Surface Plot - using a linear vector for displaying x axis

Hi everybody! I am using 'surf' to display a self-made spectrogram of a speech signal. Here's the code:

% plot spectrogram to axes2
nfft=512;
% reverse axes of vector 'data'
s = data';
% check if signal lenght is an equal number
if mod(length(s),2)
  n = length(s)-1;
  % crop the signal
  s = s(1:n);
  data = data(1:n);
else
  n = length(s);
end;
% window overlap value
numoverlap = nfft*3/4;
% create hann window
win = hann(nfft, 'periodic');
% determine amount of segments
k = fix((n-numoverlap)/(length(win)-numoverlap));
% create the time vector for entire signal
t=0:((n/fs)/(k-1)):(n/fs);
% create frequency vector
%f = fs/2*linspace(0,1,nfft/2+1); ---> original
f = fs/2*linspace(0,1,nfft/2); % my adaptation for better processing
% predefine the matrix to store the segments
segs=zeros(k,nfft);
% segmentation index
m=1;
% segmentation
for i=1:k
  % windowing of the segment
  w_s = s(m:nfft+m-1).*win';
  segs(i,1:nfft)=w_s;    
  % next segment
  m = m+nfft-numoverlap;    
end
% predefine the matrix for the spectral values of the segments
segspec=zeros(k,nfft);
% computation of the fft of the individual segments
for y=1:k
  segspec(y,1:nfft)=fft(segs(y,1:nfft));    
end
segspec=segspec';
spectral = segspec(1:nfft/2+1,1:k);

surf((log10(abs(spectral))+eps),'EdgeColor','none'); 
colormap(jet);
view(0,90);
axis([-inf k -inf nfft/2]);
xlabel('Analysis Windows');
ylabel('Frequency Bins');
title('Spectrogram','FontWeight','bold','FontSize',11);

On the x axis, I will now get the frequency bins 1:k. Is it possible to use another vector for displaying the numbers along the x axis, just as I can do with the normal plot function?

Example: t1=1:200; plot(t1,graphdata);

I have tried to look for any solutions in the documentation and online, but I've just hit dead ends on this.

I'd appreciate any suggestions.

[Honglei Chen]

You can certainly do surf(x,y,z), and it is documented. Does it work for you?

[Matthias]

I have seen that in the documentation, but that doesn't work for me, unfortunately. Somehow, I don't get the matrix dimensions right. It seems I've no idea what y and z should be if x is my 'spectral' matrix that is the spectrogram.