hi. i am trying to implement FFT in DIT for 8 pt in VHDL language as twiddle factors should be using CORDIC algorithm. i already implemented 8 -pt FFT for 8 bit preasion for singned input in c language. few days ago i found cfft core.but it is implemented in radix 4 fft. so,for my FFT in DIT for 8 pt what shoud i make changes? can you provide me some material like fsm diram or something??? i really need help in my code.plz help me. plz respond as soon as possible. thank you very much pratibha _______________________________________________ http://www.opencores.org/mailman/listinfo/cores

hi.. u didn't specify that you want fft implemented in c language as twiddle factor replace by CORDIC algorithms. anyway i am sending u fft in c language without using cordic algorithms. hope so it will work for you. pratibha --------------------------------------------------------------------- #include #include #include /***************************************************** *******************************/ void init_array(int n, double *A_re, double *A_im); void compute_W(int n, double *W_re, double *W_im); void output_array(int n, double *A_re, double *A_im, char *outfile); void permute_bitrev(int n, double *A_re, double *A_im); int bitrev(int inp, int numbits); int log_2(int n); void fft(int n, double *A_re, double *A_im, double *W_re, double *W_im); /***************************************************** *******************************/ /* gets no. of points from the user, initialize the points and roots of unity lookup table * and lets fft go. finally bit-reverses the results and outputs them into a file. * n should be a power of 2. */ int main(int argc, char *argv[]) { int n; int i; double *A_re, *A_im, *W_re, *W_im; n = 64; A_re = (double*)malloc(sizeof(double)*n); A_im = (double*)malloc(sizeof(double)*n); W_re = (double*)malloc(sizeof(double)*n/2); W_im = (double*)malloc(sizeof(double)*n/2); assert(A_re != NULL && A_im != NULL && W_re != NULL && W_im != NULL); for (i=0; i %g,%g\n", input_buf, input_buf[i+1], A_re[NumPoints], A_im[NumPoints]); */ /* printf("%g %g\n", A_re[NumPoints], A_im[NumPoints]); */ NumPoints++; } #endif for (i=0; i>= 1; } return rev; } /* returns log n (to the base 2), if n is positive and power of 2 */ int log_2(int n) { int res; for (res=0; n >= 2; res++) n = n >> 1; return res; } /* fft on a set of n points given by A_re and A_im. Bit-reversal permuted roots-of-unity lookup table * is given by W_re and W_im. More specifically, W is the array of first n/2 nth roots of unity stored * in a permuted bitreversal order. * * FFT - Decimation In Time FFT with input array in correct order and output array in bit-reversed order. * * REQ: n should be a power of 2 to work. */ void fft(int n, double *A_re, double *A_im, double *W_re, double *W_im) { double w_re, w_im, u_re, u_im, t_re, t_im; int m, g, b; int i, mt, k; /* for each stage */ for (m=n; m>=2; m=m>>1) { /* m = n/2^s; mt = m/2; */ mt = m >> 1; /* for each group of butterfly */ for (g=0,k=0; gcagliostro82 at g...> To: Date: Mon Nov 21 23:19:03 CET 2005 Subject: [oc] about cfft core---NEED HELP

> Hi, i'm working on fft too for an exam at my university (please
> forget
> my english mistakes, I'm Italian). Could you send to me your fft
> implementation in c so I can work a little on it? I have to perform
> a
> timing simulation, comparing c results with the risults of a fft
> core
> on a fpga.
> Please I really need help!!
> Thanks!!
>

_______________________________________________ http://www.opencores.org/mailman/listinfo/cores