Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!gem.mps.ohio-state.edu!apple!malcolm
From: malcolm@Apple.COM (Malcolm Slaney)
Newsgroups: comp.dsp
Subject: Re: Looking for 2D FFT code...
Message-ID: <35926@apple.Apple.COM>
Date: 25 Oct 89 23:17:05 GMT
References: <5047@orca.WV.TEK.COM> <9520008@hpsad.HP.COM> <2576@uceng.UC.EDU>
Organization: Apple Computer Inc, Cupertino, CA
Lines: 54

In article <2576@uceng.UC.EDU> mfinegan@uceng.UC.EDU (michael k finegan) writes:
>>Try "Numerical Recipies in C" page 467 for the complete code.
>But don't plan to compile it on a 16 bit machine, and don't mind
>wasting M+N elements of your N+1 by M+1 array ... Anyone actually
>translate all the indices from 1 <= j <= N, to 0 <= j < N ?

No, Numerical Recipies does not waste elements in the array.  The vector
and the matrix routines all allocate arrays where the first element is
indexed by 1, just like Fortran.  They do this by carefully subtracting
an offset (generally 1) from the value returned by malloc.  Then the user
(the 2d fft routine, for example) can do things just like Fortran.  See
the vector() and matrix() routines described in the appendix.

If you are going to complain about wasted space you should talk about the 
double indirection they do for all two dimensional arrays.  But, this is a 
pretty good way to get around the inability to dimension two dimensional 
arrays at run time in C but it does cost N elements of storage.

All in all I recommend the Numerical Recipies book.  I wish I had the book 
back when I was doing my thesis.

								Malcolm

/* nrl is number of rows (low index)
   nrh is number of rows (high index), etc.
 */

float   **matrix(nrl, nrh, ncl, nch)
int     nrl, nrh, ncl, nch;
{
        int     i;
        float   **m;

        m = (float **)malloc((unsigned int) (nch-ncl+1)*sizeof(float *));
        if (!m) nrerror("Allocation failure in matrix()");
        m -= nrl;

        for (i=nrl;i<=nrh;i++) {
                m[i] = (float *) malloc((unsigned) (nch-ncl+1)*sizeof(float));
                if (!m[i]) nrerror("Allocation failure in matrix()");
                m[i] -= ncl;
        }
        return m;
}

void free_matrix(m, nrl, nrh, ncl, nch)
float   **m;
int     nrl, nrh, ncl, nch;
{
        int     i;

        for (i=nrh;i>=nrl;i--) free((char *) (m[i] + ncl));
        free((char *) (m+nrl));
}