Re: Convolving a non-uniformly sampled signal with a Gaussian

[Quentin Spencer]

Søren Hauberg wrote:

> Hi
> I have a signal that I need to convolve with a Gaussian (or another
> similar filter). The problem is that the signal is not uniformly
> sampled, meaning I can't use conv directly. My first thought was to
> interpolate the signal, perform convolution, and then resample, but I
> don't want to do this as the function that generated isn't continuos.
>
> Does anybody know a easy way to do this, or do I have to implement
> this from scratch.

I'm not sure there's an easy way, but it seems doable as long as the
function you are convolving is defined analytically (like Gaussian,
sinc, or other similar function). For each point you would need to
compute the time separation for all other points in the range of the
convolution function and compute the function for that point, so the
computation speed would be considerably slower than normal convolution,
but it could be faster than interpolation depending on just how
non-uniform your sampling is.

-Quentin



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------