0.90
#...
·
 |
|---|
Jan Erik Vallestad:Carl Elgario: One last question. You mentioned a clone stamp, which I assume came from Photoshop, but you also said it was all handled in PI. In that case, how did you use such a tool within PI to get rid of the residual coma? And in the past, I would use BlurX because of the fact that I could never get Starnet to work with the coma, but with this application and your method, you don’t need BlurX really at all. Really is quite fantastic what you were able to get out of my data. Now it truly feels like the 7hrs I spent in the cold in Bortle 3 was worth it, knowing what the data is capable of appearing like. |
4.36
#...
·
|
|---|
|
PI has it's own clone stamp tool, PS does it better, but it is useable in PI as well. I think I used about 60% opacity and quite a bit of feathering. Yes, when I have removed stars like these myself I have only used BXT on the starless image. Trial and error made me realize that it made the star removal worse, so in stead I use morphological transformation to dim/reduce the stars before adding them back in. One thing I could mention is that lots of stars were over-exposed, the reason they had these weird colors before and during stretching, which led to me having to stretch it more than I thought I needed to. So taking separate star images with shorter exposures makes sense either way, so you may as well stop down for those anyway. Win win  Great that you're happy! |
|
0.90
#...
·
|
|---|
Jan Erik Vallestad: Interesting, I'll look into that PI clone stamp. I did not realize processing images could get so remarkably deep, but wow the deeper you go the better your images get. As for the stars, that really is a fantastic suggestion. Better stars + not overexposed! So basically, all I'd really have to do for that is get separate data that is stopped down and is shorter exposure, specifically for the stars, and use that as my star layer that I would combine my starless with? How much integration time would I need for this "star layer", though? Surely not that much, as most of the data is from within the starless layer? |
|
4.36
#...
·
|
|---|
|
Exactly, you can experiment a bit with this. With my mono setup I have used only a few minutes of RGB combined with luck. As you say, all the data is in the starless image, and stars don't require long integrations. A typical solution for me would be to get 10x30s pr filter, but you could try 10-15 minutes perhaps. Edit: I have gone as low as 7 minutes as well at F/4. Albeit with mono. |
1.20
#...
·
·
1
like
|
|---|
| -------- |
|
0.90
#...
·
|
|---|
Jan Erik Vallestad: I've been playing around with the data a bit and I am getting a bunch of "smudges" where a bunch of stars used to be. A lot more than what you were seemingly able to achieve with StarX. It makes the data look very odd once combined. Do you think you could provide a link for me to download the starless + stars layer you were able to extract with StarX, before any further processing? It will help me make my final decision on whether or not I should invest in StarX. I also sent you a message on Instagram, but I wasn't sure you saw it or not, so I decided to follow up here. Also another thing, if you don't mind. I saw you mention your use of MorphologicalTransformation to help tone down the stars. Can you explain exactly when and why you used this process? Not necessarily HOW to use it, I can do my own research on that, but rather the point in the processing you did it, and why you chose to use it. |
|
0.90
#...
·
|
|---|
Jan Erik Vallestad: Hey Jan. Scratch my first question. I bought StarX on a gamble and I'm glad I did. Wow, the starless image with StarX looks indescribably better than the starless with Starnet2. I think they used to be on par back about a year ago, but the improvements in the StarX AI have made it evidently a MUCH better tool. My second question regarding the use of MorphologicalTransformation still stands, though. |
|
4.36
#...
·
|
|---|
Carl Elgario: Sorry, I had turned in for the night. Alright, I think you will be very pleased with the tool. If you use the "full package" with large overlap it works best, no unscreening in linear as well (only in non-linear). It does a very good job of extracting everything and replacing the void with a matching noise pattern. As for MorphologicalTransformation I find it to work very well in order to make the stars take the back seat. It won't always be necessary when you get your separate star images working - Then you could control the stretch level a lot more, and might effectively negate the need to reduce them. The idea of the script is basically that it can both erode or dilate stars in an image. So, for instance, you could make an image with an entire constellation and dilate the stars that make out the constellation itself as well (Just as an example). Though you would need to mask them separately first. There are a few settings and some adjustments might be needed for different images, although you could probably figure out something that works pretty good for your particular combination of camera/scope/lens as a basic part of the workflow. I use it on the star layer just before re-combining (this is the final step for me). You can apply it to an image with stars, but then you would need a very good star mask first. |
|
4.36
#...
