Sign in
Search compdsp
Ads
Discussion Groups
See Also
Discussion Groups > Comp.DSP
Comp.dsp is a worldwide Usenet news group that is used to discuss various aspects of digital signal processing.
There are 104 threads in our archives.
You are looking at page 1 of 11.
The threads with the newest articles are listed first.
FFTW 3.3-beta1 available
Steven G. Johnson - 2011-06-27 01:22:00
Dear DSP and numerics folks,
You might be interested to know that a beta release of FFTW 3.3 is now
available from www.fftw.org, with the final 3.3 release scheduled for
July 25. This release in...
Re: matlab fft and fftw
Steven G. Johnson - 2011-06-20 23:05:00
I just tried your code snippet with FFTW and gcc on a GNU/Linux
system, and I can't reproduce your error. I get the same results as
the Matlab results that you quote (indeed, Matlab uses FFTW
inte...Re: analyzing large array separately for HF and LF, then re-combining subset of freq bins
Steven G. Johnson - 2010-09-23 20:11:00
On Sep 22, 3:41=A0pm, "all4dsp" wrote:
> I've got a verylarge(real)array, on the order of 200Mpts, that makes it
> unlikely to obtain an accurateFFTor IFFT. One thought to get around this
> i...Re: Cross correlation via Hartley Transform
Steven G. Johnson - 2010-09-18 18:13:00
On Sep 8, 8:15=A0am, "pete26" wrote:
> So I am doing the Hartleytransformon both signals H1(f) and H2(f) ... to
> get the cross correlation -> iFHT(1/2[ H1(f) * H2(f) + H1(-f) * H2(-f) +
> H...Re: Fast Floating-Point Format
Steven G. Johnson - 2009-10-22 11:57:00
On Oct 20, 11:09=A0am, Tim Wescott wrote:
> I would assume that for a well-scaled fixed point algorithm you'd want at
> least your input precision, plus log2 of your number of bins, plus two or...Re: fftw3: Correct results only on second run
Steven G. Johnson - 2009-10-05 16:19:00
On Oct 3, 10:10=A0am, Andreas wrote:
> When I'm using the "FFTW_MEASURE" flag for creating the plan, lots of
> transformations are done on the given data. So I lost my data before
> even doin...Re: fftw3 roundoff error in convolution c++
Steven G. Johnson - 2009-09-22 11:41:00
On Sep 21, 6:07=A0pm, "Kwisatch" wrote:
> you're so right about the sensitivity of the process values, and the
> differences in arrays. I've fixed this, so results come out OK, until the
> se...Re: Double precision in FFTW
Steven G. Johnson - 2009-09-10 15:07:00
On Sep 9, 2:17=A0am, "chiraag" wrote:
> I am relatively new to usingFFTW.. I have been usingFFTWto take fft for
> a few complex matrices. My problem is that I am not gettign the same valu=
es
...Re: fft in spherical coordinates
Steven G. Johnson - 2009-08-24 12:30:00
On Aug 24, 5:04=A0am, glen herrmannsfeldt wrote:
> I know for cylindrical coordinates there is a Fourier-Bessel
> transform, with sine/cos on the Z axis and theta axis, and
> bessel functions...Re: N-pt DFT where n != power of 2
Steven G. Johnson - 2009-08-13 23:08:00
On Aug 12, 6:52=A0pm, glen herrmannsfeldt wrote:
> If N is not prime then it can be reduced depending on the factors of N.
> If it is prime, then you are stuck.
Not true; for at least 40 yea...