Dear colleagues, I am looking for detailed description of the "zoom-fft" algorithm and its effective implementation in C/C++, Fortran or Matlab. Thanks in advance, Michal
Zoom-FFT algorithm
Started by ●January 10, 2005
Reply by ●January 10, 20052005-01-10
Michal Kvasnicka wrote:> Dear colleagues, > > I am looking for detailed description of the "zoom-fft" algorithm and its > effective implementation in C/C++, Fortran or Matlab. > Thanks in advance, > > Michal > >Rick Lyons' book "Understanding DSP" has a good description. My only beef (maybe 3rd edition, Rick?) is that he leaves the details of the low-pass/band-pass filter as an exercise to the reader. Without care, the cost of the filtering could outweigh the gains from the smaller FFT. I'd suggest multirate decimation -- Mark Borgerding.
Reply by ●January 10, 20052005-01-10
You could look up Chirp FFT. Robert "Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote:>Dear colleagues, > >I am looking for detailed description of the "zoom-fft" algorithm and its >effective implementation in C/C++, Fortran or Matlab. >Thanks in advance, > >Michal >( modify address for return mail ) www.numbersusa.com www.americanpatrol.com
Reply by ●January 11, 20052005-01-11
I am not sure that Chirp FFT is the same as Zomm FFT, see for example: http://www.numerix-dsp.com/zoomfft.html Michal <r_obert@REMOVE_THIS.hotmail.com> p�se v diskusn�m pr�spevku news:rde6u0t4pfpgjla9t1ogkftnskplcnn2ug@4ax.com...> You could look up Chirp FFT. > > Robert > > "Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote: > > >Dear colleagues, > > > >I am looking for detailed description of the "zoom-fft" algorithm and its > >effective implementation in C/C++, Fortran or Matlab. > >Thanks in advance, > > > >Michal > > > > > ( modify address for return mail ) > > www.numbersusa.com > www.americanpatrol.com
Reply by ●January 11, 20052005-01-11
Could you recommend me some good starting point for "multirate decimation"? The basic question is: Is Zoom FFT generally faster than standard FFT if I am looking only for narrow band part of spectrum? Michal "Mark Borgerding" <mark@borgerding.net> p�se v diskusn�m pr�spevku news:HcDEd.82573$mA3.259@fe2.columbus.rr.com...> Michal Kvasnicka wrote: > > Dear colleagues, > > > > I am looking for detailed description of the "zoom-fft" algorithm andits> > effective implementation in C/C++, Fortran or Matlab. > > Thanks in advance, > > > > Michal > > > > > > Rick Lyons' book "Understanding DSP" has a good description. > > My only beef (maybe 3rd edition, Rick?) is that he leaves the details of > the low-pass/band-pass filter as an exercise to the reader. Without > care, the cost of the filtering could outweigh the gains from the > smaller FFT. > > I'd suggest multirate decimation > > -- Mark Borgerding.
Reply by ●January 11, 20052005-01-11
Michal Kvasnicka wrote:> I am not sure that Chirp FFT is the same as Zomm FFT, see forexample:> http://www.numerix-dsp.com/zoomfft.html > > Michal > <r_obert@REMOVE_THIS.hotmail.com> p�se v diskusn�m pr�spevku > news:rde6u0t4pfpgjla9t1ogkftnskplcnn2ug@4ax.com... >> You could look up Chirp FFT. >> >> Robert >> >> "Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote: >> >>> Dear colleagues, >>> >>> I am looking for detailed description of the "zoom-fft" algorithm >>> and its effective implementation in C/C++, Fortran or Matlab. >>> Thanks in advance, >>> >>> Michal >>> >> >> >> ( modify address for return mail ) >> >> www.numbersusa.com >> www.americanpatrol.comFrom this description it seems closely related to what used (?) to be called "complex demodulation". The only ref I have to hand relates to its use in estimating bispectra: MG Godfrey (1965) An exploratory study of the bi-spectrum of ecomomic time-series, Applied Statistics, 14 (1), 48-69. This does give a first-principles outline of the approach. David Jones
Reply by ●January 11, 20052005-01-11
I am still can not believe to fact, that is not possible to find any freeware code (matlab, fortran, C) which implemented zoom-fft algorithm. Who knows any freeware code or detailed description of the zoom-fft algorithm??? Michal "David Jones" <dajxxx@ceh.ac.uk> p��e v diskusn�m p��sp�vku news:41e3a2b1$1@news.nwl.ac.uk...> Michal Kvasnicka wrote: > > I am not sure that Chirp FFT is the same as Zomm FFT, see for > example: > > http://www.numerix-dsp.com/zoomfft.html > > > > Michal > > <r_obert@REMOVE_THIS.hotmail.com> p�se v diskusn�m pr�spevku > > news:rde6u0t4pfpgjla9t1ogkftnskplcnn2ug@4ax.com... > >> You could look up Chirp FFT. > >> > >> Robert > >> > >> "Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote: > >> > >>> Dear colleagues, > >>> > >>> I am looking for detailed description of the "zoom-fft" algorithm > >>> and its effective implementation in C/C++, Fortran or Matlab. > >>> Thanks in advance, > >>> > >>> Michal > >>> > >> > >> > >> ( modify address for return mail ) > >> > >> www.numbersusa.com > >> www.americanpatrol.com > > From this description it seems closely related to what used (?) to be > called "complex demodulation". The only ref I have to hand relates to > its use in estimating bispectra: > > MG Godfrey (1965) An exploratory study of the bi-spectrum of ecomomic > time-series, Applied Statistics, 14 (1), 48-69. > > This does give a first-principles outline of the approach. > > David Jones > >
