[Pd] Toeplitz solver libraries? Is there a new fast routine anywhere?

Mathieu Bouchard matju at artengine.ca
Sat Feb 18 08:38:54 CET 2006


On Fri, 17 Feb 2006, Mathieu Bouchard wrote:

> If they are not circulant, then I guess you can fake it by considering
> them each as the sum of a circulant matrix and a simpler matrix. The
> latter is actually a triangular Toeplitz minus its transpose.

lies, lies, damn lies. It looks like:
[
   0, +a, +b, +c
  -c,  0, +a, +b
  -b, -c,  0, +a
  -a, -b, -c,  0
]

So it's not really what i said it is, but still it's a subgroup of
Toeplitz.

> because a circulant matrix is like a cyclic convolution product,

damn. let's rephrase this as: the ordinary matrix product on circulant
matrices is like a cyclic convolution product.

 _ _ __ ___ _____ ________ _____________ _____________________ ...
| Mathieu Bouchard - tél:+1.514.383.3801 - http://artengine.ca/matju
| Freelance Digital Arts Engineer, Montréal QC Canada




More information about the Pd-list mailing list