[obs-tap] Spectral axis - meters

[Anita M. S. Richards]

>
> I do not understand what is the meaning of B2.2. (Spectral Bounds):
> "Such values (it is meant spectral intervals) are expressed as frequencies 
> but using meters as units as it is easily convertible"
>
> In stellar astronomy tehe spectral interval is in A or nm - and after some 
> experiences with  SSA clients (e.g. in BAND=4583e-10/6600e-10) the meters are 
> already well accepted by the community.
>
> So the above statement referes mostly to radioastronomy (probably) as the 
> X-ray community preferes keV (energy).
>
> The Arnold's suggestion to keep Hz I feel strongly radio-centric.
> IF we have to keep one units (BTW WHY ???? I would put a reference to section 
> 17 (Spectral Bounds)) and leave the words about frequencies and easy 
> convertibility .
>
> Just comment to the meter/Hz issue - in simplified terminology about EM 
> spectrum the wavelength is the more general and easily understood concept - 
> even for radioastronomy (we have submm range and decimeter and decameter 
> ranges and we have microwaves and milimeter wave bands etc ...
> Its dificult to find reference to observation in THz band or something like 
> this....

The problem is that Hz (or energy) and m have an inverse relationship. 
Hence, if you have a bandwidth which is a large fraction of the observing 
wavelength/frequency, the spectral resolution is linear on one unit but 
changes appreciably across the bandwidth in the other unit. For example, 
radio correlators produce channels evenly spaced in frequency.  The 
resolution in wavelength at the lower end of a band e.g. at 4 GHz will be 
very different from that at 6 GHz, for a 2-GHz band.  The same would be 
true for wide X-ray bands, or in the inverse example for wide optical 
bands if the spectral resolution was evenly spaced in metres.

In other words, it is possible with trivial linear unit conversion to go 
from Angstrom to m to nm etc., and similarly for Hz/eV.  It is possible to 
express characterization of the coarse-level quantities in any unit, i.e. 
upper and lower bounds, representative wavelength, spectral resolution..

It is not possible to give a single precise value for e.g. spectral 
resolution in wavelength units, for a radio spectrum.

So, at a coarse level we can adopt one unit but for tools which require 
precision we need Hz and m

thanks

Anita