.
False. You can easily automate this with ffmpeg. It’ll just suck unless you know the cue points :p
That said, you could probably cook something up using OpenAI’s whisper (or something more recent; i’m not terribly familiar with transcription; or local models in general tbh) which you can run locally to get time stamps and then cut up and match starting points. Could be a fun project.