Reply by ●January 11, 20052005-01-11
"Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote in message news:crvq24$1shj$1@ns.felk.cvut.cz...> I am not sure that Chirp FFT is the same as Zomm FFT, see for example: > http://www.numerix-dsp.com/zoomfft.htmlMy understanding of the NumerixDSP ZoomFFT function is that it does uses quadrature frequency mixing to downconvert the area of interest to 'baseband' I/Q and then performs the FFT on that baseband. Although I haven't used that particular function within the NumerixDSP library, I have used several other parts of the library and it takes the donkey work out of re-inventing the wheel, especially if you develop cross-platform: it's convenient to be able to write and test stuff on the laptop first, and then implement it on the target platform later. Cheers, Howard
Reply by ●January 11, 20052005-01-11
Yes, NumerixDSP is really good solution, but in this time I am looking for any freeware code before I buy something better. So, do you know any freeware code for the zoom-fft??? Michal "Howard Long" <howard@howardlongxxx.com> p��e v diskusn�m p��sp�vku news:cs0v2e$95e$1@hercules.btinternet.com...> "Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote in message > news:crvq24$1shj$1@ns.felk.cvut.cz... > > I am not sure that Chirp FFT is the same as Zomm FFT, see for example: > > http://www.numerix-dsp.com/zoomfft.html > > My understanding of the NumerixDSP ZoomFFT function is that it does uses > quadrature frequency mixing to downconvert the area of interest to > 'baseband' I/Q and then performs the FFT on that baseband. > > Although I haven't used that particular function within the NumerixDSP > library, I have used several other parts of the library and it takes the > donkey work out of re-inventing the wheel, especially if you develop > cross-platform: it's convenient to be able to write and test stuff on the > laptop first, and then implement it on the target platform later. > > Cheers, Howard > >
Reply by ●January 11, 20052005-01-11
I could be mistaken, but it looks like all the freeware you need is the FFTW. The decimation and filtering is easliy implemented in a dozen or so lines of c-code. As for the filter coefficients, just use matlab to find them and then incorporate them into your code. I am guessing that the whole thing, driver and all, should come in at less than 200 lines of code. Michal Kvasnicka wrote:> I am still can not believe to fact, that is not possible to find any > freeware code (matlab, fortran, C) which implemented zoom-fftalgorithm.> > Who knows any freeware code or detailed description of the zoom-fft > algorithm??? > > Michal > > "David Jones" <dajxxx@ceh.ac.uk> p=ED=B9e v diskusn=EDm p=F8=EDsp=ECvku > news:41e3a2b1$1@news.nwl.ac.uk... > > Michal Kvasnicka wrote: > > > I am not sure that Chirp FFT is the same as Zomm FFT, see for > > example: > > > http://www.numerix-dsp.com/zoomfft.html > > > > > > Michal > > > <r_obert@REMOVE_THIS.hotmail.com> p=EDse v diskusn=EDm pr=EDspevku > > > news:rde6u0t4pfpgjla9t1ogkftnskplcnn2ug@4ax.com... > > >> You could look up Chirp FFT. > > >> > > >> Robert > > >> > > >> "Michal Kvasnicka" <michal.kvasnicka@No_s_Pam.quick.cz> wrote: > > >> > > >>> Dear colleagues, > > >>> > > >>> I am looking for detailed description of the "zoom-fft"algorithm> > >>> and its effective implementation in C/C++, Fortran or Matlab. > > >>> Thanks in advance, > > >>> > > >>> Michal > > >>> > > >> > > >> > > >> ( modify address for return mail ) > > >> > > >> www.numbersusa.com > > >> www.americanpatrol.com > > > > From this description it seems closely related to what used (?) tobe> > called "complex demodulation". The only ref I have to hand relatesto> > its use in estimating bispectra: > > > > MG Godfrey (1965) An exploratory study of the bi-spectrum ofecomomic> > time-series, Applied Statistics, 14 (1), 48-69. > > > > This does give a first-principles outline of the approach. > > > > David Jones > > > >
