I have several different nights worth of M31 data. I am hoping to cull the subs in some rational way and then move to a new directory those pp_images .fit files that I want to use (having figured out how to use windows powershell to batch rename so I can get all sequential pp_images filenames).
If I use the .csv file to help choose the pp_images files with background values below, say, 0.0052, move (what I thought would be) those files to a new directory, choose the same stars again I get:
as looking through the original raw subs I don’t see any offhand that would have stars with roundness down near 0.1. If I randomly choose a single pp_images and box a star (either one of the four or any other star), I have yet to encounter any with roundness below 0.5, and most are > 0.8, which the plot leads me to believe should be in the minority.
I realize that I would be better off doing the sequence PSF on the r_pp_images sequence (which I have), but I assume there could be yet another step of re-ordering the files, making it that much harder to identify specific pp_images files.
Thanks in advance for any insight into how this works.
Hi, I’ll try to answer your questions, but I don’t understand why you need to ask them. Siril can filter the images used for stacking, the deep-sky registration step keeps in memory the FWHM and roundness of stars. You just have to rename your images to have different numbers for pp_images of the different sessions. Sirilic automates that if you like.
The X axis is either the image number or the time if dates are available in files. In your case it’s time, JD means Julian Day, where a unit is one day. I think there’s a way to switch to number of images in the window.
In the current development version of siril, you will find a new feature: unselecting images directly from the graph and displaying information about the image where the pointer is on the graph.
A roundness of 0.1 if the images look normal means that the star you selected wasn’t detected in some images. If you use the global star registration, the roundness you’ll get will be an average of all stars used for registration, so much more reliable than the roundness of a single star with moving images.
Thanks. guess I should have gone to wikipedia first - I had googled Julian Day and was told today is 20341 - but wikipedia lists the version where today is 2459191.
As to ‘why?’, it started with some of my nights having issues with very high background at various times in the session (thanks, neighbors, I guess?). Brightness is not one of the filters that can be used, as far as I know. That led to the general question of 'is images_001 the same frame as pp_images_001? etc. And it appears that is not the case.
Meanwhile, I’m learning more about what is going on “under the hood”, which is worthwhile in itself.
In the filters, the weighted FWHM is a FWHM weighted by background if I remember correctly, so basically what you need.
The output sequences are supposed to be in the same order as input sequences. If the sequence is FITS files, indices should be the same with most operations, maybe some fill the holes, I don’t remember. With other sequence types, like a sequence in one file, holes in the sequence numbers have to be filled.
Here’s an example of what started me down the road of wondering what’s up with image ordering. These two plots are from the same sequence. The first one is the average fwhm across all registration stars (these plots have “image #” as the X-axis). I had seen similar plots before, and wondered what might be causing the sudden shift in fwhm, as I had not done any refocusing or anything else:
Here is the same sequence plotted for one selected star “psf for the sequence” (these plots have the timestamp as the X-axis):
It was only after the first time I used the psf for the sequence function that I realized that the images to the far-right in the upper plot are actually frames from the beginning of the imaging session. (if there is a way to switch between timestamp and image number for each plot, I have yet to find it)
Hi. Sorry for the delay, final exams week arrived and I need to get everyone’s grades calculated and posted (preferably correctly ).
I did start with Sirilic - I should go back and start from the beginning with just Siril to see if Sirilic is somehow copying files out of order. Note that both images above are exactly the same registered sequence. The first image is the average fwhm plot, which arranges by “image number”, the second is from choosing an individual star plus PSF for the sequence, which orders by the timestamp.
).