We're using the 8147 processor, and are trying to do some signal processing. The TI FFT and BitReverse C routines come from:
TMS320C67x DSP Library
Programmer’s Reference Guide
Literature Number: SPRU657C
January 2010
I'm seeing some strange behavior out of the BitReverse routine. Specifically, we're using dspf_sp_cfftr2(). Seeing a couple of general search posts about potential errors in the "C" FFT routine, I thought I'd ask if there are currently any known errors in these 2 routines? If not, then it's something that I'm doing in my code.
Some details:
*************** variables initialization **************
Buffer size 2048
float x[2048]
float w[2048];
************* code init **********
gen_w_r2(w, 2048);
bit_rev(w,1024);
************* code *************
for (i=0;i<2048;i++)
{
x[2*i) = (real) some function;
x[2*i + 1) = (imag) some other function;
}
dspf_sp_cfftr2_dit(x,w,1024);
bit_rev(x,512); (I've also used bit_rev(x,1024 trying to bit-reverse the output of the bit-reversed fft routine)
Input (2 channels from processor): 1 stream of real data + 1 stream of imag data
*************** question ***************
Is the above correct?
Thanks ahead of time, Don
