Bug 96247

Summary:	Add support for waterfall plot
Product:	[Applications] kst	Reporter:	Andrew Walker <arwalker>
Component:	general	Assignee:	Rick Chern <rchern>
Status:	RESOLVED FIXED
Severity:	wishlist	CC:	kst
Priority:	NOR
Version:	1.x
Target Milestone:	---
Platform:	unspecified
OS:	Linux
Latest Commit:		Version Fixed In:

Description Andrew Walker 2005-01-03 20:58:24 UTC

Version:            (using KDE KDE 3.2.1)
OS:                Linux

It would be useful to be able to display a waterfall plot, allowing greater resolution on the x-axis without increasing the widrh of the plot.

A waterfall plot displays some range of the data as is currently done for curves. If any data remains at larger x-values it is displayed offset in the negative y-direction, with the x-values continuing from where they were left off. This can be done for any number of lines, until the lines drop off the bottom of the plot.

The vertical offset should be user configurable.

Comment 1 Netterfield 2005-01-11 05:34:42 UTC

What I meant by a 'waterfall plot' in the document is a way of showing the evolution of spectral content of a data stream.

X is Time
Y is frequency
Z is fourier amplitude.

So a slice at fixed X is the power spectrum of the data in that time slice.
A slice at fixed Y is the time evolution of that spectral component.

Consult google for more information.  Most examples use 3d mesh plots, but kst would display them as an image, with Z as color.  I find this clearer anyway (though not as high-tech looking).

Comment 2 Andrew Walker 2005-01-11 19:22:05 UTC

We could achieve this using the morlet wavelet, and having the user specify the necessary parameters. The result would be an image. This would be quite flexible, though would necessarily assume that the data is evenly sampled in time (unless we want to get into unevenly sampled transforms).

Comment 3 Netterfield 2005-01-12 18:01:08 UTC

Yes.  This is a good example.  I might consider a more general case of 
apodised fourier transforms, of which this is a subset, to allow for spectral 
coadding, for the case where the user is interested in the evolution of 
incoherent noise and low q resonences, like we have in psd's now.

On January 11, 2005 01:22 pm, Andrew Walker wrote:
> ------- You are receiving this mail because: -------
> You are the assignee for the bug, or are watching the assignee.
>
> http://bugs.kde.org/show_bug.cgi?id=96247
>
>
>
>
> ------- Additional Comments From arwalker sumusltd com  2005-01-11 19:22
> ------- We could achieve this using the morlet wavelet, and having the user
> specify the necessary parameters. The result would be an image. This would
> be quite flexible, though would necessarily assume that the data is evenly
> sampled in time (unless we want to get into unevenly sampled transforms).
> _______________________________________________
> Kst mailing list
> Kst@kde.org
> https://mail.kde.org/mailman/listinfo/kst

Comment 4 Rick Chern 2005-06-14 20:09:42 UTC

Added CSD object but holding off on slave matrix creation until the matrix/datasource re-working has been agreed upon.

Comment 5 Rick Chern 2005-07-16 01:43:38 UTC

Cumulative Spectral Decays (CSDs) have been implemented with output matrices.