Anyone knows to tackle the phenomenen of color dispersion in startools? I have some stacks of globs that have it a bit, due to the fact that the object was not too high up in the sky. Is there a way to overcome this?
Ivo is of course correct, as to trying to fix this within ST. I have been known to do either or both, depending on what is needed. You can experiment with the pixel amount of highlight repair to set just enough but not too much, though it is still a bit of a desaturation technique. Likewise in Shrink, and you can maybe add a little deringing if needed also (works even with zero iterations shrinking).
Really though this is mostly an issue for the stacker, I have found. Though theoretically if you split your channels you might be able to manually adjust things, in a multi-step process, using the x/y displacement in ST Layer.
My quick solution is usually just to run the file through DSS with RGB Align on. Note this is spatial channel alignment, not neutralization or normalization. You should be able to do this with any final stack, regardless of what program created it. You just load the stack a single "lights" file to be "stacked" (even though there are no other files to stack it with). It will still "register" and "stack" but don't worry about that, all it does is spit the same file back out, but with RGB aligned. If set up during a full stacking session, RGB align is the final step anyway.
DSS has had this feature forever. But, within the last year or two IIRC, PI actually added it too. I forget what they call it (some long-winded name, no doubt), but since you now use PI for stacking that should do the trick for you also.
Hopefully either way it helps correct and preserve star colors rather than just fading the anomalies. Usually works pretty well for me.
Ivo is of course correct, as to trying to fix this within ST. I have been known to do either or both, depending on what is needed. You can experiment with the pixel amount of highlight repair to set just enough but not too much, though it is still a bit of a desaturation technique. Likewise in Shrink, and you can maybe add a little deringing if needed also (works even with zero iterations shrinking).
Really though this is mostly an issue for the stacker, I have found. Though theoretically if you split your channels you might be able to manually adjust things, in a multi-step process, using the x/y displacement in ST Layer.
My quick solution is usually just to run the file through DSS with RGB Align on. Note this is spatial channel alignment, not neutralization or normalization. You should be able to do this with any final stack, regardless of what program created it. You just load the stack a single "lights" file to be "stacked" (even though there are no other files to stack it with). It will still "register" and "stack" but don't worry about that, all it does is spit the same file back out, but with RGB aligned. If set up during a full stacking session, RGB align is the final step anyway.
DSS has had this feature forever. But, within the last year or two IIRC, PI actually added it too. I forget what they call it (some long-winded name, no doubt), but since you now use PI for stacking that should do the trick for you also.
Hopefully either way it helps correct and preserve star colors rather than just fading the anomalies. Usually works pretty well for me.
Mike,
Tried to RGG align in DSS, it seems not much help, or i did it wrong...I still need a bit of testing here.
These are all such fine tweakings...you really have to find out what tweaking for what problem...takes time i guess
No? Bummer. Well maybe the PI version of it will help.
You just showed a cropped corner I think - is the blue/red orientation the same across the entire image? It should be for dispersion, which an RGB align type technique should be good for. But if the blue/red is different in different areas of the image it could be C/A instead?
Yes the orientation is thesame across the whole image. It is no CA, it is clearly atmospheric dispersion. Strange i don't have it on my M13, allthough it was not high up at the time. The Elf has it on his M3 too, there is a thread on CN about it, it is thesame . But i also think that his working around in PI did not solve it completely either if i evaluate the pictures...
I still have to check in PI, for now i am using an older version , better switch to the latest i guess.
