[Dspforum] More Lab 5 Questions

[Joseph Haber]

Nenad,

At least one problem is likely with the documentation of the length
you should give when generating the bit_rev_index. It should be
sqrt(8192*2) = 128, I believe.

Joseph



On 10/12/09, Nenad Uzunovic <nenad at roboticresearch.com> wrote:
>
> I am trying to use single-precision bit-reverse function and input with
> (double *) is confusing me. From documentation, I understood that
> (double *) stands for two floats i.e complex pair. Just like with two
> shorts being an int in sample function. Common sense, right :)? #$%^@
>
> Though, I still get bogus data on the output. I read documentation in
> fair detail and am not sure where more to go. I think I am passing all
> the data right.
>
> By the way, documentation for cfft references functions that can be
> found in
> C:\CCStudio\c6700\dsplib\support\fft
> that took me a while to figure out.
>
> In one of the forums online I found that someone was using
> bit_rev(current_data, PROCESS_LENGTH);
> instead of
> DSPF_sp_bitrev_cplx((double*)current_data,
>                      bit_rev_index,
>                      PROCESS_LENGTH);
> I tried it, and this time I did not get a bunch of NANs. Though I don't
> think my plots are correct. I need to investigate this more.
>
> Can we get a little more information about usage of these tools?
>
> Thanks,
> Nenad
>
>
> _______________________________________________
> Dspforum mailing list
> Dspforum at echelonembedded.com
> http://echelonembedded.com/mailman/listinfo/dspforum_echelonembedded.com
>