THE AUDITORY MODELING TOOLBOX

This documentation applies to the most recent AMT version (1.6.0).

View the code

Go to function

auditoryfilterbank
Linear Gammatone auditory filterbank

Usage:

[outsig, fc] = auditoryfilterbank(insig, fs);
[outsig, fc] = auditoryfilterbank(insig, fs, flow);
[outsig, fc] = auditoryfilterbank(insig, fs, flow, fhigh);

Input parameters:

insig Acoustic signal. Size: (time x channels).
fs Sampling rate (in Hz).
flow Optional frequency (in Hz) of the lowest filter. Default: 80 Hz.
fhigh Optional frequency (in Hz) of the highest filter. Default: 8 kHz.

Output parameters:

outsig Filtered output. Size: (time x numfc x channels).
fc Center frequencies (in Hz) of the applied filterbank. The number of filters numfc is the result of flow, fhigh and bwmul. Default: 31.

Description:

auditoryfilterbank(insig, fs) applies the Gammatone filterbank to the signal insig sampled with fs. The filterbank will be composed of 31 filters from 80 Hz to 8 kHz spaced 1 ERB apart.

[outsig, fc] = auditoryfilterbank(...) additionally returns the center frequencies fc of the filterbank.

This function takes the following optional key-value pairs:

'flow',flow As flow.
'fhigh',fhigh As fhigh.
'basef',basef Center one filter exactly at basef (in Hz) and spread other filters accordingly. Default: [], which means that the filters will be spread optimally between flow and fhigh.
'bwmul',bwmul Spacing of the auditory filters in the ERB rate. Default: 1, which means that filters will be spaced one ERB apart.

Build with Bootstrap