Optimum Rejection Algorithm for Combining
Elden Paul Iverson wrote on Mar 13, 2008
I have been wondeirng what the optimal rejection algorithm is given my current setup. My professors typically use the avsigclip; however, since we have begun to calculate the gain and rdnoise for our ccds, we switched to ccdclip. This has worked very well except on a few nights when neither avsigclip nor ccdclip seem to perform very well at all (the best results that I have found for this particulat night have been while using minmax). I was wondering if it perhaps has something to do with the sensitivity noise parameter that we have been ignoring. This leads to my questions.
What is the sensitivity parameter exactly, and Is there some IRAF task that I can use to calculate it? What is the best situation for each rejection algorithm? They can't all be equally utilizable,...can they?
I want to produce the highest quality data while staying within acceptable boundaries, and it just feels wrong to use different rejection schemes on different nights and say that my data is uniform/contiguous.
Thanks.
-Paul
What is the sensitivity parameter exactly, and Is there some IRAF task that I can use to calculate it? What is the best situation for each rejection algorithm? They can't all be equally utilizable,...can they?
I want to produce the highest quality data while staying within acceptable boundaries, and it just feels wrong to use different rejection schemes on different nights and say that my data is uniform/contiguous.
Thanks.
-Paul
Francisco Valdes wrote on Mar 13, 2008
The "sensitivity" term in the noise model comes from work done at STScI. In the help page at the bottom SEE ALSO it points you to wfpc.noisemodel. This is in the STSDAS package which you may need to install. The package path is stsdas.hst_calib.wfpc. The help page and task should help you understand this.
I strongly doubt this is the source of your problem. You don't say how you decide the result is poor. I agree you should try and understand why things don't seem to work well on those nights. In general I would agree that using different methods on different nights is not a good thing. On the other hand, the goal is to get the best result out of the data and if that requires different choices then maybe that is what it takes. Again, the main factor is understanding "why" you need to use a different method.
I hope this helps.
Yours,
Frank Valdes
I strongly doubt this is the source of your problem. You don't say how you decide the result is poor. I agree you should try and understand why things don't seem to work well on those nights. In general I would agree that using different methods on different nights is not a good thing. On the other hand, the goal is to get the best result out of the data and if that requires different choices then maybe that is what it takes. Again, the main factor is understanding "why" you need to use a different method.
I hope this helps.
Yours,
Frank Valdes
Last post on Mar 13, 2008