[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: imagesc (specgram) problem
From: |
Michael Goffioul |
Subject: |
Re: imagesc (specgram) problem |
Date: |
Mon, 13 Aug 2007 15:09:37 +0200 |
On 8/11/07, Michael Grossbach <address@hidden> wrote:
> Schirmacher, Rolf wrote:
>
> > imagesc does not work as expected for me.
>
> It sometimes does and it sometimes doesn't work under Windows 2000 and
> Octave 2.9.13, with the same error message though. But if it (kind of)
> works it yields different results each time: different power in the
> upper left and lower right corners (either orange or pink), and the
> sweep sometimes stops on its way upwards and starts at the x axis again.
> The axes are not labelled properly. They seem to reflect the number of
> entries in x and y respectively instead of frequencies and time. And
> with regard to compatability: An old Matlab version (R11.2), when
> specgram was still called specgram and not spectrogram as it is today,
> the axes would be labelled 'Time' and 'Frequency'.
I don't have a solution for the gnuplot- or octplot-related problems, but
you might just want to give jhandles a try. It seems to work for me.
For your information, I also experience startup problems using octplot,
which crashes for unknown reason; I still couldn't figure out why this
happen. Usually, I can work around the problem by first creating the figure
(with "gcf", for instance), then calling the plot routines.
Concerning the gnuplot problem, I don't have any idea, I'm not really using it.
Michael.