[SciPy-user] convolve2d.. or Numeric to numarray conversion

Kasper Souren Kasper.Souren at ircam.fr
Tue Feb 18 15:51:20 EST 2003


> I'm sorry you are running into trouble.

Well, I chose to use cvs stuff myself, so you don't have to be sorry :)

> The problem is that your version of SciPy has a bug which overwrites the
> correct convolve function (that handles N-d) with the one in Numeric
> which only handles 1-d.

> The current version (in CVS) works correctly.  Or you can fix it with a
> simple line change in scipy/signal/__init__.py

I did a "cvs update" and reinstalled the stuff. The signal.convolve2d works 
fine now, but the signal.convolve still gives me troubles:

  File "test.py", line 26, in convtest
    convolved = norm(abs(signal.convolve(data, kernel)))
  File "/usr/lib/python2.2/site-packages/Numeric/Numeric.py", line 205, in 
convolve
    return cross_correlate(a,asarray(v)[::-1],mode)

data and kernel are both 2 dimensional.

I'm happy with the signal.convolve2d function working correctly now, but maybe 
the signal.convolve still needs some fixing.

bye,
Kasper



More information about the SciPy-User mailing list