Iris in LRGB

[cfpendock]

16 x 300 second subs each of LRGB.

Not a very good image in my opinion - processed in photoshop, I found the colour balance very difficult, but this is the best I could achieve.  

Better resolution here:  http://www.astrobin.com/full/225527/0/

Any comments on how it might be improved would be very gratefully received.  I've run out of ideas!

Chris

[ChrisLX200]

You did well, this is a tough object to get the colour balance right and the temptation is to pick colours individually and increase brightness/saturation of those, before long you find you're tweaking this and that, trying to correct some other bias you never noticed before and the whole process gradually goes to pot. Yep, I've been there :-)   I don't have any suggestions for photoshop, if you had Pixinsight I would suggest using Linear_Fit to balance the channels first, then combine to RGB. When you're happy with the appearance of that then add the L channel to get the detail back (L needs to be your best channel, it will contain all the detail while the RGB just colours it).

ChrisH

[MilwaukeeLion]

Id be very happy with that image, great job !

[cfpendock]

You did well, this is a tough object to get the colour balance right and the temptation is to pick colours individually and increase brightness/saturation of those, before long you find you're tweaking this and that, trying to correct some other bias you never noticed before and the whole process gradually goes to pot. Yep, I've been there :-)   I don't have any suggestions for photoshop, if you had Pixinsight I would suggest using Linear_Fit to balance the channels first, then combine to RGB. When you're happy with the appearance of that then add the L channel to get the detail back (L needs to be your best channel, it will contain all the detail while the RGB just colours it).

ChrisH

Thanks Chris.  I am encouraged that you also found it a challenge!  In fact I did exactly what you said with the different colours, and found it easier to re-start the processing from scratch each time it went to pot.  I found the L channel easily the most powerful for the detail, and I'm not unhappy with the actual nebula.  It's just the colours, particularly the background which I still find too red......

Chris

[toxic]

very nice image indeed

[laser_jock99]

Great image- I find the Iris a challenge, but mainly keeping the detail and exposure under control in the centre.

Hope you don't mind- had a very quick play with your jpeg image (just tweaking the individual channels as ChrisH puts it !) and feel you could probably get a lot more out of your data.

[ChrisLX200]

If you find you have a significant bias (as you have in the Red channel in the original image you posted) then you're not going to get rid of it using brightness/contrast/Hue&Saturation adjustments. You want to balance the histograms of the 3 channels so they occupy the same place (relatively) on the horizontal scale, but they will vary in height of course. The easy way in Photoshop is using Levels with a simple linear adjustment (the more difficult is using Curves for a non-linear adjustment). If you look at the R, G, and B levels (separately, not as the RGB combined image) you'll note the histogram for the Red channel is shifted to the right, so if you ove the bottom end slider to the right alittle then you will get this:

That by no means corrects all the problem but it does improve things. Using LinearFit in Pixinsight actually calculates the background levels and makes this adjustment for you (much more accurate that just manually moving the slider a bit!). If you make these adjustments 'by eye' the chances are when you leave it and come back to it the next day, or you make the adjustment at night then view it in daylight, or even view it on different monitors, then you'll probably think the adjustment you made wasn't correct. Oh yeah, been there too... and had to post a modded image :-)

ChrisH
 

...and had to post a modded image :-)   like now in fact... LOL

[cfpendock]

Great image- I find the Iris a challenge, but mainly keeping the detail and exposure under control in the centre.

Hope you don't mind- had a very quick play with your jpeg image (just tweaking the individual channels as ChrisH puts it !) and feel you could probably get a lot more out of your data.

Well, if you can do that with the degraded JPEG then I am seriously impressed.  I have tried re -balancing the histogram as ChrisLX200 suggested - originally I didn't do this because I found the stars became too blue.  However, revisiting it, I think my eyes were tired because of course it then became possible to stretch the image considerably more, showing much more faint stuff (while protecting the nebula).  But I find it impossible to pull out the level of detail which you achieved on the JPEG.   Did you just use curves and a masked noise reduction for this?

Or how?  

Anyway, thanks for the example of what CAN be done!

...... If you make these adjustments 'by eye' the chances are when you leave it and come back to it the next day, or you make the adjustment at night then view it in daylight, or even view it on different monitors, then you'll probably think the adjustment you made wasn't correct. Oh yeah, been there too... and had to post a modded image :-)

ChrisH

Yes, exactly.  I think leaving it alone for a time and then coming back to it makes you wonder what on earth you were doing.......

And ditto Chris about your own treatment of my degraded JPEG - incredible.  I don't have Pixinsight but normally I think I can produce a "reasonable" image with photoshop.  However, this target I find easily the most challenging I have done, and as I said above, even after re-balancing the histogram, I find it difficult to achieve the results which you gentlemen have obtained, and I am working with 16bits tif rather than a degraded JPEG!  

Thanks to both of you - it is precisely this kind of help which typifies the value of this forum.

I shall continue working on this image, but any further suggestions would be welcome.

Chris

[MarsG76]

That nebula looks great.... Like the light at the end of the tunnel or a portal into another universe... I'd be happy with the result you got.

I wanted to image this object but unfortunately it's never visible from my location.

[cfpendock]

I wanted to image this object but unfortunately it's never visible from my location.

And I could say the same about lots of objects in the southern hemisphere which are also out of range for us.......Never mind - it's an excuse to visit!

chris

[MarsG76]

And I could say the same about lots of objects in the southern hemisphere which are also out of range for us.......Never mind - it's an excuse to visit!

chris

True... How true...

[cfpendock]

So, thanks to comments from you wise people on this forum, I have now had a chance to follow the advice and have balanced the colours - this has allowed a lot more background detail to appear. I have learned - process the image and then go away from it for a time before doing anything else with it!  

