| Help plotting a 3D Carioid... [message #47498] |
Wed, 15 February 2006 08:46  |
![Steve[3] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "Steve[3] is currently offline")
Steve[3]
Messages: 7
Registered: January 2006
|
Junior Member |
|
|
I'm a newbie to IDL, and am having a bit of trouble doing stuff that is
pretty basic in other tools. In this case, I'm trying to generate a
surface plot of a 3D cartioid function. I start by defining azimuth
and elevation angle matrices, then compute the range of the cartioid
for each angle pair. I then want to convert from spherical to
rectangular coordinates and make a surface plot of the result.
In Matlab, I was able to do the coordinate transform on the matrices in
n x n form, but it appears that IDL wants the data as column vectors.
However, somewhere it appears that things are getting messed up, as the
resulting plot has a "hole" in it. However, when I use xplot3D to plot
the result, the hole is not there...
Additionally, the surface plot just doesn't have a nice look - is there
anything a bit nicer I could do, along the lines of xplot3D, but be
able to do it in a pre-specified window (I'm going to be having the
result plot in a GUI)? Best would be to have the surface shaded and
with the color set to correspond with the magnitude of each point.
I guess what I'm looking for is a little guidance on how to get this to
look a bit nicer, and what's causing the hole when I use surface...
;ang=[0:.1:2.*pi+.1]'; (this is what I used in
Matlab...)
ang=transpose(2*!PI*findgen(64)/63)
n=size(ang, /N_ELEMENTS )
ones=fltarr(n,1)+1
theta_a=(ang##ones)
;ang2=[0:.05:pi+.05]'-pi/2; (this is what Iused in Matlab)
ang2=transpose(!PI*findgen(n)/(n-1)-!PI/2) ; (should be roughly
equivalent to above)
theta_e=transpose(ang2##ones)
rcart=1+cos(!pi/2-theta_e)
theta_a_1d=reform(theta_a,1,n*n)
theta_e_1d=reform(theta_e,1,n*n)
rcart_1d=reform(rcart,1,n*n)
sph=[theta_a_1d,theta_e_1d,rcart_1d]
rectc=cv_coord(from_sphere=sph, /to_rect)
x_1d=rectc(0,*)
y_1d=rectc(1,*)
z_1d=rectc(2,*)
x=reform(x_1d,n,n)
y=reform(y_1d,n,n)
z=reform(z_1d,n,n)
surface, z,x,y
xplot3d, x,y,z
|
|
|
|
| Re: Help plotting a 3D Carioid... [message #47571 is a reply to message #47498] |
Wed, 15 February 2006 15:29  |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Steve writes:
> Thanks for your prompt and detailed reply... I just checked out your
> extensive website, and your fsc_surface routine is very close to what
> I'm looking for... However, I'm not sure if I can use it with repeated
> calls.
It was not designed for that purpose, no.
> Do you (or anyone else on here) know if it is possible (within the GUI)
> to surf something in a separate object window with fsc_surface (or
> XPlot3D), then when I push another button on the GUI, have the first
> object window close, and open a new one for the second plot (or just
> overwrite the old with the new)?
>
> Also, is it possible to simply have this object window display in the
> GUI by replacing the base with a specified base?
Everything is possible in IDL. Well, excluding truly device-independent
programs, of course. But everything within the limited imagination
of astronomers, surely. :-)
The problem is, how to do it. :-(
What you like about IDL (I can tell) is how easy it is
to use. You type a command, you see a result in a graphics
window. Brilliant! But that only works in what we call "direct
graphics." There is another whole (and completely different
and incompatible) graphics system in IDL that is MUCH more powerful
and MUCH more difficult to use. It is called the "object
graphics" system.
The FSC_Surface program appears to do what the SURFACE
command does (plus a bit more), but it is MUCH more
complicated than that. You write a least a page of
code in the object graphics system to replicate what is
already done for you in the SURFACE command. And then
you have to display object graphics output in an object
graphics window. You can't mix and match direct graphics
commands and output with object graphics commands and
output.
You can't just willy-nilly throw FSC_Surface output
in any window, because the *interaction* with the surface
depends on the window. (If you don't like the rotations,
etc., you could, of course, just throw a static rendering
of your surface in the window. This is about 3/4 of a page
of code, but easily done. You just have to build the surface,
the axes, rotate everything into your view, establish the
view, stuff everything into a model, etc., etc. It really
isn't as bad as it sounds. See Simple_Surface as a stripped
down example of FSC_Surface.)
And, of course, what you do in a widget program is
completely divorced from the graphics system you are
using. Although you have to know how to create draw widgets
that represent windows in one system or the other, depending
on which you are using. But that is trivial.
So, where does that leave you? At the moment it leaves you
writing direct graphics code (WSET, SURFACE, etc.) and wishing
you knew how to write object graphics code (IDLgrWindow, IDLgrSurface,
IDLgrAxis, etc, etc.). I'd get Ronn's book. Otherwise, you will
be at this a LONG time. Maybe longer than you have the stomach
for it. Especially as you are coming from a Matlab background.
