| Re: REBIN needs explaining [message #41921] |
Wed, 01 December 2004 04:38 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
b.hinrichsen@fkf.mpg.de writes:
> Interpolate considers these values as the new x-array which determines
> the values of the y-array. The original x-array is implicitly
> understood to be the index of A viz. [0,1,2,3].
Ah, the *X* values! Alles klar. Danke. :-)
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: REBIN needs explaining [message #41924 is a reply to message #41921] |
Wed, 01 December 2004 00:45  |
Paolo Grigis
Messages: 171
Registered: December 2003
|
Senior Member |
|
|
David Fanning wrote:
> JD Smith writes:
>
>
>> So here it has just 1 interval producing 10 points, and 10 points are
>> duplicated from the last value. The manual also recommends the
>> solution:
>>
>> If this effect is undesirable, use the INTERPOLATE function.
>
>
> Well, I *would* use the INTERPOLATE function if I could
> understand it. :-(
>
> The REBIN documentation says this:
>
> *************************************************
> ; A four point vector:
> A = [0, 10, 20, 30]
>
> ; Expand by a factor of 3:
> B = REBIN(A, 12)
>
> PRINT, B
> IDL prints:
>
> 0 3 6 10 13 16 20 23 26 30 30 30
>
> Note that the last element is repeated three times. If this effect is
> undesirable, use the INTERPOLATE function. For example, to produce 12
> equally spaced interpolates from the interval 0 to 30:
>
> B = INTERPOLATE(A, 3./11. * FINDGEN(12))
> PRINT, B
> IDL prints:
>
> 0 2 5 8 10 13 16 19 21 24 27 30
> *************************************************
>
> But,
>
> IDL> Print, 3./11. * FINDGEN(12))
> 0.000000 0.272727 0.545455 0.818182 1.09091
> 1.36364 1.63636 1.90909 2.18182 2.45455 2.72727
> 3.00000
>
> How in the world to *those* numbers, when applied to A, produce
> the result I get? Seems like magic to me. :-(
They say, a picture is worth a thousand words, therefore the
following plot should be able to enlighten the meaning of INTERPOLATE
better than a lengthy explanation...
x=findgen(4)
a=[0.,10,20,30]
z=3./11. * FINDGEN(12)
plot,x,a,psym=-4,xrange=[-0.5,3.5],yrange=[-5,35]
oplot,z,interpolate(a,z),psym=1
Ciao,
Paolo
>
> Cheers,
>
> David
>
--
____________________________________________________________ ________
Paolo Grigis
ETHZ - Institute of Astronomy email: pgrigis@astro.phys.ethz.ch
Scheuchzerstrasse 7
ETH Zentrum phone: ++41 1 632 42 20
8092 Zurich fax : ++41 1 632 12 05
Switzerland http://www.astro.phys.ethz.ch/
____________________________________________________________ ________
|
|
|
|
| Re: REBIN needs explaining [message #41925 is a reply to message #41924] |
Wed, 01 December 2004 00:38 |
b.hinrichsen
Messages: 2
Registered: January 2004
|
Junior Member |
|
|
David Fanning <davidf@dfanning.com> wrote in message news:<MPG.1c16c3fd524557b998988c@news.frii.com>...
> JD Smith writes:
>
>> So here it has just 1 interval producing 10 points, and 10 points are
>> duplicated from the last value. The manual also recommends the
>> solution:
>>
>> If this effect is undesirable, use the INTERPOLATE function.
>
> Well, I *would* use the INTERPOLATE function if I could
> understand it. :-(
>
> The REBIN documentation says this:
>
> *************************************************
> ; A four point vector:
> A = [0, 10, 20, 30]
>
> ; Expand by a factor of 3:
> B = REBIN(A, 12)
>
> PRINT, B
> IDL prints:
>
> 0 3 6 10 13 16 20 23 26 30 30 30
>
> Note that the last element is repeated three times. If this effect is
> undesirable, use the INTERPOLATE function. For example, to produce 12
> equally spaced interpolates from the interval 0 to 30:
>
> B = INTERPOLATE(A, 3./11. * FINDGEN(12))
> PRINT, B
> IDL prints:
>
> 0 2 5 8 10 13 16 19 21 24 27 30
> *************************************************
>
> But,
>
> IDL> Print, 3./11. * FINDGEN(12))
> 0.000000 0.272727 0.545455 0.818182 1.09091
> 1.36364 1.63636 1.90909 2.18182 2.45455 2.72727
> 3.00000
Interpolate considers these values as the new x-array which determines
the values of the y-array. The original x-array is implicitly
understood to be the index of A viz. [0,1,2,3].
