python-speech-enhancement icon indicating copy to clipboard operation
python-speech-enhancement copied to clipboard

Published 20 hours ago • verified publisher icon chenwj1989

snr estimate

Open rainskyfyy opened this issue 5 months ago • 2 comments

what are #1~#5 mean in the 75~79 lines of file "pns/noise_suppressor.py"? like this: #1 rough noise estimation #2 rough a priori and posteri snr estimation #3 speech presence prabability estimation #4 precise noise estimation #5 a priori and posteri snr estimation I only see codes of condition "fnz_flag==1" here. Are you going to accomplish them in the future? Actually, I want to do snr estimate with your code in this project. could you please tell me how I can do the calculation? Thanks !!!

rainskyfyy avatar Jan 10 '24 06:01 rainskyfyy

hi @rainskyfyy I think #1~5 is implemented in noise_estimator (line 82), which are the steps of IMCRA.

SouppuoS avatar Jan 20 '24 09:01 SouppuoS

Hi @rainskyfyy , sorry I've not very active on github for a while.

The nosie estimation is implemented in noise_estimator.py.
Please refer to Cohen's papers for more details on noise estimation. I. Cohen and B. Berdugo, Speech Enhancement for Non-Stationary Noise Environments, Signal Processing, Vol. 81, No. 11, Nov. 2001, pp. 2403-2418. I. Cohen, Noise Spectrum Estimation in Adverse Environments: Improved Minima Controlled Recursive Averaging, IEEE Trans. Speech and Audio Processing, Vol. 11, No. 5, Sep. 2003, pp. 466-475.

Thank you @SouppuoS for the reply.

chenwj1989 avatar Jan 31 '24 06:01 chenwj1989