Reducing star bloating from modded dslr...

[Mr Bergman]

 Hi folks, 

 I have been trying to reduce star bloating from my modded dslr. Long exposures seem to bleed out colours from really big stars, which after stretching give me large difficult halos. Doesn´t matter if I isolate the stars before stretching, since the bloat already is lurking in the background. I had a look at the colour channels and the biggest problem is in the red channel, where I also have the nebulosity. 

So how do I solve this? Well, I am not sure to be honest! The only way I can come up with is, 

1) Careful stretching when doing the non-linear, not to bloat the stars

2) Removing colour halos 

http://budgetastro.com/micro/articals/red_halos/red_halos.html 

These two is only working for stars that are not bloated in an excessive amount. The best way I found is therefore to Select a portion of neutral background close to the star, Feather it, put it above the star, use a layer Mask, reveal all and then paint it just above the star so you get a pinhole. Thus you will have a better looking star without bloating, and you can do it after stretching.  

Is this a "normal practice" for you pro´s out there, or is it regarded as "too much"?  I mean, I wouldn´t want to alter too much of an image. 

Comments, suggestions and tips are welcome! 

/Erik

[MarsG76]

Hello Erik,

The thing that promotes star bloat in subs when using a DSLR (and CCDs for that matter) is if you expose at high ISOs... I personally take my subs at no higher than ISO400 for RGB and ISO800 when doing narrowband on my modded Canon 40D. The only time I'll increase the ISO to 1600 is if a particular narrowband exposure is too dim to give me any useful data in 30 minutes exposure, but that is rare.

Even the anti blooming CCDs apparently can have star bloat, its just suppressed so its not as bad...

I know its tempting to use high ISO and shorter subs to get your nebulosity than stack multiple subs, but as you increase you gain you drop dynamic range, increase noise and are more prone to star bloat.

That said, you will always have a certain degree of bloating on brighter stars, that if unwanted, has to always be fixed in post processing.

When you stretch your levels, do not use the levels feature, in programs like Photoshop or Gimp.. always use curves. Curves will give you a lot more control over what part of the image you increase the levels on and this way you can create multiple points selectively increase the brightness of your nebulae, leaving your star levels unaltered.

To fix star bloat there are many methods people use such as to select stars and use minimum filter, make masks or duplicate unstretched stacks then use the Range select tool to leave the original star levels, but starting with curves to stretch your image, to me, is a very good starting point. Make copies of your original layers, that way you can go back to the previous step easily or use selective masks between the two and experiment with what works for you, you will find that, a lot of the time, different images require certain different processing methods.

Most astro images, especially the best ones you see, have had QUITE a bit of processing applied to them (even the Hubble images you see take the pros a few hours to produce the end result)... that's an unfortunate fact and necessity when imaging high contrast and dim objects such as celestial objects.

I know that so much processing can seem like a cheat, but as long as whats shown is the data, features and detail that was actually captured, the processing side is no different then adjusting contrast and brightness, sharpness etc in regular photography, like magazine, scenery or weddings for example applied by the photographer after the shoot.

I hope this helped you with your question.

MG

[Mr Bergman]

Thank you MG! I think the best advice then is to lower the ISO and try with longer subs as a start. Hadn't thought about that since I was recommended to use iso 1600 with my modded camera.

[ollypenrice]

The Hubble team spend a lot more than a few hours on processing, I suspect.   If an amateur like me spends a couple of days on an image then I think they'll do at least that much! And maybe a month...

Before thinking about reducing haloes and bloat, try not to create them in the first place. Think about what is going on in curves. When you lift the bottom of the curve you brighten the faint stuff more than the bright stuff. This can be useful. Learn what shaped curves produce what effects by experiment.

You can also mask stars. If you Google Star Masking you'll get some ideas to set you off. I never find they work perfectly for me so I play around with the techniques and find ones which do work for me. At best they are a help but I've never found a way to stretch hard under a star mask. A little, yes, and it's worth doing.

Olly

[Mr Bergman]

The Hubble team spend a lot more than a few hours on processing, I suspect.   If an amateur like me spends a couple of days on an image then I think they'll do at least that much! And maybe a month...

Before thinking about reducing haloes and bloat, try not to create them in the first place. Think about what is going on in curves. When you lift the bottom of the curve you brighten the faint stuff more than the bright stuff. This can be useful. Learn what shaped curves produce what effects by experiment.

You can also mask stars. If you Google Star Masking you'll get some ideas to set you off. I never find they work perfectly for me so I play around with the techniques and find ones which do work for me. At best they are a help but I've never found a way to stretch hard under a star mask. A little, yes, and it's worth doing.

Olly

Thanks Olly! Yep, you are right about doing the curves right. I am doing some progress now. But although I am very careful I get some halos at the end. Made some experiments with curves yesterday, but since the halos/bloating occurs mostly in the red channel, they get worse when I try to dig out the nebulosity which is mostly in that channel. This is not easy! I was thinking about focussing or deep sky stacker causing it, but the blue channel is perfectly focussed. Hmm...

[Merlin66]

Erik,

I think your comment:

but the blue channel is perfectly focussed

may be significant.....

What telescope optics are you using?

When you use anything with lenses involved the chromatic aberrations can cause major differences in focus across the spectrum

In my spectroscope a 200mm fl achromat can have a +2mm difference in focus between Cak and Ha.

[Mr Bergman]

Erik,I think your comment:but the blue channel is perfectly focussedmay be significant.....What telescope optics are you using?When you use anything with lenses involved the chromatic aberrations can cause major differences in focus across the spectrumIn my spectroscope a 200mm fl achromat can have a +2mm difference in focus between Cak and Ha.

I use primary focus with my dslr on a 190mn. The camera is baader modded so the baader filter is exactly as thin/thick as the original. The focus point should be the same regardless, at least of what I know! Any ideas?

[Merlin66]

Erik,

I don't think it has anything to do with the camera/ filters.

It could be that the front corrector plate on the MN is giving some chromatic aberration.

Easy way to check is to focus with a blue filter somewhere in the optical path and then do the same with a red filter. This will show any re-focusing requirements.

HTH

[Mr Bergman]

Ah ok. Will check the focussing. Thx!

[MarsG76]

Thank you MG! I think the best advice then is to lower the ISO and try with longer subs as a start. Hadn't thought about that since I was recommended to use iso 1600 with my modded camera.

I think just experiment with different settings... I did see many images taken at ISO1600 and they looked great although they were all RGB, natural color.... I'm going more for the lower ISO because doing 30 min subs on a uncooled DSLR at high ISO levels introduces way more noise, than I do need that time since I'm doing NB... For RGB the exposures are much shorter and I try to get various levels of gain at times where either I don't have blown out stars and/or the light pollution doesn't blowout the sub...

After stacking I have a easier time to get the faint object without blowing out stars.

I'm sure that cooled CCD principles are different...

[Mr Bergman]

Ok, so I tried a few new processing types, and actually made my stars a little bit better. If you have a look at this one, and then compare with previous version on astrobin, I am sure that this bloat is normal and due to my lack of  processing experience. Never mind the colour, might be off a little bit. 

http://www.astrobin.com/full/203940/B/

/Erik

[MarsG76]

That look very good Erik... I think your processing is just fine.