- Home
- AI Noise Cancellation Software
- Noise Suppression for Voice
Noise Suppression for Voice
What it is
Noise Suppression for Voice is a free real-time microphone noise-suppression plugin based on Xiph’s RNNoise. It targets computer fans, office and crowd sounds, aircraft, cars, trains, and construction noise. The project supports VST2, VST3, LV2, LADSPA, AU, and AUv3 formats. Its minimal interface includes parameters and diagnostic statistics, including voice-activity detection threshold and grace-period controls. Voice Activity Detection can silence audio below a chosen voice-probability threshold, while grace periods help reduce cutoffs around word beginnings and endings. Audio must be 16-bit at 48000 Hz, with one or more channels supported. Documentation covers Equalizer APO on Windows and PipeWire or PulseAudio on Linux; the maker also reports testing with Carla and Audacity on Linux. macOS is listed, but setup instructions are still marked as TODO. The project is licensed under GPL-3.0. The maker cautions that poor microphone quality may not improve and that speech can be classified as noise. Echo cancellation is not provided.
For and against
For
- Free GPL-3.0 plugin
- Supports six plugin formats
- Voice-activity threshold and grace controls
- Windows and Linux setup documented
- Targets multiple environmental noise types
Against
- Requires 16-bit 48000 Hz audio
- MacOS setup instructions are unfinished
- Poor microphones may not improve
- Speech can be mistaken for noise
The verdict
Noise Suppression for Voice provides configurable RNNoise-based suppression across common plugin formats and desktop systems. Check the audio requirements and platform documentation first, especially if using macOS or a weak microphone.
Who it suits
This plugin suits microphone users on Linux, macOS, or Windows who need real-time suppression of listed environmental noises. It is also useful for recording and compatible host applications supporting its plugin formats.
On the record
- Audio requirements
- The plugin requires 16-bit audio at 48,000 Hz and supports one or more channels.github.com · read 3 Oct 2026
- Compatibility
- The maker reports testing with Equalizer APO v1.2 x64, PipeWire on Arch Linux, Carla on Linux, and Audacity on Linux.github.com · read 2 Oct 2026
- Controls
- Release v1.21 lists Wet/Dry controls, an updated RNNoise model, and runtime CPU detection for AVX2, SSE, or default paths.github.com · read 3 Oct 2026
- Integrations
- The README says the plugin was tested with Equalizer APO on Windows and PipeWire, Carla, and Audacity on Linux.github.com · read 3 Oct 2026
- Latest release note
- Version 1.21 lists an updated RNNoise model, wet/dry controls, and runtime CPU detection for AVX2, SSE, and default paths.github.com · read 2 Oct 2026
- License
- The project is licensed under GNU General Public License v3.0.github.com · read 2 Oct 2026
- License and contributions
- The project is licensed under GPL-3.0, and the README welcomes improvements and asks contributors to open an issue before sizeable contributions.github.com · read 3 Oct 2026
- Limit
- The maker cautions that the plugin may not improve poor microphone audio and could misclassify voice as noise.github.com · read 3 Oct 2026
- Linux integration
- The README documents setup with PipeWire and PulseAudio.github.com · read 2 Oct 2026
- Linux setup
- The README provides setup instructions for PipeWire and PulseAudio, including creating a noise-reduced microphone source.github.com · read 3 Oct 2026
- macOS status
- The README’s macOS section says contributions are welcome and setup instructions are TODO.github.com · read 3 Oct 2026
- Noise types
- The project says it is meant to suppress noise such as computer fans, office and crowd noise, and sounds from airplanes, cars, trains, and construction.github.com · read 3 Oct 2026
- Plugin formats
- The README lists VST2, VST3, LV2, LADSPA, AU, and AUv3 formats.github.com · read 3 Oct 2026
- Purpose
- It is a real-time voice noise suppression plugin based on Xiph’s RNNoise.github.com · read 3 Oct 2026
- Security
- The repository's Security and quality tab shows 0 security items.github.com · read 2 Oct 2026
- Support
- The maker welcomes improvements and asks contributors planning sizeable changes to open an issue first.github.com · read 2 Oct 2026
- Voice detection
- Voice Activity Detection settings can silence audio below a configurable voice probability threshold, with grace periods to help avoid cutting off word endings or beginnings.github.com · read 3 Oct 2026
- Windows integration
- On Windows, the README describes enabling the VST2 plugin in Equalizer APO.github.com · read 2 Oct 2026
