search

CIRADA-Tools / RM-Tools

RM-synthesis, RM-clean and QU-fitting on polarised radio spectra
MIT License
42 stars 23 forks source link

rmsynth3d imaginary FDF files have wrong Stokes axis values #127

Closed Cameron-Van-Eck closed 4 months ago

Cameron-Van-Eck commented 4 months ago

I realized that both the real and imaginary dirty FDF files have CRVAL4 = 2 (for the Stokes axis, indicating Stokes Q). That's not correct, and should be fixed (especially for products going into the CADC archive). Imaginary/Stokes U cubes should have Stokes axis CRVAL = 3.

5 bucks says I copied the Stokes Q/real header and re-used it for Stokes U without thinking about it. Not sure yet if rmclean3d is also affected, but good chance that it is.

I don't have time in the next month or so to fix this, but I'll try to get to it after that if no one else has time to take a look in the meantime.

Cameron

ErikOsinga commented 4 months ago

Are you sure that's the case? The dirt FDF cubes that I have running the most recent version of the pol pipeline show CRVAL3=2 for the real part and CRVAL3=3 for the imaginary part.

(Note that it's also CRVAL3 instead of the 4 you mention?)

Cameron-Van-Eck commented 4 months ago

Ah, I already fixed this (6 months ago!). My bad. I looked at the Pilot 1 pipeline products after our discussion with Sharon, and they have this mistake in the FITS headers. So ideally we should correct those before ingesting the pilot 1 products. But RM-Tools was already fixed, so this issue is redundant. Whoops!