Better version on http://www.astrobin.com/full/227927/0/

[toxic]

hope you dont mind but here is a very simple way to bring out detail just remove the stars in photoshop with anies astro actions - then save  it as no stars or what ever, then on the image with no stars go to filter / other / high pass and select aound 90 then apply and save as say highpass. now have you original image open and your highpass image (make sure they are both rgb/16 bit or at least the same) then while th original is highlighted go to image / apply image -in the box that appears select highpass image then merge as soft light click ok and your done.if it is a bit grainy then you can soften the highpass image a bit with noise reduction.hope this helps.

no other adjustments where made

[ollypenrice]

The long and the short of it is that the dusty stuff is infernally difficult and if anyone knew an instant processing fix they'd be rich! The first step is to take four times as much data. Twice as much won't cut it. Then you need to colour balance the true background sky. Humph, what might that be?? You have to make a decision and tell DBE. Then you can apply local contrast stretches in Ps, or at least that's what I do. Or you could ask Harel Boren how he does his dusty images because he blows most of us out of the water...

lly

[Mr Bergman]

Nice image! One thing you could try iis to download Regim, and tell it which object you are imaging, and then it will calibrate it to a G2V star and normalize your background. It works well and it is also freeware.

/Erik

[StuartJPP]

Good work there Chris, but as mentioned I think there is more data available than you are extracting...however one thing I don't like is when people make the faint nebulosity as bright as the bright nebulosity...faint stuff is meant to be that, faint.

[ollypenrice]

Good work there Chris, but as mentioned I think there is more data available than you are extracting...however one thing I don't like is when people make the faint nebulosity as bright as the bright nebulosity...faint stuff is meant to be that, faint.

Yes, though we only have so much dynamic range. I think there's a case for an imager saying, 'The point of my image is to show ... ' and then specify, perhaps, some very faint things we might not previously have appreciated even if the global brightnesses are rather distorted.

Olly

[StuartJPP]

Yes, though we only have so much dynamic range. I think there's a case for an imager saying, 'The point of my image is to show ... ' and then specify, perhaps, some very faint things we might not previously have appreciated even if the global brightnesses are rather distorted.

Olly

Agree about the dynamic range issue Olly (imagine M42 with no layering) but once the faint stuff is clearly visible then I'd stop selectively stretching them until they were brighter than the bright stuff that was masked off...I only know of a few people who go overboard with this and they aren't on this forum (that I am aware of). If something like an OIII region is to be emphasised then that would be okay.

At the end of the day what we create are "pretty pictures" so artistic license is allowed I suppose.

[cfpendock]

Thank you all very much again for the helpful advice.  I will try to get some time to make some more experiments - I think most is down to lack of experience.

hope you dont mind but here is a very simple way to bring out detail just remove the stars in photoshop with anies astro actions - then save  it as no stars or what ever, then on the image with no stars go to filter / other / high pass and select aound 90 then apply and save as say highpass. now have you original image open and your highpass image (make sure they are both rgb/16 bit or at least the same) then while th original is highlighted go to image / apply image -in the box that appears select highpass image then merge as soft light click ok and your done.if it is a bit grainy then you can soften the highpass image a bit with noise reduction.hope this helps.

Thanks Chris.  That is really useful information which I will be trying....

The long and the short of it is that the dusty stuff is infernally difficult and if anyone knew an instant processing fix they'd be rich! The first step is to take four times as much data. Twice as much won't cut it. Then you need to colour balance the true background sky. Humph, what might that be?? You have to make a decision and tell DBE. Then you can apply local contrast stretches in Ps, or at least that's what I do. Or you could ask Harel Boren how he does his dusty images because he blows most of us out of the water...

lly

Thanks Olly.  I was pretty sure lack of data (excepting lack of processing technique!) would be the case.  I am too impatient, but I guess I can try to increase the subs when the weather permits.  I also have an excuse in that my sky is not at all dark, and I find it difficult to more than 5 minute subs in LRGB.

Good work there Chris, but as mentioned I think there is more data available than you are extracting...however one thing I don't like is when people make the faint nebulosity as bright as the bright nebulosity...faint stuff is meant to be that, faint.

Thanks Stuart.  I agree exactly with that.  Yes, for sure there IS more stuff to be shown, but it is easy to locally overstretch the data which then makes the image ridiculously noisy -(need more data), as well as over-distorting balance of the image.  (this is just my excuse).....

One thing I can say for sure, processing this image has way exceeded my anticipated processing needs.  It has certainly been very worthwhile.  One day I might even pluck up enough courage to try M42.......

Chris

[ollypenrice]

Thank you all very much again for the helpful advice.  I will try to get some time to make some more experiments - I think most is down to lack of experience.

Thanks Chris.  That is really useful information which I will be trying....

Thanks Olly.  I was pretty sure lack of data (excepting lack of processing technique!) would be the case.  I am too impatient, but I guess I can try to increase the subs when the weather permits.  I also have an excuse in that my sky is not at all dark, and I find it difficult to more than 5 minute subs in LRGB.

Thanks Stuart.  I agree exactly with that.  Yes, for sure there IS more stuff to be shown, but it is easy to locally overstretch the data which then makes the image ridiculously noisy -(need more data), as well as over-distorting balance of the image.  (this is just my excuse).....

One thing I can say for sure, processing this image has way exceeded my anticipated processing needs.  It has certainly been very worthwhile.  One day I might even pluck up enough courage to try M42.......

Chris

M42 is great if you follow the helpful Mr Lodigruss... http://www.astropix.com/HTML/J_DIGIT/LAYMASK.HTM

Olly