madhavmk/Noise2Noise-audio_denoising_without_clean_training_data

Source code for the paper titled "Speech Denoising without Clean Training Data: a Noise2Noise Approach". Paper accepted at the INTERSPEECH 2021 conference. This paper tackles the problem of the heavy dependence of clean speech data required by deep learning based audio denoising methods by showing that it is possible to train deep speech denoising networks using only noisy speech samples.

47
/ 100
Emerging

This project helps audio engineers and researchers improve the clarity of speech recordings by removing background noise. It takes noisy audio files as input and outputs cleaner speech. This is especially useful for those working with diverse languages where obtaining perfectly clean speech samples for training is challenging or expensive.

208 stars. No commits in the last 6 months.

Use this if you need to train a robust speech denoising model using only noisy audio data, particularly in environments with complex or low signal-to-noise ratio conditions.

Not ideal if you already have access to abundant clean speech recordings for training traditional denoising models.

audio-enhancement speech-processing noise-reduction sound-engineering linguistic-data-collection
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 21 / 25

How are scores calculated?

Stars

208

Forks

42

Language

Jupyter Notebook

License

MIT

Last pushed

Sep 01, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/madhavmk/Noise2Noise-audio_denoising_without_clean_training_data"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.