There really aren't any transparent fast fft convolutions in SciPy. The
closest thing is in signaltools, fftconvolve, and if you ask it to convolve,
say, sequences whose length add up to 7902, then it will do a size 7901
transform.


BTW, is this really a glitch ? I think there is two schools there :
- those who think the software must do something
- those who think it is the programmer's responsibility.

If I give to a fftconvolve sequences that add up this way, I want it to use
a 7901 transform, not a 2^n transform. So you understood I'm in the second
group.
If people want to use a convolution with fft, they _should_ know a little
about this algorithm. I'm a moderator on a French programming forum, and in
the algorithm area, there are people that even don't know the FFT algorithm
that want to make complicated thing with it, it is bound to fail. I suppose
that indicating this "limitation" in the docstring is enough, so that people
make the decision of is it enough or not

Matthieu
_______________________________________________
Numpy-discussion mailing list
[email protected]
http://projects.scipy.org/mailman/listinfo/numpy-discussion

Reply via email to