How to display images for apples-to-apples comparison?

Jason Schella

I have subs from a B7 site and have recently moved to a B3 site. I'm trying to a comparison to illustrate the difference/benefit of the B3 data.

If I use the SNR script it shows that, in order to achieve a similar SNR for the B3 data, I require ~10 times the integration time for the B7 data.  This number makes sense.

However, when I look at a 10 minute exposure of the B3 data and compare it to a 90 minute total integrated image from the B7 data I tend to see more faint detail in the B7 data. I use the same stretch on both images (normalizing the background level).

Any thoughts on how to best do this?

Thanks,

Jason

BlockHead

Well... it is complicated.

First I would likely just compare subframes- or a very small group of the best subframes from each.

By looking at the integrated images you are mixing the relative image qualities.

Even better, perhaps... Put all of the frames into WBPP and output the registered files. You could look at the weights (of PSF Signal Weight) or if you use NSG you would would looking at PSF Scale SNR. This would show you the relative quality (weights) of the subframes. 

I think this is the best way to proceed since it is measuring the stars to produce a quality measure. If both sets of subframes are taken under similar conditions (good quality for their respective sites) then I think I think it will be clear the dark site will show you faint detail *with less noise* (that is the key).

Finally - there going to be a difference in the sets of data with respect to the number of frames. Simply averaging many (probably more) frames to make a 90 minute masterlight will likely have better SNR relative to the 10minute under the dark site. That isn't a far comparison either. I would compare equal exposures. I think the better metric is the SNR and not thinking in terms of relative exposure time (which is related... but it includes sources of light and noise).

-the Blockhead

P.S. Just to add... the brighter sky adds Poisson noise to your B7 images. This is the issue you cannot stop from happening...it is just the nature of light.

Jason Schella

Hi Adam, thanks for the thoughtful response. I did put the subs through NSG and found that the weightings were about 6 times higher for the B3 site. I think this makes sense given that I was using a 5nm Ha filter. I've yet to have the opportunity to image with the LRGB filters, but I would guess I should see an even bigger difference.

I was trying to find a way to display, or stretch the images, so that I could see the "real" differences between the images. I think this may be too complicated to do in a repeatable way. For instance, a bright star should be the same brightness in both images, but the background will be different.

Thanks again for the input,

Jason