DeepFilterNet vs NoiseTorch-ng
DeepFilterNet
github.comNoiseTorch-ng
github.comSide by side
| Figure | DeepFilterNet | NoiseTorch-ng |
|---|---|---|
| Price from | Free | Free |
| Free plan | Yes | Yes |
| Free trial | — | — |
| Open source | Yes | — |
| Windows | Yes | — |
| macOS | Yes | — |
| Linux | Yes | Yes |
| API | Yes | — |
| Self-hosted | Yes | — |
| Plans published | 2 | 1 |
| Spec score · AI Noise Cancellation Software | 9.0 (#1) | 7.4 (#11) |
| Spec score · AI Speech Enhancement Software | 8.7 (#4) | 7.2 (#24) |
| Spec score · Noise Cancellation Software | 8.9 (#4) | 7.4 (#8) |
| Latency | The LADSPA plugin has a minimum latency of 20 ms from STFT processing, plus latency from the LADSPA host. | Microphone filtering introduces 10 ms of inherent latency, while output filtering is approximately 100 ms with PulseAudio. |
| Purpose | DeepFilterNet is a low-complexity speech enhancement framework for full-band 48 kHz audio using deep filtering. | NoiseTorch-ng is an easy-to-use open-source application for Linux that suppresses microphone noise in real time. |
Plans
In their own words
DeepFilterNet
DeepFilterNet is a free, open-source framework for speech enhancement on full-band 48 kHz audio, using deep filtering. Its command-line tool suppresses noise in WAV files, with 48 kHz as its only supported sampling rate. A LADSPA plugin provides real-time noise suppression and can be used with PipeWire filter chains to create a virtual microphone or audio output, including for VoIP apps such as Zoom or Discord. The plugin's minimum latency is 20 ms from STFT processing, plus any delay added by the LADSPA host. The Rust command-line tool defaults to DeepFilterNet2; the Python interface documents DeepFilterNet and DeepFilterNet2. An optional post-filter may slightly over-attenuate very noisy sections. The framework supports Linux, macOS and Windows, though training is tested only under Linux. The Python package installs as deepfilternet, and an optional train extra adds data-loading functionality for Linux training. Code is available under either the MIT License or Apache License 2.0.
NoiseTorch-ng
NoiseTorch-ng is a free, open-source Linux application that reduces microphone noise in real time. It uses RNNoise and works with PulseAudio or PipeWire, creating a virtual “Filtered Microphone” for applications such as OBS, Mumble, and Discord. The project describes setup as a two-click process and distributes the app as a small, statically linked binary. A Voice Activation Threshold can be set as high as 95%, or set to 0% to dampen noise without deactivating the microphone. Optional output filtering routes applications through a virtual “Filtered Headphones” device. Microphone filtering adds 10 ms of inherent latency; output filtering is approximately 100 ms with PulseAudio. The project reports a suspected update-server and repository breach, says no malicious code was found, and says updates are retrieved from GitHub releases. GitHub reports no SECURITY.md file or published security advisories. Troubleshooting guidance, including how to collect debug logs, is in the project wiki. The latest listed release is v0.12.2, released June 17.