Question on order merging with echelle.scombine

John wrote on Dec 02, 2008

Hi there,

i have a question on the order merging process when using
echelle.scombine.

Is it possible to apply weighting in the merging process based on an adjacent sigma spectrum? Suppose i have a normalised echelle spectrum with an associated sigma image that holds the noise characteristics for each pixel (or wavelength bin). Can scombine use the noise info from the sigma image to properly apply weights on each individual wavelength bin in the overlapping part of the orders?
If not, is there another IRAF task that can do this?

The scombine help page is not 100% clear to me in this respect.

thanks,
John

---

Francisco Valdes wrote on Dec 02, 2008

What you are interested in would be a good thing. However, none of the current crop of NOAO combining task (scombine, imcombine, etc.) provide weighting on a per pixel basis (i.e. with a weight array). The only way I can think of is to use imexpr to make your own specific expression for the combined pixels. The difficulty is that this task doesn't work in wavelength space like scombine. So you would have to use a spectral task, such as dispcor, to put all the orders onto a common pixel scale. You would also need to take care of all the other things that go into combining within the expression (scaling, masking, etc.).

I'm sorry this is not easier.

Yours,
Frank Valdes

---

Last post on Dec 02, 2008