search

desihub / fastspecfit

Fast spectral synthesis and emission-line fitting of DESI spectra.
https://fastspecfit.readthedocs.org
BSD 3-Clause "New" or "Revised" License
13 stars 2 forks source link

improve velocity dispersion estimates #67

Closed moustakas closed 2 years ago

moustakas commented 2 years ago

One medium-term goal is to improve the velocity dispersion estimates. But first we need to fix issues like this, which suggest that velocity dispersion isn't being handled correctly at all. In detail, with the nominal vdisp=150 km/s, the Ca H+K lines in the model are clearly over-broadened. However, the implied broadening is more like 20-30 km/s, which can't be right for a relatively luminous & massive system like this. So, I need to check that I'm doing the broadening correctly in the first place.

fastspec /global/cfs/cdirs/desi/spectro/redux/fuji/tiles/cumulative/80605/20210205/redrock-3-80605-thru20210205.fits  \
  -o fastspec.fits --targetids 39627676662630998

fastspec-80605-cumulative-39627676662630998

moustakas commented 2 years ago

I'm working on this ticket here by comparing my vdisp estimates against those derived by Khaled Soliman using pPXF-- https://data.desi.lbl.gov/desi/users/ioannis/fastspecfit/pv-vdisp/fujilupe/qa/

moustakas commented 2 years ago

Lots of bugs (both algorithmic and in the QA code) now fixed in #69. Also, vdisp is now computed if solve_vdisp=True or if: (S/N_B>3 or S/N_R>3) and (redshift < 1), where S/N_B and S/N_R is the signal-to-noise ratio per pixel in the B and R camera, respectively.