You can't be thinking pleasant thoughts about IDL at the moment. :-)
Cheers,
David
P.S. There is another possibility. You could try iTools.
Almost none of us here can figure out how to program
the damn things, but there is always the small change they
work the way you want them to work. Try iSurface. Let us
know how it goes. :-)
P.S.S. If it is any consolation to you, I feel for you,
I really do.
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: Help plotting a 3D Carioid... [message #47573 is a reply to message #47498] |
Wed, 15 February 2006 14:36 |
Steve[3]
Messages: 7
Registered: January 2006
|
Junior Member |
|
|
Thanks for your prompt and detailed reply... I just checked out your
extensive website, and your fsc_surface routine is very close to what
I'm looking for... However, I'm not sure if I can use it with repeated
calls.
Do you (or anyone else on here) know if it is possible (within the GUI)
to surf something in a separate object window with fsc_surface (or
XPlot3D), then when I push another button on the GUI, have the first
object window close, and open a new one for the second plot (or just
overwrite the old with the new)?
Also, is it possible to simply have this object window display in the
GUI by replacing the base with a specified base?
Thanks...
-Steve
P.S. The following is along the lines of how I am attempting the
plot...
pro ant_test_event, event
widget_name=widget_info(event.id, /uname)
;print, widget_name
case widget_name of
'pushit': begin
;print, 'in fslider1_roll'
widget_control, event.top, get_uvalue=global_data
;ang=[0:.1:2.*pi+.1]';
ang=transpose(2*!PI*findgen(14)/13)
n=size(ang, /N_ELEMENTS )
ones=fltarr(n,1)+1
theta_a=(ang##ones)
;ang2=[0:.05:pi+.05]'-pi/2;
ang2=transpose(!PI*findgen(n)/(n-1)-!PI/2)
theta_e=transpose(ang2##ones)
r_o=fltarr(n,n)+1
rcart=1+cos(!pi/2-theta_e)
;rcart=r_o
theta_a_1d=reform(theta_a,1,n*n)
theta_e_1d=reform(theta_e,1,n*n)
rcart_1d=reform(rcart,1,n*n)
sph=[theta_a_1d,theta_e_1d,rcart_1d]
rectc=cv_coord(from_sphere=sph, /to_rect)
x_1d=rectc(0,*)
y_1d=rectc(1,*)
z_1d=rectc(2,*)
x=reform(x_1d,n,n)
y=reform(y_1d,n,n)
z=reform(z_1d,n,n)
fid=global_data.window1_id
wset, fid
surface, z,x,y;, AX=45, AZ=45
;shade_surf, z;,x,y;AX=-30, AZ=-45
;xplot_sl, x,y,z
;fsc_surface, z,x,y
end
'Quit': begin
print,'Got into Quit Case'
widget_control, event.top, /destroy
end
else:
endcase
end
;----------------------------------------------------------- --------------------------------------
pro ant_test
fund_base=widget_base(column=2, title="Plotting test GUI")
N=256
draw1=widget_draw(fund_base, xsize=N, ysize=N, uname="window1")
pushit=widget_button(fund_base, value="Push to Plot",
uname="pushit")
quit_button=widget_button(fund_base, value="Quit", uname="Quit")
widget_control, fund_base, /realize
widget_control, draw1, get_value=window1
global_data={draw1_id: draw1, window1_id: window1 }
widget_control, fund_base, set_uvalue=global_data
xmanager, 'ant_test', fund_base
event_handler='ant_test_event'
end
|
|
|
|
| Re: Help plotting a 3D Carioid... [message #47591 is a reply to message #47498] |
Wed, 15 February 2006 09:34 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Steve writes:
> Additionally, the surface plot just doesn't have a nice look - is there
> anything a bit nicer I could do, along the lines of xplot3D, but be
> able to do it in a pre-specified window (I'm going to be having the
> result plot in a GUI)? Best would be to have the surface shaded and
> with the color set to correspond with the magnitude of each point.
>
> I guess what I'm looking for is a little guidance on how to get this to
> look a bit nicer, and what's causing the hole when I use surface...
Hard to say what is happening to SURFACE. I note that
rotating the surface slightly gets around the problem:
surface, z, x, y, az=45
This is probably an artifact from the fact that it is
really a 2.5D representation of a surface. But, it looks
like a bug, for sure.
XPlot3D uses an object graphics surface (IDgrSurface), rather
than a direct graphics surface. This is a true 3D surface,
so I am not surprised this works correctly. You can
put an object graphics surface into an object graphics
window, but the whole process is quite a bit more
trouble than it is with direct graphics. :-)
If this is something you want to pursue, I'd recommend
getting ahold of Ronn Kling's book, Power Graphics in IDL.
The chances of learning to use object graphics from
the IDL documentation is slim and none, I believe, unless
you have some graphics experience from somewhere else.
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
