How much data is too much?

DONALD VANDERDOES

I am less than a year new to acquiring data and was just thinking about how much data to capture. What is the point of diminishing return, if there is one? Or do you just get as much data as possible?

Clear Skies

BlockHead

Hi Donald,

This is always a hard question to answer- but you need to satisfy a few things:

1. You need to collect enough frames to get good statistics in order to process the data in terms of weighting, normalization and most especially rejection. This statistical number of frames (note, I said NOTHING about exposure time) is around 20-30 frames. 

2. Given the above you want to expose as long as you can (see below) in each frame that results in an image that is not degraded in any form. For example, tracking errors could be a problem for some people if they do 10 minute exposures. Another issue deals with the fact that some sensors (especially new CMOS) have shallow bit depth. This means the stars "saturate" quickly since they are bright. I personally think people (especially beginners) make a big deal out of this unnecessarily and they purposefully degrade their images by trying to accommodate the issue.

3. There is also a limit in terms of exposure per frame that is possible under light pollution. Under a dark sky only the first two points matter. For a bright sky, shortening exposures may be necessary.

Once you digest the above- then for a given total exposure time (given points 1-3 above) you simply look at the faint limit you have captured for any particular time. Then in your mind DOUBLE that time- you will NOT see more detail and the result will NOT be strikingly different. Doubling the time will lessen the graininess of the image- but it will be necessary to DOUBLE the time to get to a point that the graininess when displaying the faintest things is lessened. If the faintest things you want to show do not look grainy- and the total integrated result looks smooth in what you intent to show... you can stop taking more data. More exposures can help with the contrast a little... but most people would say you have reached that point of diminishing return... even doubling the time will not make the image easily different without looking very very closely.

I would recommend you keep this reply and come back to it. My answers are based on several different advanced concepts as well as my experience. You should do the experiment. Double the time... see what happens. For your given sensor the result is the same no matter what object you are looking at. My statement "the faintest things you want to show..." is a very deep comment. Consider how that statement *would* affect how much data you would capture on the star cluster compared to a galaxy- or compared with one nebula to another. 

-the Blockhead