NoiseTorch README

The project's README from github.com/noisetorch/NoiseTorch. As one markdown file: README.md.

NoiseTorch-ng

Noise Supression Application for PulseAudio or Pipewire

Licence Latest

NoiseTorch-ng is an easy to use open source application for Linux with PulseAudio or PipeWire. It creates a virtual microphone that suppresses noise in any application using RNNoise. Use whichever conferencing or VOIP application you like and simply select the filtered Virtual Microphone as input to torch the sound of your mechanical keyboard, computer fans, trains and the likes.

Don't forget to leave a star ⭐ if this sounds useful to you!

Regarding the recent security incident

Due to a suspected security breach of the update server and code repository, there's been a concerted effort by the NoiseTorch community to ensure the source code and binaries are free from malicious code.

No malicious code has been found.

You can read more about the audit that was done here and here. Updates will now be retrieved from the project's releases page to avoid any risk of this reoccurring. We thank everyone for their trust and the love that they've shown towards the project in this unpleasant time.

Screenshot

Then simply select "Filtered" as your microphone in any application. OBS, Mumble, Discord, anywhere.

Demo

Linux For Everyone has a good demo video here.

Features

Download & Install

Download the latest release from GitHub.

Unpack the tgz file, into your home directory.

tar -C $HOME -h -xzf NoiseTorch_x64_v0.12.2.tgz

This will unpack the application, icon and desktop entry to the correct place.
Depending on your desktop environment you may need to wait for it to rescan for applications, or tell it to do a refresh now.

With gnome this can be done with:

gtk-update-icon-cache

You now have a noisetorch binary and desktop entry on your system.

Give it the required permissions with setcap:

sudo setcap 'CAP_SYS_RESOURCE=+ep' ~/.local/bin/noisetorch

If NoiseTorch-ng doesn't start after installation, you may also have to make sure that ~/.local/bin is in your PATH. On most distributions e.g. Ubuntu, this should be the case by default. If it's not, make sure to append

if [ -d "$HOME/.local/bin" ] ; then
    PATH="$HOME/.local/bin:$PATH"
fi

to your ~/.profile. If you do already have that, you may have to log in and out for it to actually apply if this is the first time you're using ~/.local/bin.

Uninstall

rm ~/.local/bin/noisetorch
rm ~/.local/share/applications/noisetorch.desktop
rm ~/.local/share/icons/hicolor/256x256/apps/noisetorch.png

Troubleshooting

Please see the Troubleshooting section in the wiki.

Usage

Select the microphone you want to denoise, and click "Load", NoiseTorch-ng will create a virtual microphone called "Filtered Microphone" that you can select in any application. Output filtering works the same way, simply output the applications you want to filter to "Filtered Headphones".

When you're done using it, simply click "Unload" to remove it again, until you need it next time.

The slider "Voice Activation Threshold" under settings, allows you to choose how strict NoiseTorch-ng should be in only allowing your microphone to send sounds when it detects voice.. Generally you want this up as high as possible. With a decent microphone, you can turn this to the maximum of 95%. If you cut out during talking, slowly lower this strictness until you find a value that works for you.

If you set this to 0%, NoiseTorch-ng will still dampen noise, but not deactivate your microphone if it doesn't detect voice.

Please keep in mind that you will need to reload NoiseTorch-ng for these changes to apply.

Once NoiseTorch-ng has been loaded, feel free to close the window, the virtual microphone will continue working until you explicitly unload it. The NoiseTorch-ng process is not required anymore once it has been loaded.

FAQs

Latency

NoiseTorch-ng may introduce a small amount of latency for microphone filtering. The amount of inherent latency introduced by noise supression is 10ms, this is very low and should not be a problem. Additionally PulseAudio currently introduces a variable amount of latency that depends on your system. Lowering this latency requires a change in PulseAudio.

Output filtering currently introduces something on the order of ~100ms with pulseaudio. This should still be fine for regular conferences, VOIPing and gaming. Maybe not for competitive gaming teams.

Alternatives

Building from source

Install the Go compiler from golang.org. And make sure you have a working C++ compiler.

 git clone https://github.com/noisetorch/NoiseTorch # Clone the repository
 cd NoiseTorch # cd into the cloned repository
 make # build it

To install it:

mkdir -p  ~/.local/bin
cp ./bin/noisetorch ~/.local/bin/
cp ./assets/noisetorch.desktop ~/.local/share/applications
cp ./assets/icon/noisetorch.png ~/.local/share/icons/hicolor/256x256/apps

Special thanks to

NoiseTorch (noisetorch/NoiseTorch)

Searching Database…

10.3K ·GitHub
NoiseTorch
What is this?A Linux app that creates a virtual microphone with real-time noise suppression (RNNoise). Select it as input in Discord, OBS, Mumble or any app, and it filters out keyboard, fan and background noise.
Is it a GUI app or a command-line tool?It's a GUI desktop app, not a CLI tool — setup takes two clicks. It ships as a single, small, statically linked, self-contained binary.
Do I need PulseAudio, or does PipeWire work too?It needs a Linux system running PulseAudio or PipeWire; both are supported. To build from source you'd additionally need the Go compiler and a working C++ compiler.
Will it add any lag to my calls?Mic filtering inherently adds just 10ms of latency, which is negligible. Output filtering adds around 100ms with PulseAudio — fine for calls and regular gaming, not for competitive teams.
What does the Voice Activation Threshold do?It's a strictness slider: how sure NoiseTorch must be that you're speaking before the mic sends anything. Keep it as high as possible (95% with a decent mic), lower it if you cut out mid-sentence, and reload to apply changes.
Can it filter audio output too, not just my microphone?Yes — point the apps you want cleaned up at 'Filtered Headphones' and output gets filtered the same way, though it adds roughly 100ms of latency with PulseAudio.
How is it different from noise-suppression-for-voice or Easy Effects?Both are RNNoise-based alternatives that are more complex to configure: noise-suppression-for-voice requires more terminal work, and Easy Effects adds many audio effects but only supports PipeWire.
Is it safe to install after the security incident I may have read about?After a suspected breach of the update server and repository, the community audited the source and binaries and found no malicious code. Updates now come straight from the GitHub releases page.
What are similar projects?easyeffects, BackgroundMusic, MicYou, Handy, lmms, Detect-It-Easy, pyannote-audio, speech-swift
Which topics is it in?hacktoberfest2023, voice-activity-detection, voice, app, audio, utility, hacktoberfest, go, gui, free, desktop-app, linux
Creator
noisetorch
Size
6 MB
Stars
10.3K
Updated
CommitJan 13Releasev0.12.2Jun 17
Age
6y
Commits · 90 Days