User's Guide.txt

User’s guide

For MBSS:
1.	The main program is named as mbss.m in MBSS folder
2.	The functions included are weightedSpectralAVG.m, deltaFunction.m, thresholdLevel.m, speechDetection.m, 
floorEnhaced.m, alphaFunction.m and overlapAddFunction.m
3.	Run the program and two input parameters are asked.
Input the corrupted speech file name with path within ‘’ and then enter the percentage overlap for 
windowing process. The results in this report are simulated by taking 50% of overlap.
4.	The other parameters beta and threshold values are set to 0.01 and 10 respectively in the code. These can be varied as required.
5.	 The original corrupted sound is played first and then the enhanced speech file is played.
6. 	The output file is saved as testMBSS.wav in the same folder.



For PSS:
1.	The main program is named as pss.m in PSS folder.
2.	The functions included are weightedSpectralAVG.m, thresholdLevel.m, speechDetection.m, floorEnhaced.m, alphaFunction.m and overlapAddFunction.m
3.	Run the program and two input parameters are asked.
Input the corrupted speech file name with path within ‘’ and then enter the percentage overlap for windowing process. The results in this report are simulated by taking 50% of overlap.
4.	The other parameters beta and threshold values are set to 0.01 and 10 respectively in the code. These can be varied as required.
5.	 The original corrupted sound is played first and then the enhanced speech file is played.
6. 	The output file is saved as testPSS.wav in the same folder.

Group Contribution and predefined tools used:

We have used MATLAB to program the PSS and MBSS methods. 
Predefined commands used in MATLAB are: Hamming from signal processing toolbox, FFT, IFFT, wavread, floor and repmat are from basic MATLAB toolbox.  
The code for PESQ score and the recommendation for subjective tests ITU-P.835 were downloaded from the web.