By the way Mike - you have an exceptional eye for colour.
I could never find anything NQR with your processing.
cheers
Allan
Well actually, I have a healthy penchant for blue myself and have been guilty of showing it too strongly at times, not as much these days probably but you ask the IIS old timers about Mike and his bloody blueness
I agree - I think over did my repro - & lost too much of that lovely blue.
Here's 50/50 blend with the original.
cheers
Allan
Well, Allan,
Hmmm......I'm debating this still but I'm pretty sure I have my answer. I could take my image and "adjust" the colours to just about anything I wanted. I could make things green (yikes, sorry Mike & Trish but this isn't narrowband- so no offense!), but in doing so I think I should give up astrophotography and take up painting.
I don't dislike the blend but I do have this problem: The bright blue star to me looks like the right blue, maybe too intense, but the "right" blue. The blended image is now the "wrong" blue because my colour balancing said "this is the blue colour your imaging is producing" and not the "fiction" of messing with the image until it fits one's imagination. So, the blended version has two different blues when the camera saw just one blue. I find it odd that the star has a different blue and I'm pretty sure I don't want the star the blue colour of the galaxy.
In conclusion I am willing to adjust the intensities of red or blue showing, but I'm unwilling to change the basic colours. I think what I produced is closer to what my cameras saw. The other is fiction.
Hmmm......I'm debating this still but I'm pretty sure I have my answer. I could take my image and "adjust" the colours to just about anything I wanted. I could make things green (yikes, sorry Mike & Trish but this isn't narrowband- so no offense!), but in doing so I think I should give up astrophotography and take up painting.
I don't dislike the blend but I do have this problem: The bright blue star to me looks like the right blue, maybe too intense, but the "right" blue. The blended image is now the "wrong" blue because my colour balancing said "this is the blue colour your imaging is producing" and not the "fiction" of messing with the image until it fits one's imagination. So, the blended version has two different blues when the camera saw just one blue. I find it odd that the star has a different blue and I'm pretty sure I don't want the star the blue colour of the galaxy.
I have spent my career as a classical musician, a cellist. There are two artistic camps that are quite divided by principal. There are those that subsume their own artistic creativity to the composer doing their upmost to bring the score to life. The other camp says to the composer "thank you for writing this lovely piece but actually you got some things quite wrong so I will just change what you wrote in order to be exciting (and so more people will like me)." I'm quite sure you know which camp I belong to.
In conclusion I am willing to adjust the intensities of red or blue showing, but I'm unwilling to change the basic colours. I think what I produced is closer to what my cameras saw. The other is fiction. Sorry, maybe I'm deluded but I have for a lifetime know very much where to hang my hat.
Peter
Dear Peter,
it's hard to know what the correct colours are.
Also- I didn't spend much time on it -
I was looking at the galaxy not the stars.
In any case - you have an amazing picture from Sydney.
I only wish I could find the time, energy & weather to take some of my own.
My humble 10" f4 Newt. is just sitting there waiting to be used.
Those comparison images just go to show how far amateur imaging has advanced. Multi million dollar scopes and CCDs sitting on a near perfect mountain top site and its only a bit more detailed than Peters effort.
I'm enjoying this discussion! Thanks everyone for chiming in.
OK, here is a nice graph comparing the freq. responses of many cameras in use today. Is there any question as to why a Trius 694 image might look more blue than the comparison cameras? So I put it thusly. Who says the other cameras have got it right? Is it just that the majority of images on the web were not taken with a 694 CCD so we now think "reality" is not blue? Maybe the only way out of this conundrum is a set of RGB filters specifically designed for a specific CCD (remove-with the goal being that the filters should match the human eye's frequency response-remove***). Didn't Don Goldman design his series 1 and 2 RGB filters with a certain CCD in mind? I'm not sure anyone has designed a set of filters around the 694 CCD.
I don't think that it's a good idea to have the human eye response built into the imaging system. Then, when you look at the image with your human eyes, you will have applied the human eye response twice. I think that the best bet is to have as flat a response as possible in the imaging system, so that the resulting image will represent the source light distribution with as little spectral distortion from the imager as possible.
On that basis, the 694 would appear to have a distinct advantage and maybe the colours in your lovely image are closer to "right", whatever that means.
OK, here is a nice graph comparing the freq. responses of many cameras in use today. Is there any question as to why a Trius 694 image might look more blue than the comparison cameras? So I put it thusly. Who says the other cameras have got it right? Is it just that the majority of images on the web were not taken with a 694 CCD so we now think "reality" is not blue?
Collecting more blue photons improves the SNR of your blue master but doesn't necessarily make the overall image more blue. It may tempt you to make it bluer, of course.