·
·
1
like
|
|---|
| Also, as Bryan mentions above, if the lens can't reach infinity focus then this might be an issue to look into as well. I haven't had any problems with my copy regarding focus, in fact I find the images very sharp, so I haven't bothered to modify it. Some people have copies where there focus mark don't reach infinity or a bit beyond though. |
2.11
#...
·
·
1
like
|
|---|
|
There's a lot of messages so I don't know if what I'm going to say has already been said but in my experience with this lens: 1- Take great care of the connection between the lens and the camera (or whatever is between the lens and the sensor). F/2 or even F/2.8 is extremely sensitive to the bac focus and the tilt. So take the time to adjust the back focus at +/- 0.1mm (that's the tolerance for f/2.8 when you are demanding) Preferably, use a system which allows you to tweak the tilt, especially if you shoot full frame. I've managed to get very good results by tweaking the tilt (with the cost of a long test session). I purchased few years ago this adapter which allows to tweak the back focus and tilt : https://astromechanics.org/  2- I never use it @ f/2. The gain of light doesn't worse the image degradation. IMHO, f/2.8 is the best combo light vs quality with this lens. 3- Use good brackets which handle the weight of every elements: ring around the lens, ring around the camera, preferably near the fixations. You should be good to go, without needing to cheat with post treatment. (If you still want to go that way ... you can shoot only few short subs with another scope / or with the 135mm closed to f/8 and add the star layer to your starless image taken a wide aperture...) Here is the setup I used few years ago (for the bracket system):  |
|
1.20
#...
·
·
1
like
|
|---|
| --------- |
|
4.36
#...
·
·
1
like
|
|---|
Guillaume Seigneuret: All very good points, but as the OP shoots with a DSLR there isn't really much you can do with neither tilt nor backfocus. I agree that shooting at F/2 won't give you any additional advantage unless you use high speed filters if you are using mono though. Many people also use RASA and Hyperstar, so I see no difference between that and using the Samyang wide open. I don't think shooting stars at F/4 is considered cheating, nor is it that much work to do. Most of us do separate star images anyway, especially when using mono and NB filters. |
|
1.20
#...
·
|
|---|
| ------- |
1.20
#...
·
|
|---|
| I have gone through two copies of the Rokinon and I have never been able to get sharp stars to the corners at apertures other than f/4 with my ASI2600MC (APS-C sensor) camera. The first one was horrific. The second one was better, but only tolerable. f/2.8 looks terrible. Wish I could use it. |
|
1.20
#...
·
|
|---|
| -------- |
|
4.36
#...
·
|
|---|
Brian Fulda: I've yet to test it on my 2600MM as I only recently received it. Bray Falls use his at F/2.2 with a 6200MM I believe, I can only assume that is due to some sort of black magic. |
|
4.36
#...
·
·
1
like
|
|---|
BryanHudson: I'm probably going to do this myself, although my copy focuses slightly past infinity already. Seems simple enough and there isn't really any reason not to do the mod I suppose. Edit: Did you remove the hard stop or simply move it further beyond? |
|
1.20
#...
·
|
|---|
| -------- |
2.71
#...
·
|
|---|
| I thought the modification to the ring was only necessary if you couldn't get a v-curve with the lens. Am I mistaken? |
|
1.20
#...
·
|
|---|
| ---------- |
 #...
·
|
|---|
I thought the modification to the ring was only necessary if you couldn't get a v-curve with the lens. Am I mistaken? if your getting a v curve over the range of your operating temps then you have some wiggle room around the CFZ and shouldnt need mess with it i would of thought. |
|
2.71
#...
·
|
|---|
I thought the modification to the ring was only necessary if you couldn't get a v-curve with the lens. Am I mistaken? Yeah, I thought it was only needed if the lens could not autofocus because the lens could not be pushed past the infinity mark. |
|
1.20
#...
·
|
|---|
| ------- |
#...
·
|
|---|
BryanHudson: you can rig these up for autofocus,but the point is if you can go either side of CFZ either manually or automatcially then your good to go. |
|
4.36
#...
·
|
|---|
| I use the EAF as well as other optics, only takes a bit of printing and a belt really. Some are noticing that during cold temperatures the critical focus zone might change, I guess HFD or HFR might be the best judge though. Or a bathinov for anyone not using NINA or the Asiair etc. I noticed some people simply moved the hard stop a bit, either way it's a pretty simple solution. I wouldn't expect that mod alone to result in round stars at F/2 though. |
