dodiku/noise_reduction

Speech noise reduction which was generated using existing post-production techniques implemented in Python

noise_reduction

See test results on: https://dodiku.github.io/noise_reduction/

Audio enhancements feature tests in Python3

Installation

To install:

1. $ brew install sox
2. $ brew install vorbis-tools
3. Create a virtualenv
4. Install dependencies in one of two options:

• manually (recommended):
  $ pip3 install librosa
$ pip3 install pysndfx

• or automatically using pip:
  $ pip3 install -r requirements.txt

To run:
$ python3 noise.py

Interesting resources:

• LibROSA (documentation + repository + paper)
• Think DSP (book + repository)
• Pyo (blog post + repository)
• pysndfx (repository)

A bit less relevant papers:

• Noise Cancellation Method for Robust Speech Recognition (PDF)
• Robust Features for Noisy Speech Recognition using MFCC Computation from Magnitude Spectrum of Higher Order Autocorrelation Coefficients (PDF)
• Improving the Noise-Robustness of Mel-Frequency Cepstral Coefficients for Speech Processing (PDF)

noise_reduction's Issues

when I run noise.py, met follows problem, can help find some causes?

(py3virtualenv) [root@node1 noise_reduction]# python noise.py
Traceback (most recent call last):
File "noise.py", line 230, in
y_reduced_centroid_s = reduce_noise_centroid_s(y, sr)
File "noise.py", line 64, in reduce_noise_centroid_s
y_cleaned = less_noise(y)
File "/home/yunbin.zhang/virtualenvs/py3virtualenv/lib/python3.6/site-packages/pysndfx/d
sp.py", line 276, in call raise RuntimeError(stderr.decode())
RuntimeError: sox FAIL bass: usage: gain [frequency(100) width[s|h|k|q|o]]

[WinError 2] The system can not find the file specified

There is a error of '[WinError 2] The system can not find the file specified' when I run the program in the line of y_clean = less_noise(y), but this line code has not the file operation.

RuntimeError: sox FAIL bass: usage: gain [frequency(100) [width[s|h|k|q|o]](0.5s)]

im using reduce_noise_power()
i also changed threshold_l vale to 1E-08
but still getting this error.

RuntimeError: sox FAIL bass: usage: gain [frequency(100) width[s|h|k|q|o]]

sox --version
sox: SoX v14.4.2

type errors

System give a error: the parameter of data_expanded is float type, but it should be int type. So I switch the type to int, system also give a error: the parameter of data_expanded is int type, but it should float type. And I'm fuzzed!

RuntimeError: sox FAIL bass

Hi, when i launch your tool, system return this error

Traceback (most recent call last):
  File "noise.py", line 230, in <module>
    y_reduced_centroid_s = reduce_noise_centroid_s(y, sr)
  File "noise.py", line 64, in reduce_noise_centroid_s
    y_cleaned = less_noise(y)
  File "/home/user/Documenti/RESEARCH/noise_reduction/venv/lib/python3.5/site-packages/pysndfx/dsp.py", line 276, in __call__
    raise RuntimeError(stderr.decode())
RuntimeError: sox FAIL bass: usage: gain [frequency(100) [width[s|h|k|q|o]](0.5s)]

How could I resolve this problem?
Thanks

ModuleNotFoundError: No module named 'numba.decorators'

Hi, I ran into this problem after I installed numba correctly.Could you please tell me how to solve it?
The error message is:
from numba.decorators import jit as optional_jit
ModuleNotFoundError: No module named 'numba.decorators'

The program collapse when the audio file is big

I tested a wav file with length around 2 hours.
The memory of the computer is 16GB.
Program always collapse, but not every time.