Ideally, you'll be using some form of colour calibration to determine the R:G:B ratios. There are many ways to do this and some even have a roughly scientific basis, e.g. G2V calibration or eXcalibrator.
Quote:
Originally Posted by PRejto
Maybe the only way out of this conundrum is a set of RGB filters specifically designed for a specific CCD
Unfortunately, the filters and sensor characteristics aren't the only factor in play here. Atmospheric conditions, especially extinction, play a significant role as well. Having filters which flatten the response of the sensor wouldn't remove the need to do colour calibration.
Collecting more blue photons improves the SNR of your blue master but doesn't necessarily make the overall image more blue. It may tempt you to make it bluer, of course.
Ideally, you'll be using some form of colour calibration to determine the R:G:B ratios. There are many ways to do this and some even have a roughly scientific basis, e.g. G2V calibration or eXcalibrator.
Unfortunately, the filters and sensor characteristics aren't the only factor in play here. Atmospheric conditions, especially extinction, play a significant role as well. Having filters which flatten the response of the sensor wouldn't remove the need to do colour calibration.
Cheers,
Rick.
agreed, the spectral sensitivity of the CCD has no direct effect after colour calibration. However, if the SNR in the blue channel is poor, there is a good reason for winding the blue contribution down - lots of images on the web have distinct "minus blue" appearances, presumably because that way of processing has yielded a smoother and more detailed looking result (a couple of images of this galaxy look quite red overall). The other area where system spectral response makes a difference is where colour balancing is carried out on an RGB and this is then overlaid with a Lum channel. If the system has low blue sensitivity, the Lum will show reduced brightness in predominantly blue areas and that will not be compensated for by the colour calibration - low blue sensitivity in Lum will show up as reduced blue in the final result. Come to think of it, this is a bit of a worry, since the recommended PI LRGB process relies on colour calibration only on the RGB...maybe that needs a rethink.
I understand that Peter used automatic colour balancing and got the result he posted - he has resisted the urge to make it look more like other images, so it looks quite blue .
I guess that, along with atmospheric extinction, galactic extinction/dust can have an effect as well. Then if you balance on the galaxy, the stars will be wrong (and vice versa).
Hi Peter
I had another look at your image. The histogram says a lot:
1. All the R, G, B peaks line up perfectly, which indicates good colour balance.
2. There is some room to raise the black point.
3. The histogram peaks are very narrow, indicating that more stretching may be helpful.
Geoff
I like the sharp detail Peter and I'm one of those people who likes to see blue in a galaxy. It may be better with the blue slightly toned down. I'll have a go when I'm next at my computer.
Geoff
Quote:
Originally Posted by Geoff45
Hi Peter
I had another look at your image. The histogram says a lot:
1. All the R, G, B peaks line up perfectly, which indicates good colour balance.
2. There is some room to raise the black point.
3. The histogram peaks are very narrow, indicating that more stretching may be helpful.
Geoff
Thanks very much Goeff! I see what you mean about the narrow peaks. I've subsequently tried to stretch a bit more but I'm paying a very high price with noise. It increases dramatically. I've already pushed noise suppression as far as I'm able to. As for the black point, do you think I've set it too high? I like to leave some room and not have things too dark. My monitor might be misleading me, however.
Quote:
Originally Posted by gregbradley
I am not a fan of others reprocessing your image for you. Its a bit like someone redoing your painting for you. You are the artist, you make the decisions good or bad. Technical points like too much noise, needs more exposure, stars aren't round are fine. But subjective points like colour and framing and methods of processing are a bit crossing the line in my opinion.
There are so many ways to processing an image that its not really granting the artist and their work any respect by redoing it for them. If you don't like the result that is the area where the audience can say - no, don't like it.
If the artist is a newbie and is after advice and help that's a different story. But that is not the case here.
Your original image is fine and needs nothing.
Greg.
Thanks so much Greg! I think you are right with what you have said. What I posted I have been working on (3 reprocesses) for more than 2 weeks. On the one hand I don't mind seeing what can be done with the same data when starting from scratch and I'malways looking for ways to do better...like most of us I'm sure. Allan has certainly made me think about my reasons and my choices. And, at the end of the day that isn't bad either. So, I don't mind his effort in the whole context. It provoked a most interesting discussion and I've already learned a lot.
Quote:
Originally Posted by strongmanmike
Not a bad effort there Pete, some nice details visible. Not exactly sure about the colour because as you say what is the right colour? If I go with aesthetics alone I recon combining your result with ReproAl's might give the best of both worlds
Mike
Mike, I have appreciated your comments along the way and any compliment from you is gratefully accepted!! See, look what 2 telescopes can do when they decide to cooperate (for once!!). It "only" took me 13 hours of imaging instead of 26! That's nice considering the weather. With only one scope I wouldn't have this image that's for sure! You are officially "forgiven" now for suggesting this nonsense of putting 2 scopes on one mount!
