splot wavelength calibration

heather wrote on May 10, 2011

I'm looking at spectra which were originally SDSS stellar spectra, rebinned to a linear wavelength scale by IDL code in the SSPP (segue pipeline)

Looking at the header of the fits file, CRVAL1 = 3850 and CD1_1 = 1. (both Angstroms)

When I plot the data in splot, the first pixel has wavelength 3851 A. And all subsequent pixels have wavelengths one A more than they should.

I'm using IRAF 2.14.1 on a Linux box running Redhat Enterprise 5 (64 bit)

The fits file can be found at
http://astronomy.case.edu/heather/boss.lineindex.fit (spectra in HDU 0) and a plot of the wavelength of the first pixel in splot is at
http://astronomy.case.edu/heather/4iraf.jpg

Any suggestions would be much appreciated.

thanks
Heather Morrison

---

Mike Fitzpatrick wrote on May 10, 2011

Try adding a CRPIX1 = 1 keyword to the headers to specify the starting pixel. Otherwise the value is assumed to be zero and since the image index starts at one .......

Frank may have additional comments about a better set of keywords to add.

---

heather wrote on May 10, 2011

That fixed it. Thanks very much for the quick reply, Mike.

I'll tell the SSPP folks that they should add that keyword to the fits files they produce.

Heather

---

Last post on May 10, 2011