Code generation specific output array size

hello, I am trying these codes that below

 function [ Mshifted, f] = Func1( m ,ts,df)
 fs = 1/ts;
 Mshifted = zeros(1,4096);
 f = zeros(1,4096);
 [M,m1,df1]=fftseq(m,ts,df); 
 M=M/fs;
 f=[0:df1:df1*(length(m1)-1)]-fs/2; 
 Mshifted = abs(fftshift(M));
 end

and I obtain that prototype

    void Func1(const double m[4096], double ts, double df, emxArray_real_T *Mshifted,emxArray_real_T *f);

How can I get array that it has 4096 size.Because in C ,this is problem especially in MCU you know because of fragmentation.emxArray_real_T include in struct is data* and if I use any emxArray_real_T abc; definition ,program give an interrupt about array size and shut down.By the way I am trying in Qt Creature those codes.Do you know any book about code generation examples and applications,and any offer about specific errors and workings

Answer 1

[Denis Gurchenkov]

I think your question is "How can I get Mshifted and f to have size 4096 in the generated C code"?

With that presumption, if you use subscripted assignment, the size of the variable will be preserved. Change the last two lines of your code to:

f(:) = [0:df1:df1*(length(m1)-1)]-fs/2; 
Mshifted(:) = abs(fftshift(M));

and now f and Mshifted are both 1-by-4096 because of the previous assignments with explicit sizes.

Another way to solve this is to modify fftseq so that it returns a fixed-size array, using same technique -- preallocate the output of the right size, and then use sub-assign to assign the first part of the array the output of fft.

hth,

Denis

[Emrah Duatepe]

Hi Denis, I know this is very late response but thank you for help :). I was just checking my questions list and I have seen that no response ,sorry for that.