How to conserve memory for Fast Fourier Results?

I have a 16384 vector FFT which is exactly 2^14. This vector is multiplied by another vector to get the results I need.

This takes up a lot of memory in Matlab. Since the FFT is symmetric around the middle, I want to be able to take advantage of this and reduce the size of my vector by 2 and still be able to calculate the same results.

Is it possible to achieve this?

[Matt J]

If 128K takes up "a lot of memory in MATLAB", you need to upgrade your computer, my friend.

Answer 1

[Matt J]

Here's a way to do it, bus as I commented above, I'm not sure your priorities are well-reasoned.

    %%Fake data
    a=5:-1:1;
    x=[a, 3, fliplr(a(2:end))];
    N=numel(x);

    %Method 1

    y1=fft(x),  

     %Method 2 - use only half the array

      aa=x(1:N/2+1); 
    y2=ifft(aa,N,'symmetric')*N