In late august, nights in northern germany finally get dark again after two months of astrophotographic deprivation. That's my anual season to forget my intentions of doing something different for once and to image the veil nebula instead. So here's the 2018 version. I tried to work very thoroughly and used only 108 out of around 250 lightframes that I took.

At my location, it was the last 200 years' dryest summer and the temperatures were fairly high during the nights as well. So I got the idea to experiment with noise reduction techniques for my uncooled DSLR.



First, I started taking dark frames, which I had not done for a few years because my autoguider's dithering function is supposed to make the use of darks obsolete. The idea was to test whether noise reduction benefited from the additional use of darks - it did not. When using a sufficient numer of dithered lightframes (in my case around 100), the SNR increase was something around 0.001. On the opposite, when calibrating single shots (I did this with an image of a perseid meteor), dithering is not possible and darks are the only option. So when using dithering, there obviously is a negative correlation between the number of lightframes used and the effect of the use of dark frames.

The second thing that I tried was using two similar cameras, in my case modified 6Ds. Since the sensor characteristics differ significantly among individual units, this does have an effect on the identification of outliers. One of my 6Ds produces ampglow gradients in the upper left while the other one has a higher overall noise. As expected, both effects were reduced when combining both cameras' images.

This image of the cygnus loop is the result of my noise reduction experiments. If anybody has ideas, thoughts or criticism (especially concerning my approach on noise reduction), I would be happy if you shared them with me.