Code

Programming Languages: C

License: BSD License

Repositories

browse code, statistics http://kissfft.hg.sourceforge.net:8000/hgroot/kissfft/kissfft (read-only)

Show:

What's happening?

  • Broken #include guard in kissfft.hh

    kissfft.hh begins with the line: #ifndef KISSFFT_CLASS_HH but it never defines such macro, so the include guard doesn't actually work. Including kissfft.hh twice will result in an error.

    2012-05-20 09:21:33 PDT by lanurmi

  • Followup: RE: Dynamically Update the FFT Size

    Nothing jumps out at me. Perhaps stackoverflow.com would be a better forum to get advice on your threading woes.

    2012-05-02 09:25:53 PDT by mborg

  • Followup: RE: Dynamically Update the FFT Size

    That's what I was thinking, but it doesn't seem to be working, maybe I'm doing something wrong: I added a function in the thread where I do the FFT: [code] void processingThread::updateFFTSize(nfft) { dataMutex.lock() // dataMutex is a qMutex //class variable for fftSize fftSize = nfft; free(rawData); // complex rawData - kiss_fft_cpx* free(cfg);...

    2012-05-02 08:35:59 PDT by https://www.google.com/accounts

  • Followup: RE: Dynamically Update the FFT Size

    Pretty much everything in the cfg structure depends on the FFT size. As far as your threading issue, perhaps you could update an FFT size variable from the GUI thread and the processing thread could reallocate the structure at its earliest convenience.

    2012-05-01 19:06:24 PDT by mborg

  • Dynamically Update the FFT Size

    Is it possible to update the FFT Size without free'ing and re-allocating the memory space for the rawdata, cfg, and fftResult arrays? I do my FFT processing in a separate thread (from the GUI) and I need to be able to update the FFT size from the GUI. I didn't see anything in the tools that would help me do this. Am I missing something? Is there an example of doing this out there (Google...

    2012-05-01 18:41:41 PDT by https://www.google.com/accounts

  • Followup: RE: Inputting complex data to kiss_fft

    Just so that it's here: rawData[i].i = (x[i]/0x1000); rawData[i].r = floor((double)(x[i]/0x1000)); instead of: rawData->i[i] = stuff; rawData->r[i] = stuff; Thank you! ~ A.

    2012-03-22 08:40:30 PDT by https://www.google.com/accounts

  • Followup: RE: Inputting complex data to kiss_fft

    Mark, FWIW, the email notifications with the square brackets that I got came through fine.

    2012-03-21 21:09:05 PDT by rogerwo

  • Followup: RE: Inputting complex data to kiss_fft

    I can see from the email that got sent that you are putting your square brackets in the wrong place. I've tried two different ways to post a reply to this but the dumb bbcode markup won't allow square brackets. Sourceforge is trying hard to become irrelevant. Ask future questions on stackoverflow.com -- much better forums.

    2012-03-21 19:48:32 PDT by mborg

  • Inputting complex data to kiss_fft

    I have a complex signal that I want to pass into the kiss_fft function, however when I try to populate the real and imaginary parts of the kiss_fft_cpx struct I get compile errors (subscript requires array or pointer type) I may be doing this completely wrong - but this is what I attempted: [code//In a header kiss_fft_cfg cfg; kiss_fft_cpx *fftResult; kiss_fft_cpx *rawData; //...

    2012-03-21 19:27:27 PDT by https://www.google.com/accounts

  • Followup: RE: Plus or minus i

    It so happens that Numerical Recipes was the implementation I was comparing against, and after a little more digging, it seems clear that they are the ones who are backwards. Wikipedia and FFTW and apparently everyone [i]except[/i] Numerical Recipes uses the same convention as Kiss FFT. Thank you again, and have a great day!

    2012-02-03 20:59:29 PST by Vector76

Our Numbers