libaudiomask - Simultaneous audio mask threshold estimation library

1.0

Authors:

Dr Matt R. Flax <flatmax@>

Academic

This code also implements the roex auditory filters.

Please read the article located on the debian file system file:///usr/shar/doc/libaudiomask-1.0/Flax.2000.Improved.Auditory.Masking.Models.pdf

It may also be found on the web http://www.assta.org/sst/SST-00/cache/SST-00-Chapter9-p2.pdf

Introduction

This example shows how to use Dr M.R. Flax's (2000) hybrid simultaneous audio masking class to find the masking threshold of a time domain signal.

The compilation of this file is demonstrated in Makefile. Run this file : ./AudioMaskerExample View the results of this file using www.octave.org by running the script view.m

• simply type view once octave has started and you are in the suitable directory.

The input audio should be stored in the file INPUTFILENAME in text format - each sample seperated by a white space.

========================= HOWTO ===============================

     // First find the masking threshold
     AudioMasker masker(sampleFreq, count); // Create the audio masker class using fs=sampleFreq and count filters
     masker.excite(input, sampleCount); // find the mask for the array of input data which has sampleCount time samples.
 
     // Now do something with the masking threshold ...
 
     // The frequency domain mask is now located here
     for (int j=0; j<count;j++)
         masker.mask[j]; // This is the mask at each of the count frequencies of interest
 
     // A more sophisticated example - find the threshold for each Fourier bin
     double fact=(double)sampleFreq/((double)sampleCount-1.0); // convert from an index to the equivalent * Fourier bin frequency
     for (int j=0; j<halfSampleCount;j++){
         cout<<"finding for freq "<<j*fact<<'\t'; // The frequency we are inspecting
         double threshold=masker.findThreshold(j*fact); // The masking threshold
         20*log10(threshold); // The threshold in decibels (dB)
     }