
... 

... 
12 
%

12 
%

13 
% `[g,a,fc]=erbfilters(fs)` constructs a set of filters *g* that are

13 
% `[g,a,fc]=erbfilters(fs)` constructs a set of filters *g* that are

14 
% equidistantly spaced on the ERBscale (see freqtoerb) with bandwidths

14 
% equidistantly spaced on the ERBscale (see freqtoerb) with bandwidths

15 
% that are proportional to the width of the auditory filters

15 
% that are proportional to the width of the auditory filters

16 
% audfiltbw. The filters are intended to work with signals with a

16 
% audfiltbw. The filters are intended to work with signals with a

17 
% sampling rate of *fs*

17 
% sampling rate of *fs*.



18 
%



19 
% Because the downsampling rates of the channels must all divide the



20 
% signal length, filterbank will only work for multiples of the



21 
% least common multiple of the downsampling rates. See the help of



22 
% filterbanklength.

18 
%

23 
%

19 
% `[g,a,fc]=erbfilters(fs,L,'fractional')` constructs a filterbank with

24 
% `[g,a,fc]=erbfilters(fs,L,'fractional')` constructs a filterbank with

20 
% fractional downsampling rates *a*. The rates are constructed such

25 
% fractional downsampling rates *a*. The rates are constructed such

21 
% that the filterbank can handle signal length that are multiples of

26 
% that the filterbank can handle signal length that are multiples of

22 
% *L*.

27 
% *L*, so the benefit of the fractional downsampling is that you get to



28 
% choose the value returned by filterbanklength.

23 
%

29 
%

24 
% `[g,a,fc]=erbfilters(fs,'uniform')` constructs a uniform filterbank

30 
% `[g,a,fc]=erbfilters(fs,'uniform')` constructs a uniform filterbank

25 
% where the downsampling rate is the same for all channels.

31 
% where the downsampling rate is the same for all channels.

26 
%

32 
%

27 
% `erbfilters` accepts the following optional parameters:

33 
% `erbfilters` accepts the following optional parameters:
