werman/noise-suppression-for-voice
Noise Suppression For Voice
Features
- Audio Denoising Models - Implements a neural network model that removes background noise from voice audio in real time, available in multiple plugin formats.
- Real-Time Audio Filters - Processes 48 kHz multi-channel audio through a trained neural network to suppress background noise in real time.
- Audio Noise Suppression - Implements a real-time neural network filter that suppresses background noise from 48 kHz multi-channel audio.
- Plugin Format Implementations - Ships a real-time voice noise suppression plugin in VST2, VST3, LV2, LADSPA, AU, and AUv3 formats.
- Linux Audio Routings - Chaining noise suppression into PipeWire or PulseAudio to create a system-wide denoised microphone source on Linux.
- Voice Denoising Tools - Provides an open-source tool for removing background noise from voice communications, suitable for streaming and conferencing.
- Audio Plugin Architectures - Builds on the JUCE framework to provide consistent audio plugin behavior across platforms.
- Cross-Platform Plugin Implementations - Ships a cross-platform audio plugin that runs on Linux, Windows, and macOS for system-wide noise cancellation.
- Multi-Format Plugin Suites - Provides a suite of plugin binaries in VST2, VST3, LV2, LADSPA, AU, and AUv3 formats for audio workstations and system pipelines.
- Neural Network Denoisers - Provides a deep learning-based audio filter that suppresses background noise from microphone input and recordings in real time.
- Equalizer APO VST2 Hosting - Loads the VST2 plugin inside Equalizer APO on Windows to apply noise suppression to the entire microphone input path.
- System-Wide Denoising Integrations - Integrates with PipeWire, PulseAudio, and Equalizer APO to apply noise cancellation to the entire microphone input path.
- Audio Pipeline Integrations - Integrates a neural network noise suppression filter into the Windows Equalizer APO audio pipeline for system-wide microphone denoising.
- Audio - Wraps the denoising engine behind a unified interface that compiles into multiple audio plugin formats.
- Audio Plugin Hosting - Loads as a standard audio plugin in VST2, VST3, LV2, LADSPA, AU, and AUv3 formats for use in digital audio workstations.
- Multi-Standard Plugin Loading - Loads as a standard audio plugin in VST2, VST3, LV2, LADSPA, AU, and AUv3 formats for use in digital audio workstations.
- Voice Activity Detection - Applies a configurable voice activity detection gate with probability thresholds and grace periods to mute non-voice segments.
- Gating Configurations - Applies configurable probability thresholds and grace periods to mute non-voice segments without clipping speech.
- Threshold Tunings - Adjusts probability thresholds and grace periods to control when non-voice sounds are silenced without cutting off speech.
- Intelligent Noise Suppression Alternatives - Provides a neural network-powered alternative to traditional noise gates, intelligently removing background noise while preserving voice clarity.
- PipeWire Filter Chains - Configures the plugin as a PipeWire filter chain to create a noise-canceling audio source device.
- PipeWire Integrations - Exposes the denoising plugin as a filter chain node in PipeWire or PulseAudio for system-wide noise cancellation.
- PulseAudio Sinks - Creates a PulseAudio sink and loopback to route microphone audio through the noise suppression plugin.
