[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: Re: [APML] NGC 6992 uploaded
Ray, thanks for that important upgrade. I knew that the theory was you needed three to median combine. I am not sufficiently expert to have gotten into that so skipped over it. The main point though is that averaging two x three hours vs. adding them tosses out a lot of signal that he has already smoothed by using the multiple scan technique. I think.
Stuart "who wishes he knew more about this stuff (and sorry about your glasses - happened to me too!)" Heggie
>
> From: Ray Butler <ray.butler@nuigalway.ie>
> Date: 2005/07/26 Tue AM 05:25:07 EST
> To: Discussion of Film Astrophotography <astro-photo@seds.org>
> Subject: Re: [APML] NGC 6992 uploaded
>
> Hi Stuart,
>
> This statement of yours caught my eye:
>
> >Third, averaging the two scans of each negative will reduce the noise so why
> >not ADD the two resultant images vs. median combine - this way you get the
> >benefit of twice as much signal.
> >
> >
> It depends on how "median combine" is implemented. Strictly speaking you
> need a minimum of 3 samples (images in this case) to get a median, since
> it is defined as the mid-point of the range of values. What is the
> mid-point of 2 pixel values? It can't be the higher value and it can't
> be the lower value, since they are both extreme values - neither is in
> the middle. So the median is undefined. How this is dealt with depends
> on the software; it may default to a straight average of the two values,
> it may pick either the higher or lower values by default, or it may spit
> the 2 images back at you with some nasty error and question your
> intelligence! The most logical approach would be to take the straight
> average of the 2, since this is how the median should be computed when
> the number of samples is large (>3) but even in number: there is no
> single middle value, so the _two middle_ values would be averaged.
>
> What I'm saying is that the user needs to find out what their software
> actually does when asked to do a median combination of only two images.
> It is possible that it does actually default to an average, in which
> case all the signal is being used after all.
>
> Ray "who vaguely remembers an old scientists joke along the lines of '1
> and 2 are fine, but once you get to 3 you have to do statistics' " Butler
>
> --
> Dr. Ray Butler
> Lecturer, Physics Department & Computational Astrophysics Laboratory,
> National University of Ireland - Galway,
> University Road, Galway, Ireland.
> Web: www.nuigalway.ie/physics/ Email: ray.butler-AT-nuigalway.ie
> Tel: +353-91-493788 FAX: +353-91-494584
>
> _______________________________________________
> Astro-Photo mailing list
> Astro-Photo@seds.org
> http://seds.org/mailman/listinfo/astro-photo
>
_______________________________________________
Astro-Photo mailing list
Astro-Photo@seds.org
http://seds.org/mailman/listinfo/astro-photo