| Re: Stupid question regarding BUFFER keyword in NG output..... [message #73862] |
Wed, 08 December 2010 02:43  |
lecacheux.alain
Messages: 325
Registered: January 2008
|
Senior Member |
|
|
On 8 déc, 00:57, Chris Torrence <gorth...@gmail.com> wrote:
> Hi Paul,
>
> Well, you have a couple of different options. You could use the
> Refresh method to disable drawing to the window. It will still create
> the window, but you won't see any changes to the plot until you call
> Refresh again.
>
> Another crazy way is to use the CopyWindow method to get a screen
> dump, and then throw it into direct graphics:
>
> p = plot(/test,/buffer)
> tv,p.copywindow(),/true
>
> Here's another one for fun:
> p = plot(/test,/buffer)
> for i=0,100 do begin & p.color = !color.(i) & tv,p.copywindow(),/true
>
> Hope this helps.
> Cheers,
> Chris
> ITTVIS
But if I am doing :
p = plot(/test)
p2 = plot(/test,/buffer)
q = image(p2.copyWindow())
the plot referenced by q is different (looks like undersampled) from
the plot referenced by p.
Why ?
alx.
|
|
|
|
| Re: Stupid question regarding BUFFER keyword in NG output..... [message #73863 is a reply to message #73862] |
Tue, 07 December 2010 15:57  |
chris_torrence@NOSPAM
Messages: 528
Registered: March 2007
|
Senior Member |
|
|
Hi Paul,
Well, you have a couple of different options. You could use the
Refresh method to disable drawing to the window. It will still create
the window, but you won't see any changes to the plot until you call
Refresh again.
Another crazy way is to use the CopyWindow method to get a screen
dump, and then throw it into direct graphics:
p = plot(/test,/buffer)
tv,p.copywindow(),/true
Here's another one for fun:
p = plot(/test,/buffer)
for i=0,100 do begin & p.color = !color.(i) & tv,p.copywindow(),/true
Hope this helps.
Cheers,
Chris
ITTVIS
|
|
|
|
| Re: Stupid question regarding BUFFER keyword in NG output..... [message #73869 is a reply to message #73863] |
Tue, 07 December 2010 09:00 |
Michael Galloy
Messages: 1114
Registered: April 2006
|
Senior Member |
|
|
On 12/7/10 7:38 am, Paul van Delst wrote:
> Hi Mike,
>
> Michael Galloy wrote:
>>
>> I believe plots made in a buffer are meant to produce some type of
>> output and then be destroyed, i.e., like
>>
>> IDL> p = plot(/test, /buffer)
>> IDL> p->save, 'test.png'
>> IDL> obj_destroy, p
>
> Nutz! I was hoping there was some sort of undocumented method that displayed the plots, e.g.
> p.display
> (which I tried by the way :o) Sort of like doing DG to a pixmap and then sticking it in a window.
Not that I know of, but that doesn't mean it's not in there somewhere.
Mike
--
www.michaelgalloy.com
Research Mathematician
Tech-X Corporation
|
|
|
|
|
|
| Re: Stupid question regarding BUFFER keyword in NG output..... [message #73876 is a reply to message #73871] |
Mon, 06 December 2010 17:00 |
Michael Galloy
Messages: 1114
Registered: April 2006
|
Senior Member |
|
|
On 12/6/10 2:37 PM, Paul van Delst wrote:
> Hello,
>
> I just created a bunch of plots like so:
>
> p = PLOT( f, r, $
> TITLE=sensor_id + channel_string + ' threshold cutoff discrepancy', $
> XTITLE='Frequency (cm!U-1!N)', $
> YTITLE='Relative Response', $
> SYMBOL='diamond', $
> BUFFER=Buffer )
>
> I returned all the created plot references (lots of 'em in a list of lists in a hash) so I can modify and output if
> necessary. And they appear to contain actual plots:
>
> IDL> p=((tref[5])[23])[0]
> IDL> help, p
> P PLOT<323024>
>
> But, and here is the stupid question, how do I now display these buffer-ised plots?
>
> Thanks for any info.
>
> cheers,
>
> paulv
I believe plots made in a buffer are meant to produce some type of
output and then be destroyed, i.e., like
IDL> p = plot(/test, /buffer)
IDL> p->save, 'test.png'
IDL> obj_destroy, p
Mike
--
www.michaelgalloy.com
Research Mathematician
Tech-X Corporation
|
|
|
|
| Re: Stupid question regarding BUFFER keyword in NG output..... [message #73956 is a reply to message #73863] |
Wed, 08 December 2010 06:54 |
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
Hi Chris,
Chris Torrence wrote:
> Hi Paul,
>
> Well, you have a couple of different options. You could use the
> Refresh method to disable drawing to the window. It will still create
> the window, but you won't see any changes to the plot until you call
> Refresh again.
That's how I did it at first[*], but since I am creating, let's see, 5x32+16x16=416 plots I don't want to create 416
windows on my desktop. I figured I would create them under the hood, return the references in a hash, and then pick and
choose which one to display/modify based on other info from the processing.
> Another crazy way is to use the CopyWindow method to get a screen
> dump, and then throw it into direct graphics:
>
> p = plot(/test,/buffer)
> tv,p.copywindow(),/true
>
> Here's another one for fun:
> p = plot(/test,/buffer)
> for i=0,100 do begin & p.color = !color.(i) & tv,p.copywindow(),/true
While that is a pretty cool technique, I would need to have "regular" access to the plots I choose to display (e.g.
annotate them, add legends, zoom into regions, save them to file etc).
What are the odds for a method to display buffer-ised graphics in a future IDL release?
cheers,
paulv
[*] See a previous post of mine where I was turned onto this technique (using refresh) after I whinged about the
slowness of NG graphics using the regular overplot technique shown in the plot example docs. And the difference in speed
is like night and day. The last two days I've been creating all manner of plots using NG and using them to easily
inspect data, annotate plots, and save them for inclusion in documents. I haven't used DG graphics for a few weeks now.
Oh, and I still think better PS/EPS output support for NG graphics is crucial in future IDL releases -- I have been
playing with the PNG or PDF output options in NG and they just do not cut it quality-wise when the plot has to look
*good* both on-screen (at, say 250% zoom) and in a printout. For publication-quality (including being scaleable)
graphics, I still think DG PS output trumps NG (any method).
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
