From dobos at pha.jhu.edu  Sat Nov 12 14:34:31 2011
From: dobos at pha.jhu.edu (Laszlo Dobos)
Date: Sat, 12 Nov 2011 23:34:31 +0100
Subject: Characterization question
Message-ID: <005301cca18b$42c562e0$c85028a0$@jhu.edu>

Hi group,

 

How do I specify overlapping axes in characterisation?

 

Let's say I have a spectrum and want to give the spectral axis in both
angstroms and velocities. Do I just add two axes and mark them both SPECTRAL
and assume that they overlap?

 

What if I have an image with two spatial axes? Do I add two SPATIAL axes
then or just one and assume that the resolution etc. is the same in both RA
and Dec?

 

What if I have overlapping spatial axes, let's say ra,dec and (for a not too
good example) also Mpc? How do I associate one Mpc axis with Ra and the
other with Dec.

 

I'm sorry, if I misunderstood something about the whole thing. (Samples on
the CDS server are unfortunately not available).

 

What I'm actually working on is a generic toolkit to characterize data cubes
with any number of dimensions, and I want to allow using multiple scales
along the axes.

 

 

Thanks,

 

 

-Laszlo

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/dm/attachments/20111112/85fe110b/attachment.html>

From t.jenness at jach.hawaii.edu  Sat Nov 12 16:44:12 2011
From: t.jenness at jach.hawaii.edu (Tim Jenness)
Date: Sat, 12 Nov 2011 14:44:12 -1000
Subject: Characterization question
In-Reply-To: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
References: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
Message-ID: <CA+G92Rfr2SLheDpssBH_VXnHzeiVW38teMnxSe_KusZ67TeRGA@mail.gmail.com>

2011/11/12 Laszlo Dobos <dobos at pha.jhu.edu>

> Hi group,****
>
> ** **
>
> How do I specify overlapping axes in characterisation?****
>
> ** **
>
> Let's say I have a spectrum and want to give the spectral axis in both
> angstroms and velocities. Do I just add two axes and mark them both
> SPECTRAL and assume that they overlap?****
>
> **
>

You should start by taking a look at the FITS formalisms for world
coordinate systems. Paper III covers spectral coordinates and that shows
how you can specify a single wavelength frame that includes details of how
to convert to velocity.

The FITS papers also show to specify multiple coordinate systems but I'm
not suire how this is implemented in the IVOA data model.


> **
>
> What if I have an image with two spatial axes? Do I add two SPATIAL axes
> then or just one and assume that the resolution etc. is the same in both RA
> and Dec?****
>
> ** **
>
> What if I have overlapping spatial axes, let's say ra,dec and (for a not
> too good example) also Mpc? How do I associate one Mpc axis with Ra and the
> other with Dec.****
>
> ** **
>
> I'm sorry, if I misunderstood something about the whole thing. (Samples on
> the CDS server are unfortunately not available).****
>
> ** **
>
> What I'm actually working on is a generic toolkit to characterize data
> cubes with any number of dimensions, and I want to allow using multiple
> scales along the axes.****
>
>
>
The Starlink GAIA and SPLAT software are applications that know how to
convert from wavelength to velocity (and many other schemes).

-- 
Tim Jenness
Joint Astronomy Centre
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/dm/attachments/20111112/13ee2a23/attachment.html>

From arots at head.cfa.harvard.edu  Mon Nov 14 06:09:03 2011
From: arots at head.cfa.harvard.edu (Arnold Rots)
Date: Mon, 14 Nov 2011 09:09:03 -0500 (EST)
Subject: Characterization question
In-Reply-To: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
Message-ID: <201111141409.pAEE93Vr015704@xebec.cfa.harvard.edu>

Actually, in the first example you are talking about two separate
axes: Spectral and Redshift. We treat them separately.

The spatial axes are 1, 2, or 3-D.

It's all laid out in the STC standard.

  - Arnold

Laszlo Dobos wrote:
[ Charset ISO-8859-2 unsupported, converting... ]
> Hi group,
> 
>  
> 
> How do I specify overlapping axes in characterisation?
> 
>  
> 
> Let's say I have a spectrum and want to give the spectral axis in both
> angstroms and velocities. Do I just add two axes and mark them both SPECTRAL
> and assume that they overlap?
> 
>  
> 
> What if I have an image with two spatial axes? Do I add two SPATIAL axes
> then or just one and assume that the resolution etc. is the same in both RA
> and Dec?
> 
>  
> 
> What if I have overlapping spatial axes, let's say ra,dec and (for a not too
> good example) also Mpc? How do I associate one Mpc axis with Ra and the
> other with Dec.
> 
>  
> 
> I'm sorry, if I misunderstood something about the whole thing. (Samples on
> the CDS server are unfortunately not available).
> 
>  
> 
> What I'm actually working on is a generic toolkit to characterize data cubes
> with any number of dimensions, and I want to allow using multiple scales
> along the axes.
> 
>  
> 
>  
> 
> Thanks,
> 
>  
> 
>  
> 
> -Laszlo
> 
--------------------------------------------------------------------------
Arnold H. Rots                                Chandra X-ray Science Center
Smithsonian Astrophysical Observatory                tel:  +1 617 496 7701
60 Garden Street, MS 67                              fax:  +1 617 495 7356
Cambridge, MA 02138                             arots at head.cfa.harvard.edu
USA                                     http://hea-www.harvard.edu/~arots/
--------------------------------------------------------------------------

