https://github.com/seanghay/cleanvoice
A Fast Speech Enhancement toolkit using Conv-TasNet
https://github.com/seanghay/cleanvoice
convtasnet source-separation speech-enhancement speechenhancement
Last synced: about 2 months ago
JSON representation
A Fast Speech Enhancement toolkit using Conv-TasNet
- Host: GitHub
- URL: https://github.com/seanghay/cleanvoice
- Owner: seanghay
- Created: 2024-04-02T08:13:40.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2024-04-04T02:32:18.000Z (about 2 years ago)
- Last Synced: 2024-08-09T20:10:04.253Z (over 1 year ago)
- Topics: convtasnet, source-separation, speech-enhancement, speechenhancement
- Language: Python
- Homepage: https://pypi.org/project/cleanvoice/
- Size: 17.1 MB
- Stars: 3
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## CleanVoice
A Fast Speech Enhancement toolkit using [Conv-TasNet (Yi Luo, Nima Mesgarani)](https://arxiv.org/abs/1809.07454)
> Only works with 16kHz audio.
## Install
```shell
pip install cleanvoice
```
## Usage
Audio file
```python
from cleanvoice import cleanvoice
from scipy.io.wavfile import write as write_wav
audio_data = cleanvoice("audio.wav")
write_wav("audio-cleaned.wav", 16000, audio_data)
```
NumPy Array as input
```python
from cleanvoice import cleanvoice
from scipy.io.wavfile import write as write_wav
import librosa
y, _ = librosa.load("audio.wav", mono=True, sr=16000)
audio_data = cleanvoice(y)
write_wav("audio-cleaned.wav", 16000, audio_data)
```
## License
`Apache-2.0`