[SciPy-user] scipy cvs : bug in fftconvolve ?

Xavier Gnata gnata at obs.univ-lyon1.fr
Mon Feb 20 10:54:37 CST 2006


Hi,

(If it is a repost, please just ignore).

Current scipy CVS complains about a
"NameError: global name 'fftn' is not defined"
calling
scipy.signal.fftconvolve(I,K)
where I and K are 2d numpy arrays.

Looking at signaltools.py, you can see that we only import the following
fft related stuff:
from scipy.fftpack import fft, ifft, ifftshift, fft2, ifft2

Looks like fftn is just missing. Anyway, the code is buggy because fftn
is not defined anywhere.

BTW, is fftconvole the best way to perform a 2D fft-based convolution
with scipy? (If not, naming conventions are not very intuitive ;))

Xavier

-- 
############################################
Xavier Gnata
CRAL - Observatoire de Lyon
9, avenue Charles André
69561 Saint Genis Laval cedex
Phone: +33 4 78 86 85 58
Fax: +33 4 78 86 83 86
E-mail: gnata at obs.univ-lyon1.fr
############################################ 



More information about the SciPy-user mailing list