>
> How in the world to *those* numbers, when applied to A, produce
> the result I get? Seems like magic to me. :-(
>
> Cheers,
>
> David
Greetings,
Bernd
|
|
|
|
| Re: REBIN needs explaining [message #41926 is a reply to message #41925] |
Wed, 01 December 2004 00:23 |
Chris Lee
Messages: 101
Registered: August 2003
|
Senior Member |
|
|
In article <MPG.1c16c3fd524557b998988c@news.frii.com>, "David Fanning"
<davidf@dfanning.com> wrote:
...
> INTERPOLATE(A, 3./11. * FINDGEN(12))
> PRINT, B
> IDL prints:
> 0 2 5 8 10 13 16 19 21 24 27 30
> ************************************************* But,
>
> IDL> Print, 3./11. * FINDGEN(12))
> 0.000000 0.272727 0.545455 0.818182 1.09091
> 1.36364 1.63636 1.90909 2.18182 2.45455 2.72727
> 3.00000
> How in the world to *those* numbers, when applied to A, produce the
> result I get? Seems like magic to me. :-( Cheers,
> David
>
It looks like IDL is flooring the numbers (or just dropping the floating
point part) to make them INTs.
IDL> a=[0,10,20,30]
IDL> print, interpolate(a,findgen(12)*3/11.)
0 2 5 8 10 13 16 19 21
24 27 30
IDL> print, interpolate(a*1.0,findgen(12)*3/11.)
0.00000 2.72727 5.45455 8.18182 10.9091 13.6364
16.3636 19.0909 21.8182 24.5455 27.2727 30.0000
IDL> print, interpol(a, 12)
0.00000 2.72727 5.45455 8.18182 10.9091 13.6364
16.3636 19.0909 21.8182 24.5455 27.2727 30.0000
Chris.
|
|
|
|
| Re: REBIN needs explaining [message #41927 is a reply to message #41926] |
Tue, 30 November 2004 16:53 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
JD Smith writes:
> So here it has just 1 interval producing 10 points, and 10 points are
> duplicated from the last value. The manual also recommends the
> solution:
>
> If this effect is undesirable, use the INTERPOLATE function.
Well, I *would* use the INTERPOLATE function if I could
understand it. :-(
The REBIN documentation says this:
*************************************************
; A four point vector:
A = [0, 10, 20, 30]
; Expand by a factor of 3:
B = REBIN(A, 12)
PRINT, B
IDL prints:
0 3 6 10 13 16 20 23 26 30 30 30
Note that the last element is repeated three times. If this effect is
undesirable, use the INTERPOLATE function. For example, to produce 12
equally spaced interpolates from the interval 0 to 30:
B = INTERPOLATE(A, 3./11. * FINDGEN(12))
PRINT, B
IDL prints:
0 2 5 8 10 13 16 19 21 24 27 30
*************************************************
But,
IDL> Print, 3./11. * FINDGEN(12))
0.000000 0.272727 0.545455 0.818182 1.09091
1.36364 1.63636 1.90909 2.18182 2.45455 2.72727
3.00000
How in the world to *those* numbers, when applied to A, produce
the result I get? Seems like magic to me. :-(
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: REBIN needs explaining [message #41928 is a reply to message #41927] |
Tue, 30 November 2004 15:56 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Tue, 30 Nov 2004 16:32:03 -0700, David Fanning wrote:
> Hi Folks,
>
> Can someone (Craig?) explain this to me.
>
> Consider:
>
> v = [0, 10]
>
> I understand this:
>
> Print, rebin(v, 20, /Sample)
> 0 0 0 0 0 0 0 0 0
> 0 10 10 10 10 10 10 10 10
> 10 10
>
> What I don't understand is this:
>
> Print, rebin(v, 20)
> 0 1 2 3 4 5 6 7 8
> 9 10 10 10 10 10 10 10 10
> 10 10
>
> Aren't there an awful lot of 10s in there!? Why doesn't this come
> out like this:
>
> 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8
> 9 9 10 10
>
> I notice REBIN, CONGRID, and even CMCONGRID does this same thing. Why?
From the manual:
When expanding an array, REBIN interpolates, it never
extrapolates. Each of the n-1 intervals in the n-element input array
produces m/n interpolates in the m-element output array. The last m/n
points of the result are obtained by duplicating element n-1 of the
input array because their interpolates would lie outside the input
array.
So here it has just 1 interval producing 10 points, and 10 points are
duplicated from the last value. The manual also recommends the
solution:
If this effect is undesirable, use the INTERPOLATE function.
A good question is why they consider the need to avoid extrapolation
by duplicating elements, rather than just interpolating on all
internal intervals. I'm not sure for what problems the current
behavior would be desirable.
JD
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
