Download All

No BSD License  

Highlights from
Wavelet Subband coding for speaker recognition

image thumbnail
from Wavelet Subband coding for speaker recognition by raghu ram
fn extracts wavelet feature vector which could be used for speaker recogniton from speech signal.

f=enframe(x,win,inc) 
function f=enframe(x,win,inc)
%ENFRAME split signal up into (overlapping) frames: one per row. F=(X,WIN,INC)


nx=length(x);
nwin=length(win);
if (nwin == 1)
   len = win;
else
   len = nwin;
end
if (nargin < 3)
   inc = len;
end
nf = fix((nx-len+inc)/inc);
f=zeros(nf,len);
indf= inc*(0:(nf-1)).';
inds = (1:len);
f(:) = x(indf(:,ones(1,len))+inds(ones(nf,1),:));
if (nwin > 1)
    w = win(:)';
    f = f .* w(ones(nf,1),:);
end

Contact us at files@mathworks.com