|
Hello e'bdy I'm using the cfft_32 from the dsplib and I don't why it doesn't work as it should be. Can anybody help me? here is the part of my code, I'va red all the documents about how top use this function and I think my code is right. I don't understand why I've such bad results! //data aligned in the cmd file SECTIONS { .signal1 : align(4096) } //data declaration in the main C program #pragma DATA_SECTION (signal,".signal1") long signal[2048] ; //32 bits cfft code cbrev32(signal,signal,1024); cfft_32(signal,1024,noscale); I try first tu use it as a 32 bits real FFT so : signal[2*i] = my data signal[2*i+1] = 0 And I try to align the data with different value, to compute cfft_32 on several points and it still doesn't wotk ! So please if somebody knows how to use it, or what I made wrong... Or if somebody has a real 32 bits fft working on a C5416... Thanks for answer Seb |
problem using cfft_32 on TMS320C5416
Started by ●February 24, 2005
Reply by ●February 24, 20052005-02-24
|
I remember a problem with some of the DSPLIB functions from TI. You should check their errata to see if this is addressed: http://focus.ti.com/general/docs/buglist/bugsearch.tsp?templateIdQ21&navigationId83 I hope this helps. --- In , "rabogoss" <rabogoss@h...> wrote: > > > Hello e'bdy > I'm using the cfft_32 from the dsplib and I don't why it doesn't work > as it should be. > Can anybody help me? > here is the part of my code, I'va red all the documents about how top > use this function and I think my code is right. I don't understand why > I've such bad results! > > //data aligned in the cmd file > SECTIONS > { > .signal1 : align(4096) > } > > //data declaration in the main C program > #pragma DATA_SECTION (signal,".signal1") > long signal[2048] ; > > //32 bits cfft code > cbrev32(signal,signal,1024); > cfft_32(signal,1024,noscale); > > I try first tu use it as a 32 bits real FFT so : > signal[2*i] = my data > signal[2*i+1] = 0 > > And I try to align the data with different value, to compute cfft_32 > on several points and it still doesn't wotk ! > So please if somebody knows how to use it, or what I made wrong... > Or if somebody has a real 32 bits fft working on a C5416... > Thanks for answer > > Seb |
Reply by ●February 25, 20052005-02-25
|
yeah...there were lot of bugs in the earlier versions of the dsplib...to be on the safer side, download and use the latest dsplib from the following link and see whether you are still having problems or not.. :http://focus.ti.com/docs/toolsw/folders/print/sprc099.html --- ajmassa <> wrote: > I remember a problem with some of the DSPLIB > functions from TI. You > should check their errata to see if this is > addressed: http://focus.ti.com/general/docs/buglist/bugsearch.tsp?templateIdQ21&navigationId83 > > I hope this helps. > --- In , "rabogoss" > <rabogoss@h...> wrote: > > > > > > Hello e'bdy > > I'm using the cfft_32 from the dsplib and I don't > why it doesn't work > > as it should be. > > Can anybody help me? > > here is the part of my code, I'va red all the > documents about how top > > use this function and I think my code is right. I > don't understand why > > I've such bad results! > > > > //data aligned in the cmd file > > SECTIONS > > { > > .signal1 : align(4096) > > } > > > > //data declaration in the main C program > > #pragma DATA_SECTION (signal,".signal1") > > long signal[2048] ; > > > > //32 bits cfft code > > cbrev32(signal,signal,1024); > > cfft_32(signal,1024,noscale); > > > > I try first tu use it as a 32 bits real FFT so : > > signal[2*i] = my data > > signal[2*i+1] = 0 > > > > And I try to align the data with different value, > to compute cfft_32 > > on several points and it still doesn't wotk ! > > So please if somebody knows how to use it, or what > I made wrong... > > Or if somebody has a real 32 bits fft working on a > C5416... > > Thanks for answer > > > > Seb > __________________________________ |
