Hi folks! I've been developing this simple plugin that makes use of Xiph's RNNoise library for speech denoising. It works more as a gate with mild noise reduction over voice present segments and it detects voice presence beautifully as RNNoise uses AI in order to recognize voice in the audio. One thing that it cannot do yet is configuring the amount of reduction to apply so that's why it sounds similar to a gate on very noisy recordings. However I feel that could be very useful for audio production, podcasting, video editing and more. Let me know if you find it useful .
---->>> https://github.com/lucianodato/speech-denoiser
speech-denoise an lv2 plugin based on RNNoise library
Moderators: MattKingUSA, khz
- lucianodato
- Established Member
- Posts: 156
- Joined: Sat May 15, 2010 9:00 pm
- Has thanked: 3 times
- Been thanked: 16 times
speech-denoise an lv2 plugin based on RNNoise library
Last edited by lucianodato on Tue Apr 24, 2018 4:11 pm, edited 1 time in total.
Arguy (IRC)
Re: speech-denoise an lv2 plugin based on RNNoise library
I'll try it out next time I do an interview for the OSM podcast. Thanks!
_ssj71
music: https://soundcloud.com/ssj71
My plugins are Infamous! http://ssj71.github.io/infamousPlugins
I just want to get back to making music!
music: https://soundcloud.com/ssj71
My plugins are Infamous! http://ssj71.github.io/infamousPlugins
I just want to get back to making music!
Re: speech-denoise an lv2 plugin based on RNNoise library
HOT!lucianodato wrote:It works more as a gate with mild noise reduction over voice present segments and it detects voice presence beautifully as RNNoise uses AI in order to recognize voice in the audio.
- lucianodato
- Established Member
- Posts: 156
- Joined: Sat May 15, 2010 9:00 pm
- Has thanked: 3 times
- Been thanked: 16 times
Re: speech-denoise an lv2 plugin based on RNNoise library
It's the possibilities what is exciting about this library. You can perfectly train the learning network with guitar samples and get a perfect gate out of it. The library comes with a prebuilt training based on voice samples but as I said could be anything. I haven't explored that yet but I might in the future.
Arguy (IRC)