Scilab 3.0 release candidate version announcement
Serge Steer
version is now available on the www.scilab.org Web site.
1/ Please use this Release Candidate version in order to help us to
built a stable final version 3.0
2/ Please mention bugs in our Bugzilla database:
http://scilabsoft.inria.fr/cgi-bin/bugzilla_bug/index.cgi (Select
Scilab 3.0 Release Candidate in the subsection of the product)
Foreword:
- The on line help french version is not provided in this Release
Candidate
- Matlab to Scilab Translator: most Matlab builtin functions have been
translated. The others, particularly Matlab graphics functions, will
be translated for release 3.0
- Please enter "help graphics_entities" at Scilab prompt to have help
for new graphics. Most graphics functions helps have not been
rewritten with the philosophy of new grapics. It will be done for
3.0
New features :
Windows version has been rewritten:
- Windows look & feel with association of Scilab files with Scipad editor
- Dramatic improvement of the speed of Scilab
- Compilation using C and Fortran Intel compilers or Visual C++ .NET
2003 project
Graphics: Object oriented graphics version (default mode - See the
foreword section before):
- 3-D surfaces are taken into account
- 2-D plots can be mapped into 3-D plots
The Matlab to Scilab converter has been completely rewritten:
- Large set of Matlab functions converted
- Matlab cells and structs converted into Scilab mlists
Miscellaneous:
- New and fast implementation of hypermatrices (thanks to B.Pinçon)
- Scipad improved: shortcuts, language handling, colorization (thanks
to E. Segre & F. Vogel)
- New library of 1-D, 2-D and 3-D interpolation functions (thanks to
B. Pinçon)
- Mex functions library improvement
- Scicos: large improvement
- Beta version of implicit Scicos
The Scilab team
email: sci...@inria.fr
Richard Owlett
> We are happy to inform you that the Scilab 3.0 Release Candidate
> version is now available on the www.scilab.org Web site.
>
> 1/ Please use this Release Candidate version in order to help us to
> built a stable final version 3.0
>
> 2/ Please mention bugs in our Bugzilla database:
> http://scilabsoft.inria.fr/cgi-bin/bugzilla_bug/index.cgi (Select
> Scilab 3.0 Release Candidate in the subsection of the product)
>[snip]
I'm just getting oriented to Scilab 2.7 and like what I've learned so far.
I did not see any cautions on what should be done to have 2.7 and
3.0RC coexist. So,,, I'm paranoid ;]!
I use WinXP Pro SP1 with 1.2 GB ram and Pentium 4 CPU
Side question:
Does Scilab 3 have capability to do 3D surface plots of experimental
data *rather than* just analytic functions?
Or did I get myself lost in the 2.7 documentation ;) ?
I've done some @d plotting in Scilab with some psudeo-3D using
multiple plot overlays. Someone pointed me to gnuplot. I suspect I'll
use it for initial overviews. But I think Ill need Scilab's "zoom"
feature eventually. Scilab makes number crunching portion of my
problem trivial. THANKS!
Stéphane Mottelet
> Does Scilab 3 have capability to do 3D surface plots of experimental
> data *rather than* just analytic functions?
No. But if you can describe us the way your data can be structured, it
could help. I suppose that you have a set of 3d points that more or
less represent a surface. The best situation would be if you had a
triangulation
of you data (then you could use the trimesh or trisurf function of the
plotlib).
I suppose you don't have such a triangulation, but scilab's "mesh2d"
function
could help, if your surface can be easily projected on a plane.
S.
--
Utilisant M2, le client e-mail révolutionnaire d'Opera :
http://www.opera.com/
Richard Owlett
>
>> Does Scilab 3 have capability to do 3D surface plots of experimental
>> data *rather than* just analytic functions?
>
>
> No. But if you can describe us the way your data can be structured, it
> could help. I suppose that you have a set of 3d points that more or
> less represent a surface. The best situation would be if you had a
> triangulation
> of you data (then you could use the trimesh or trisurf function of the
> plotlib).
> I suppose you don't have such a triangulation, but scilab's "mesh2d"
> function
> could help, if your surface can be easily projected on a plane.
>
I have a VERY long .WAV file.
I use Scilab to do FFT on 100 mSec portions
[ therefor each spectra has ~4k data points ]
The input data for each FFT is offset by a constant amount
Are you saying I can easily plot this mess?
[ give me any reasonable data restriction and I can probably meet it ]
Stéphane Mottelet
a écrit:
Ok. A brute force solution would be to gather all FFT's in a big matrix
and then use the plot3d function, but the size is problematic.
How many FFT's do you want to represent at the same time ?
Anyway, you could subsample the FFT after the computation (for the
representation), since a few hundred points would give something
already good looking.