Audacity GuideFirst, here is great guide using Audacity:
https://audiobookbay.lu/forum/guides/reducing-background-noise-with-audacity/Noise Reduction GuideDisclaimer: Not expert. You are an expert? Please advise.
Use case:You found good audiobook. Happy times. But good audiobook is bad tape copy of tape copy of tape copy? Background noise is much?
Listening too much modern quality audio, your ears are picky. They can not bear it. Simple solution: remove the noise!
How: Web (50MB per file only):Drag and drop.
https://audioalter.com/noise-reducerHow: Windows with NVIDIA GTX:Lucky you.
https://github.com/amirldn/rtx-voice-script/tree/masterHow: Linux:#1 sox:
Not AI. Require noise sample (no voice, noise only). Guide here:
https://unix.stackexchange.com/a/427343Works great with "The Road Less Travelled".
https://audiobookbay.lu/abss/then-road-less-traveled-and-beyond-m-scott-peck/Track 1 sample 00:00:00-00:00:06. Can sample first few seconds every track.
#2 RNNoise:
AI*. Not need sample.
1. Download model here (
https://github.com/GregorR/rnnoise-models)
2. Which model? Audio contains music? It will harm music. But Conjoined Burgers, Marathon Prescription are best options. No music? Leavened Quisling, Beguiling Drafter. Choose one.
3. Use ffmpeg. Replace model path with path for .rnnn file. No include ><.
ffmpeg -i in.mp3 -af 'arnndn=m=<MODEL PATH>' out.mp3
#3 DeepFilterNet:
AI*. Not need sample.
https://github.com/Rikorose/DeepFilterNetQuestion? Ask.
*neural network, not sentient
