Wodaski's chart is only true if the noise is Gaussian and the signal is
constant. Thermal noise should be Gaussian. Other sources of noise may not
be Gaussian so that the improvement with number of frames may vary faster or
slower than shown by the graph.

The graph is representative of noise improvement as a function of exposure
time - for a linear device like a CCD camera.

Randy

----- Original Message -----
From: "Bob Nanz" <bob@...>
To: <SDAA_AISIG@yahoogroups.com>
Sent: Thursday, July 02, 2009 2:44 PM
Subject: Re: [SDAA_AISIG] Signal to noise calculations


> That is a little misleading.
> Even though there is little improvement per frame if you double the number
> of frames (100 vs. 200) you improve your S/N by as much as you do in going
> from 1 frame to 2 frames.
> Bob
> ----- Original Message -----
> From: San Diego Paul
> To: SDAA_AISIG@yahoogroups.com
> Sent: Thursday, July 02, 2009 12:18 PM
> Subject: [SDAA_AISIG] Signal to noise calculations
>
>
>
>
>
> RE: the point of diminishing returns, where taking more subframes has
> little or no value on the SN ratio of the final picture.
>
> Thanks to Rod Wodowski: http://tinyurl.com/mche85
>
> The bottom row is the # of subframes you take, the vertical axis is the
> approximate % improvement in S/N by adding ONE more image
> to the existing stack.
>
> Very interesting!
>
> --Paul B
>
>
>
>
>
> [Non-text portions of this message have been removed]
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>