Re: Noise values in latest DRP outputs

[Nick Konidaris]

Hi Karl,

Thank you for your email. I actually have been quietly working (hard) on this problem. There are at least several bugs in the way the pipeline currently handles these outputs. My hope is that the next release will squash these bugs!

~n

On Thu, Apr 3, 2014 at 7:00 PM, Karl Glazebrook <kglaz...@swin.edu.au> wrote:

Hi Nick,

I just wanted to follow-up on the issue of the noise (sd) arrays in the latest version of the DRP that Themiya has been talking to you about. In particular I have been looking myself closely at a test case myself and I have some data that might help you,

I show a screenshot of a 120s single pair (A&B) reduced using the latest 14 Mar DRP to make an eps_*.fits file (top left DS9 panel). This is the usual (A-B) + (B-A)*  (*=shifted). The star has peak counts of (-4, +8, -4) electrons, which is consistent with the counts in the raw frame converted to electrons and divided by the exposure time. (220 ADU -> 4e/s)

The RMS in the blank region near the star defined by the green box is ~ 0.13 e/s. In the raw A frame (bottom right) the median in the same box (matched approximately as it is not rebinned) is 42 ADU. I can then predict the RMS in the eps_ frame using the usual poisson + RN eqn:

sqrt(42*2.15+ 5.8**2) * sqrt(4) /120    = 0.185547238669234

which is pretty close. (2.15 = e/ADU, 5.8 = RN, sqrt(4) for 4 effective frames combined and 120s exposure)

HOWEVER the median in the sd_*.fits frame (top right) in the same box is 10.7 (whatever units?) which is out by a factor of 100 compared to the actual RMS.  This is suspiciously close to the intergration time so I wonder if the code has forgotten to normalise this output? 

hope this is useful, can supply the files if you want

Karl

----------------
Karl Glazebrook
Director & Distinguished Professor
Centre for Astrophysics & Supercomputing, Swinburne University of Technology
Contact: +61-3-9214-4384  kglaz...@swin.edu.au
astronomy.swin.edu.au/karl galacticturmoil.org @karlglazebrook

--
You received this message because you are subscribed to the Google Groups "MOSFIRE Data Reduction PIpeline" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mosfire-drp...@googlegroups.com.
To post to this group, send email to mosfi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mosfire-drp/9F708B05-F636-49C6-97A0-9EBC3749F2D5%40swin.edu.au.
For more options, visit https://groups.google.com/d/optout.

--
   +1 831 704 6425

[Karl Glazebrook]

Thanks Nick.Happy to test.

[Nick Konidaris]

Hi Karl + crew,

Thank you for posting all your bug reports, and helping me debugging the pipeline. The most recent version of the pipeline addresses these changes.

~nick