Affiliation
American Association of Variable Star Observers (AAVSO)
Fri, 03/08/2024 - 16:33
Greetings,
I have been submitting measurements to AAVSO for some time. However, I have never submitted SNR or Error values because I have not been able to determine them correctly. I capture from my DSLR using Ekos and process the images in Siril. The problem is that I get wildly different values for SNR in Siril and AstroimageJ. For example in an image linked below of V117 Tau, I get a value of about 727 with AstroimageJ and 22.3 in Siril (I used a gain of 0.45 in each). I would appreciate any help in this.
https://drive.google.com/file/d/1rRVZIsk_BdgSmHwgAeNzQ8B0CUyv-2WR/view?usp=drive_link
Best
Riyadh
Hi, I wish I could help because I've had the same problem for a long time. Normally I use MaximDL but I have also used another one called ASTAP. They give widely different SNRs and errors for the same image. When I submit observations I use the values calculated by MaximDL. I was advised to do that by someone on this forum. I've contacted the developer of ASTAP about the problem but with no satisfactory answer. I hope someone can answer your problem.
Steve
Riyadh,
For accurate SNR in AIJ, gain, dark current and read out noise of the camera detector must be entered in the Aperture Photometry Settings panel. The latter can be accessed by clicking the "change aperture settings" icon above an image.
You mention you use DSLR images for photometry. There are problems with assessing dark current accurately in some DSLR cameras because it does not increase as a linear function with increasing exposure as is the case with CCD cameras and (within the appropriate exposure range) monochrome CMOS cameras (I have not determined dark current in any colour CMOS camera).
You should be able to approximate SNR in AIJ by calculating source-sky/square root of source-sky for each star.
Roy
After some trials with Siril and AIJ, it seems that the SNR calculation does not work on 32bit images in AIJ - requiring the pixel values to be in ADUs rather than the normalized 0.0 to 1.0 values in 32bit FITS images. When comparing the values for 16 bit images in AIJ with those given by Siril, they are in fact not too far from each other. My mistake was that I failed to notice that the values given by Siril are in DB while those in AIJ are straight Signal to Noise ratios. When I converted using SNR = 10^(DB/10) the values were similar. The reason I have to use Siril to process my images is that as far I could tell, AIJ does not have features for image de-rotation and alignment prior to stacking - which I need as I am using an alt/az mount.
Best