Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
In this Discussion
WBPP results in uneven field - How I solved it
Hi Fellow Blockheads,
This started as a post for help, but as I was gathering data to post my problem, I applied a little "Adam Logic" and set about eliminating possible root causes one by one. As a result I have figured it out, but I still wanted to share in case others had advice or could benefit from my errors. Thanks to Adam for his carefully explained videos!
I've been watching Adam's fantastic video series on WBPP 2.x - and I've tried the process (using 2.2) but my results are not good. A sample of my LUM filter integration shows the outcome (the other filters are similar - not the same, but similar).
First step was to eliminate a potential issue from having very old calibration data. I took fresh bias and darks (they were from 2019 and during one of Adam's troubleshooting videos he pointed out the folly of using old bias and thermal frames). I have taken, but did NOT apply dark flats - as far as I can tell I don't need them, as what I gather is that this is really only required if you're darks are showing amp glow and cannot reliably scale. Besides, my darks match my lights for time and temp. Would love some feedback here. Here's a typical dark for consideration.
At any rate, using fresh bias and darks made no difference (in my uneven background at least - I'm sure it had some effect), so I do not think those are the issue - the integrated lights look the same as with the old calibration frames.
So, I'm left with the flats. Now, here I have to admit that I just set up my flat session in SGP and let it rip. I only look at the first couple of them and then go about my business. Turns out, my flats - despite automation using an Alnitak Flip-Flat - had some outliers for mean values.
I discovered this as I loaded the flats for each session and filter into Blink and applied a consistent histogram from one frame to the others. That's when I spotted the outliers - cheeky little buggers.
Here's an animated GIF of Blink - note the black frame as well as a couple other brighter than average frames.
After eliminating the obvious outliers I got a much better result (note, this was only one set of multi-night data - the LUM frames from one night and corresponding flats). Not perfect, but getting there. I'm planning now to more closely compare the flats and toss more outliers. Also, I am going to try and figure out why I'm getting inconsistent results in my flats session in SGP.
Comments