Thread

1. thenoblesunfish · 1790939127 · focus · HN ↗

   Task I want but have been too lazy - I have a favorite podcast, that I often fall asleep to. Certain parts with music etc. wake me up. I'd like to chop those parts out as automatically as possible. Can I do that here by e.g. giving some example edits to a few files, and asking to remove similar stuff from other files?

   1. playfultones · 1790940183 · focus · HN ↗

      Sounds like something a better model armed with ffmpeg would already be able to do. Run an analysis on loudness along the track, detect when speech begins/ends (with whisper) around loud segments, and compress or cut those parts out
   2. vltmrkls · 1790948158 · focus · HN ↗

      an interesting use case... the analysis could detect rhythmic or not and then edit the rhythmic parts away. i will look into, added to my backlog. thanks for your feedback!

      1. thenoblesunfish · 1791042863 · focus · HN ↗

         Cool! As the other comments suggest, I could think of plenty of ways you could do this, probably much more efficiently than having a GUI, once you figured it out. But I would still like something that felt like less software engineering or data science and more "this is how I would do it manually in Audacity, do this for me automatically on a batch of files"
   3. tervem6istus · 1790993714 · focus · HN ↗

      Actually there are models for classifying sections in audio files as speech or music, for example Silero VAD or YAMNet. I am rather confident that any decent LLM can one-shot a script that downloads the model, feeds it all the podcast episodes to generate timestamps where music is playing, and uses ffmpeg to cut out those parts. For a nice touch maybe replace each music section with a few seconds of silence, add a few second fade-out to the part preceding the silence and a few second fade-in after.

Unofficial Hacker News client; not affiliated with Y Combinator.