Post B2fIH206FjZJRtxXCS by Elliptica@poa.st
(DIR) More posts by Elliptica@poa.st
(DIR) Post #B2f9Zsv1LOldCFUyx6 by karna@poa.st
2026-01-26T02:27:14.301558Z
6 likes, 5 repeats
IC 443 (Jellyfish Nebula)14 hours and 10 minutes of integration timeDual narrow-band filter for H-alpha and O-IIIProcessed using the foraxx palette - its closeish to the SHO palette but a little more "dynamic". it may not have been the best image for this palette, but i liked the way it combined the "natural" look of the nebula with the blues of the HOO palette. I played around with drizzling since my setup is technically undersampled (and also it can be better than regular debayering)I had a nasty halo around the bright star right above the nebula which I could not get rid of without resorting to clone-stamping it out - so that is the most "fake" part of the image.#astrophotography #DSO #nigger #rape #finance
(DIR) Post #B2f9wQDmCsqKqy1Bdg by SuperSnekFriend@poa.st
2026-01-26T02:31:48.720901Z
1 likes, 0 repeats
@karna So pretty!
(DIR) Post #B2fAcMn7LjzGNUY7hA by karna@poa.st
2026-01-26T02:39:21.453920Z
1 likes, 0 repeats
@SuperSnekFriend thank you friend! i tried to keep some of the natural look since all of you guys seemed to like it when i polled about it some days back
(DIR) Post #B2fB0HbiH72NFZcSJM by karna@poa.st
2026-01-26T02:43:40.434071Z
0 likes, 0 repeats
Color Credits: I got the palette idea from this youtube video https://youtu.be/Gy42AeZ_XB4 which in turn got it from this blog: https://thecoldestnights.com/2020/06/pixinsight-dynamic-narrowband-combinations-with-pixelmath/
(DIR) Post #B2fCjdXbfkdCGIt9zk by Elliptica@poa.st
2026-01-26T03:03:03.512152Z
1 likes, 0 repeats
@karna That's an amazing result. Great Job!I want to start doing Deep space imaging again, but my telescope is currently set up for planetary. Once Jupiter settles down, I will switch back to doing that.
(DIR) Post #B2fFgcb51dPUKOlczQ by karna@poa.st
2026-01-26T03:36:11.312659Z
1 likes, 0 repeats
on drizzling: when I was stacking, I tried some different approaches to see what would help with both improving detail and reducing noise. I tried a few things:regular stacking (no drizzling)drizzling 1x upscale; 1x pixel fraction; square kerneldrizzling 1x upscale; 1x pixel fraction; lanczos2 kerneldrizzling 2x upscale; 0.9x pixel fraction; square kernel; then 0.5x downsampledrizzling 2x upscale; 0.8x pixel fraction; square kernel; then 0.5x downsampledrizzling 2x upscale; 0.7x pixel fraction; square kernel; then 0.5x downsampledrizzling 2x upscale; 0.5x pixel fraction; square kernel; then 0.5x downsampleThe biggest improvement was going from no drizzling to a 1x1 square drizzle (less noise and smaller stars) which makes sense since the CFA drizzle is supposed to be better then regular debayering anyways. The lanczos and 2x1 square were pretty similar. and decreasing the pixel fraction to 0.5x made the stars smaller but also increases the noise (not as much the original though). The siril docs said a pixel fraction that is the reciprocal of the upscaleing factor is a good place to start, hence the 0.5 attempt. but on other forums i saw people say 0.7-0.9 is best.Ultimately, i went with 2x0.8 which seemed like a good compromise between noise and star size. though the drizzle really shined after deconvolution/noise reduction steps which I am not showing here.PS all the pic rels are just linear osc stacks with an unlinked autostretch in siril.
(DIR) Post #B2fGL7bx8R6I6FUHSq by karna@poa.st
2026-01-26T03:43:30.497313Z
0 likes, 0 repeats
@Elliptica thank you friend. jupiter shots sound fun - I wish I had a telescope big enough to get enough resolution on planets. How does the data gathering process go with jupiter rotating so fast, does that mess up the stacking (if any) since the rotation would be quite apparent (I presume)?
(DIR) Post #B2fIH206FjZJRtxXCS by Elliptica@poa.st
2026-01-26T04:05:10.695629Z
0 likes, 0 repeats
@karna Data goes as fast as you can get it. I usually I aim to get something between 150 and 200 fps, though I've heard some people say that frame rate is not as important as good observing conditions, and that there is diminishing returns past 30 fps.Because Jupiter rotates so fast (around 1 degree every 2 minutes) I try to only take a video for 3 minutes, max, and then I stack only the best 15% of the images from that run. Because the best images usually come from a few single moments of really good seeing, and because I'm aggressively selecting only the best frames, the rotation isn't much of a problem for single images, especially for a color camera.The hard part is when you want to add filters. There is a different telescope I am working on at the moment that uses a mono camera, and they want to have rgb filters plus an IR long pass filter (ir is less affected by bad seeing) take images. So that's 10 to 12 minutes of imaging. If all filters go for that time, Jupiter rotates around 6 degrees, which is noticeable. When you try to line the images from each filter, the image gets really blurry.The "trick" that I haven't managed to get working yet is to de-rotate the image, so that all the images line up. The best software solution for that currently is a program called WinJUPOS.