From mireille.louys at unistra.fr  Tue Nov 15 09:50:26 2011
From: mireille.louys at unistra.fr (Mireille Louys)
Date: Tue, 15 Nov 2011 18:50:26 +0100
Subject: Characterization question
In-Reply-To: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
References: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
Message-ID: <20111115185026.83fkbznzsw80wokw@webmail.u-strasbg.fr>

Hi Laszlo,

As Arnold already answered, you mention an example with two axes on  
spectral content, distinguished in SpectralAxis ( Ansgtrom) and  
RedShiftAxis (vel).
The Characterisation DM partly re-uses the STC definitions.
Characterisation v2.0 covers the description of RedShiftAxis (on progress).
I will be glad to work on an example of your metadata to check the mapping.
more below...

Laszlo Dobos <dobos at pha.jhu.edu> a ?crit?:

> Hi group,
>
>
>
> How do I specify overlapping axes in characterisation?
>
>
>
> Let's say I have a spectrum and want to give the spectral axis in both
> angstroms and velocities. Do I just add two axes and mark them both SPECTRAL
> and assume that they overlap?
>
>
>
> What if I have an image with two spatial axes? Do I add two SPATIAL axes
> then or just one and assume that the resolution etc. is the same in both RA
> and Dec?
>
>
CharDM considers the Spatial axis is 2D or 3D . As in STC, resolution  
will have the dimentionality of the Coordinate : 2D or 3D.
>
> What if I have overlapping spatial axes, let's say ra,dec and (for a not too
> good example) also Mpc? How do I associate one Mpc axis with Ra and the
> other with Dec.
>
CharDM V1 and 2 consider the same units on all spatial dimensions.
Interesting to consider your use-case.
>
>
> I'm sorry, if I misunderstood something about the whole thing. (Samples on
> the CDS server are unfortunately not available).
>
Sorry. Thanks for the feedback , will fix this.
>
>
> What I'm actually working on is a generic toolkit to characterize data cubes
> with any number of dimensions, and I want to allow using multiple scales
> along the axes.
>
Great! What about circulating an example of your metadata so that we  
could check how it fits to what we have in existing data models and  
what can evolve.

Many thanks, Mireille.




From dtody at nrao.edu  Sun Nov 27 14:16:13 2011
From: dtody at nrao.edu (Douglas Tody)
Date: Sun, 27 Nov 2011 15:16:13 -0700 (MST)
Subject: Characterization question
In-Reply-To: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
References: <005301cca18b$42c562e0$c85028a0$@jhu.edu>
Message-ID: <alpine.OSX.1.00.1111271504360.388@colorado.tuc.noao.edu>

As Tim J. also mentioned, we should not forget the existing FITS WCS
model which is relevant for parts of this, e.g. associating image axes
with world coordinate systems, or associating two image axes with a
spatial coordinate system.  In general we should distinguish between
higher level characterization of a dataset (Char), and actually defining
the precise WCS for the physical axes of the dataset.  Char for example
can characterize the spectral attributes of a dataset but it is a WCS
model which specifies which physical image axis is the logical spectral
axis, and the exact mapping to spectral coordinates.  There is a first
cut at this in the SIAV2 draft (which addresses cube data).  We retain
and use Characterization but also add a "Mapping" model which is based
directly on FITS WCS but which is defined as an abstract model
independent of serialization.

 	- Doug


On Sat, 12 Nov 2011, Laszlo Dobos wrote:

> Hi group,
>
>
>
> How do I specify overlapping axes in characterisation?
>
>
>
> Let's say I have a spectrum and want to give the spectral axis in both
> angstroms and velocities. Do I just add two axes and mark them both SPECTRAL
> and assume that they overlap?
>
>
>
> What if I have an image with two spatial axes? Do I add two SPATIAL axes
> then or just one and assume that the resolution etc. is the same in both RA
> and Dec?
>
>
>
> What if I have overlapping spatial axes, let's say ra,dec and (for a not too
> good example) also Mpc? How do I associate one Mpc axis with Ra and the
> other with Dec.
>
>
>
> I'm sorry, if I misunderstood something about the whole thing. (Samples on
> the CDS server are unfortunately not available).
>
>
>
> What I'm actually working on is a generic toolkit to characterize data cubes
> with any number of dimensions, and I want to allow using multiple scales
> along the axes.
>
>
>
>
>
> Thanks,
>
>
>
>
>
> -Laszlo
>
>