Resampling vector from the middle sample (Keeping the symmetry)

Hi All,

I would like to re-sample a vector using different sample rates. But I would want to keep the symmetry as it is.

E.g. h = [ 0.1 0.3 0.7 0.3 0.1]; % Original vector with 5 samples h2 = resample(h,4,5); % Re-sampled vector into 4 samples

But the above command doesn't give a symmetric vector as the original one. Is there any way to re-sample keeping the symmetry around the center sample ?

Thanks very much.

[Dishant Arora]

And how would you define a centre element in case of even length sampled array?

[Dan]

Good question, in that case I assume the center length as the average (or interpolated) sample, between the mid two. Could you please provide a way to solve this ?

Thanks in advance.

Answer 1

[Image Analyst]

Try interp1(), or if you have the Image Processing Toolbox, imresize():

% Method using interp1()
h = [ 0.1 0.3 0.7 0.3 0.1]
x = 1 : length(h);
newNumberOfSamples = 4;
xq = linspace(x(1), x(end), newNumberOfSamples);
h2 = interp1(x, h, xq)
% Method using imresize (slightly different)
h3 = imresize(h, [1, newNumberOfSamples])

In the command window:

h =
          0.1          0.3          0.7          0.3          0.1
h2 =
          0.1      0.43333      0.43333          0.1
h3 =
      0.11644      0.48579      0.48579      0.11644

The numbers are somewhat different because where each puts the "end point" x values is different. Interp1, the way I did it, pins down the first and last x value while imresize picked the "in between" x values (1.5, 2.5, 3.5, 4.5)

Use whichever method you want.