Hi Jonathan -
I agree that this would be useful for photometry points; it is just that it appeared to be suggested for spectra as well, and if so it would be good to examine more carefully whether it is needed or desirable for spectra.
For filter transmission curves, one thing we should look at is the JHU filter profile service; see http://www.voservices.net/filter/
It could be useful to expose filter transmission curves via an SSA interface, so that we don't need something separate for filter data, but we should first see what metadata is required to describe filters. I haven't looked carefully at the JHU filter database, but I see that it contains data for about 100 standard filters.
On Sun, 22 Oct 2006, Jonathan McDowell wrote:
> Doug,
> I don't now recall who requested this at the Tucson meeting.
> The idea was that the Spectrum model is also very close to being
> a model for photometry, and especially for SEDs we would like to
> be able to mix spectra and photometry points.
> However, photometry observers like to report their data
> in magnitudes, even when it is absolutely calibrated. They want
> to say:
> I measured a J magnitude of 11.8, and to convert my J to Jy
> (ignoring color term effects) you do F = F0 * 10**(-0.4 * J)
> where F0 is given in the header
> rather than calculating F themselves and putting that directly
> in the data.
> Of course you'd like to point to the whole transmission curve
> yourself; then you can calculate the flux F from the magnitude J
> making your *own* (client side) assumptions about the object spectrum.
> Hence the URI.
> I guess the real science justification is that J is a calibrated
> value; F is more than that, it's a modelled flux making an assumption
> about the source spectrum. So better to publish J than F as long as
> you can get F easily.
> It seems a small thing to add to the model, and wins us a lot of
> archival photometry. We discussed this around the table in Tucson
> and you didn't seem bothered then...
>
> Jonathan
>
Received on 2006-10-23Z04:02:00