Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/nuniz/mlx_stft
Short-Time Fourier Transform (STFT) for MLX (Apple)
https://github.com/nuniz/mlx_stft
apple audio ml-explore mlx signal-processing speech stft
Last synced: about 1 month ago
JSON representation
Short-Time Fourier Transform (STFT) for MLX (Apple)
- Host: GitHub
- URL: https://github.com/nuniz/mlx_stft
- Owner: nuniz
- License: mit
- Created: 2024-02-10T09:19:27.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2024-07-16T15:48:39.000Z (4 months ago)
- Last Synced: 2024-09-18T23:15:04.621Z (about 2 months ago)
- Topics: apple, audio, ml-explore, mlx, signal-processing, speech, stft
- Language: Python
- Homepage:
- Size: 241 KB
- Stars: 5
- Watchers: 2
- Forks: 0
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Short-Time Fourier Transform (STFT)
This module implements the Short-Time Fourier Transform (STFT) in Python using MLX.
It is designed to inherit from the nn.Module, which allows it to be used either as a standalone module or as part of a larger neural network architecture.
## Installation
You can install mlx_stft directly from pypi:
```bash
pip install mlx_stft
```
Or you can install it directly from the source code:
```bash
git clone https://github.com/nuniz/mlx_stft.git
cd mlx_stft
python setup.py install
```
## Usage
```python
from mlx_stft import STFT
# Create an instance of the STFT module
stft = STFT(n_fft=1024, win_length=256, hop_length=128, return_db=True, onesided=True)
# Compute STFT of an audio signal
y = stft(x)
```
```
x: mx.array [batch, length]
y: mx.array [batch, n_fft // 2, size of fold] if one_sided else [batch, n_fft, size of fold]
where size of fold = 1 + length // hop_length
```
## Arguments
n_fft: Number of Fourier transform points.
win_length: Length of the STFT window.
hop_length: Number of audio samples between adjacent STFT columns.
window: Type of window function to apply (default is "hann").
onesided: Whether to return only the non-redundant part of the spectrum (default is False).
return_db: Whether to return the result in decibels (default is False).
## Example
The STFT is illustrated in the picture below using the [NOIZEUS](https://ecs.utdallas.edu/loizou/speech/noizeus/) dataset's sp09.wav file.
### One-Sided STFT
### Dual-Sided STFT
