explaingit

facebookresearch/denoiser

Analysis updated 2026-07-17 · repo last pushed 2023-03-14

★ 1,900PythonAudience · generalComplexity · 3/5DormantSetup · moderate

TLDR

A real-time speech denoising tool that strips background noise from your microphone on a regular laptop, so video calls and recordings sound clean.

Mindmap

mindmap
  root((repo))
    What it does
      Remove background noise
      Clean speech in real time
      Reduce echo
    Tech stack
      Python
      Machine learning model
    Use cases
      Clean video call audio
      Enhance podcast recordings
      Train custom models
    Audience
      Remote workers
      Content creators
      Audio researchers

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Enable live denoising on your microphone so video calls sound clear despite a noisy room.

USE CASE 2

Clean up a podcast or video recording after the fact using the pre-trained model.

USE CASE 3

Route the denoiser's live output into your OS's audio system so any call app picks up the cleaned mic.

USE CASE 4

Train a custom denoising model on your own noisy dataset using the included training code.

What is it built with?

PythonPyTorch

How does it compare?

facebookresearch/denoisertencent-hunyuan/hy-world-2.0nvidia-nemo/datadesigner
Stars1,9001,9111,859
LanguagePythonPythonPython
Last pushed2023-03-14——
MaintenanceDormant——
Setup difficultymoderatehardmoderate
Complexity3/55/53/5
Audiencegeneralresearcherdeveloper

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 30min

Real-time use requires routing audio through the OS, pre-trained models work out of the box without training.

No license information was mentioned in the explanation.

Copy-paste prompts

Prompt 1
Show me how to install facebookresearch/denoiser and run its pre-trained model to clean up a noisy .wav recording.
Prompt 2
Walk me through setting up denoiser for live, real-time microphone denoising during video calls on my laptop.
Prompt 3
Help me route denoiser's real-time output as a virtual microphone input on macOS so Zoom picks up the cleaned audio.
Prompt 4
Using denoiser's training code, explain what a custom dataset needs to look like if I want to train a model for a specific type of background noise like traffic.

Frequently asked questions

What is denoiser?

A real-time speech denoising tool that strips background noise from your microphone on a regular laptop, so video calls and recordings sound clean.

What language is denoiser written in?

Mainly Python. The stack also includes Python, PyTorch.

Is denoiser actively maintained?

Dormant — no commits in 2+ years (last push 2023-03-14).

What license does denoiser use?

No license information was mentioned in the explanation.

How hard is denoiser to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is denoiser for?

Mainly general.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.

Alternatives & similar repos