[SciPy-Dev] scikit-signal or Similar

Stuart Mumford stuart@mumford.me...
Wed Feb 1 03:53:47 CST 2012


Hello all,


> *That's great. Have you been following the discussion that's happened
> about this package earlier on this list? Here's a summary I made -
> http://brocabrain.blogspot.in/2012/01/scikit-signal-python-for-signal.html
> *
>

I have read the discussion and your blog post. I think that development in
scikit-signal is a good thing as long as we keep open the possibility of
merging bits (all) of it into other places later. It really depends on
where the project goes, as you said in your blog post, I don't really
understand the intricacies of namespaces either so I am just happy to work
on some code.

*But I think the above shows that this really belongs in scipy. I think we
> should either improve scipy.signal.wavelets or look at merging pywavelets
> into scipy. This particular wheel gets reinvented way too often.*
>

I agree that scipy.signal.wavelets needs improving, the reason myself and
my friend started developing this wavelet code, was the only piece of
Continuous Wavelet Transform code we could find was the piece we based what
is now in the GitHub on. Even that had major omissions to what we needed
and therefore we have spent time making the code fit our needs.

As for pyWavelets that seems to be a good standalone project and appears to
be good at Discrete Wavelet Transforms, which I have not looked into. Again
there is no need to reinvent the wheel so I don't think implementing a DWT
into SciPy is necessary, however with the amount of applications for CWT I
feel it would be better off in SciPy when it is ready.

*We'd also be interested in having wavelet code in scikits-image (
> http://skimage.org), since we need it for denoising *


I am interested in this application for wavelets, and how to expand the
current 1D CWT into 2D. Do you know the advantages / disadvantages for CWT
/ DWT for your applications?

*Wow, I took a look at the wavelet.py code. I, for one would learn like to
> learn from you. I want to learn to start coding like that.*
>

I am flattered! I have never been taught OOP I just sort of blunder through
so I hope what I produce is decent code!

*I don't think pyHHT will be a part of scikit-signal for some time, both
> are projects in their infancy.*
>

pyHHT certainly needs a lot of work and scikit-signal needs some code, but
I think eventually this could be the aim.

*Right now I'm working on time-frequency analysis (for the scikit-signal).*
>

Cool, what?

*Although HHT too is ultimately a tool for time-frequency analysis, we need
> to create enough motivation for using the HHT over other conventional
> methods.*
>

>From my (very limited) knowledge of what is good for what in signal
processing, HHT to me is pretty impressive in what it can do. But as for
using something like that it's all about what data you have and what you
want to learn from it. I am studying the Sun for my PhD and the primary
reason I wish to use HHT (well EMD) is to calculate the periods of
oscillation in my very short data sets. [And I will go to extreme lengths
to avoid using IDL]

*But of course, as an independent project, you are welcome to contribute.
> I've put a crude version up at https://github.com/jaidevd/pyhht*
>

I have already cloned it and started tinkering, but I need to do some more
theoretical research first as I don't fully understand it.

Stuart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/scipy-dev/attachments/20120201/f3d864ff/attachment.html 


More information about the SciPy-Dev mailing list