Cull Subs Fast With PixInsight's SubFrame Selector

[@JerryNewman-w8p]

Great video. I’ve seen other videos about SFS but none that talk about the specific properties to use to cull your subs. I also found that you can use FWHMSigma

[@aleixandrus]

I also use that for a faster first selection, but I manually check those subs in the proximity of the 2 sigma mark. Hard to say bye to 300s or 600s subs 😅

[@SKYST0RY]

It's a good idea. I often manually review measurements to make sure they are worth culling or keeping even despite the statistical measurements. Ex. If all the subs are good but some are exceptional, a statistical measure would show the good subs to be of less quality in comparison even though they are entirely worth using. So it is a good idea to watch the discreet numbers and even do a visual inspection prior to using SFS.

[@HeavenlyBackyardAstronomy]

Thanks for the great tutorial. This should come in quite handy.

[@ThePetrus112]

Great video as with everything else on this channel. truly love it! Do you calibrate or debayer subframes before running SFS?

[@SKYST0RY]

Thank you! All I do is a visual inspection prior to running SFS. Regardless of the myths that WBPP will remove satellite trails, satellites will leave "scars" in images, even if faint, so I manually remove them. And SFS doesn't catch the occasional double image. (I'm still not sure what causes the double image. I get one or two per night. I think perhaps a bug where NINA auto centers while shooting.) So a manual review is good. It only take a few minutes with Irfanview, then straight into SFS for final culling. Then into WBPP.

[@ThePetrus112]

Do you debayer before visual assesment? The origin of my question is that my subframes before debayering looks so blocky i even doubt SFS is doing calculations the right way

[@SKYST0RY]

@@ThePetrus112 I am shooting with a mono camera, so no.

[@AmatureAstronomer]

Nice.

[@IronMan-2024]

Thanks for finding this very useful tool in PI. It would be nice if the tool could summarize the first and second deviation value so one would not need to manually find the value. Unless I’m missing something from the charts. Also you need to be a little careful on your wording, the expressions are in the approval window so you are really trying to select data within the gray regions and cull samples outside the expression.

[@NotAmishAstronomer13]

I thought that WBPP does this for you, does it not?

[@SKYST0RY]

I get okay, but not great, results with WBPP's culling. I prefer to keep the process much more hands-on. In fact, unless I have thousands of images, I give them all a visual inspection before even running them through SFS and manually remove images with satellite trails or vibration-based errors. Automation still promises more than it can deliver.

[@nikaxstrophotography]

Cull subs FAST after watching video it takes LONGER than normal.

[@SKYST0RY]

Fast is relative. In this case, faster than culling by eye.

[@nikaxstrophotography]

@@SKYST0RY Or use astropixel processor and their graphs to cull

[@SKYST0RY]

@@nikaxstrophotography Interesting. I will have to take a look at this some time.