Quote:
Originally Posted by Placidus
Hi, Peter,
The sharpness of your stars and galactic detail is quite superb. It speaks for a combination of excellent seeing (over the Pacific Ocean?), together with perfect technique. You've gone quite deep, too.
Regarding colour, I like an image to show the physics of what is going on. Strongly interacting galaxies will have powerful bursts of star formation, and consequently will have regions that are very blue compared with an elliptical galaxy. Also in general the core of a spiral galaxy should be much cooler and therefore less blue than the star-forming regions. Your image shows all those "internal" features.
But what about absolute colour? Should one aim to show the ellipticals and distant, bland, background spirals as a warmer colour to match their known physics, or is it fine to say that they are unimportant and worrying about them is missing the point of the image? I could be persuaded either way.
So a lovely, deep, startlingly sharp image that says a lot about the interacting pair. Really well done.
Best,
Mike
Hi Mike,
You are too kind! Indeed, occassionally the Pacific Ocean produces some unbelievable seeing. Sadly the light pollution extracts a heavy price. The one thing I can say about my observatory is that it is very convenient and has provided the perfect opportunity to learn. I think I've progressed some over the past few years. Many thanks again for your kind words! Now what I really want is a dark site to put my equipment!
Quote:
Originally Posted by gregbradley
Those comparison images just go to show how far amateur imaging has advanced. Multi million dollar scopes and CCDs sitting on a near perfect mountain top site and its only a bit more detailed than Peters effort.
Wow.
Greg.
I guess it goes to show that we are mostly seeing limited. I had a few exceptionally steady nights. That together with FocusLock allowed me to take full advantage. Perhaps it also shows that with enough imaging time certain disadvantages of light pollution can be dampened down.
Quote:
Originally Posted by Ryderscope
Fantastic to see what can be produced from the suburbs of Sydney.
Thanks!!! Yes, a worthwhile project.
Quote:
Originally Posted by Shiraz
I don't think that it's a good idea to have the human eye response built into the imaging system. Then, when you look at the image with your human eyes, you will have applied the human eye response twice. I think that the best bet is to have as flat a response as possible in the imaging system, so that the resulting image will represent the source light distribution with as little spectral distortion from the imager as possible.
On that basis, the 694 would appear to have a distinct advantage and maybe the colours in your lovely image are closer to "right", whatever that means.
Thanks Ray. Of course you are right. Another moment of not thinking the problem all the way through. Your image of NGC1532 is terrific. I notice how similar they are...I see similar star colours, small galaxy colours, outer arms & core. The main difference seems to be in the dust lanes where yours are consistently brown where mine are redder in one area with a kind of purple haze around the dust lanes. I've no idea why there is this difference when so much else is similar. Do you see the 694 chip as "flat" over the visual range? It doesn't look too flat to me, but I suppose it might be so in comparison to other chips. Thanks again for your helpful comments in my initial stages of processing!
Collecting more blue photons improves the SNR of your blue master but doesn't necessarily make the overall image more blue. It may tempt you to make it bluer, of course.
Ideally, you'll be using some form of colour calibration to determine the R:G:B ratios. There are many ways to do this and some even have a roughly scientific basis, e.g. G2V calibration or eXcalibrator.
Unfortunately, the filters and sensor characteristics aren't the only factor in play here. Atmospheric conditions, especially extinction, play a significant role as well. Having filters which flatten the response of the sensor wouldn't remove the need to do colour calibration.
Cheers,
Rick.
Hello Rick,
Thanks very much for the educational input. There is so much to consider!
Am I correct in thinking that if I'm collecting "blue" down to 380nm I am going to be "seeing" uv-blue where our eyes could never see blue? Would that not make certain areas of the imager look blue that perhaps ought to be a different colour (or white)?
I'd really like to know if you personally would keep the galaxy arms as blue as I've presented?
Thanks Ray. Of course you are right. Another moment of not thinking the problem all the way through. Your image of NGC1532 is terrific. I notice how similar they are...I see similar star colours, small galaxy colours, outer arms & core. The main difference seems to be in the dust lanes where yours are consistently brown where mine are redder in one area with a kind of purple haze around the dust lanes. I've no idea why there is this difference when so much else is similar. Do you see the 694 chip as "flat" over the visual range? It doesn't look too flat to me, but I suppose it might be so in comparison to other chips. Thanks again for your helpful comments in my initial stages of processing!
Peter
I agree, they are remarkably similar. I cannot guess why the dust lanes are slightly different, but both look fine to me.
no, the 694 is not really flat, but it is much more nearly so than competing designs. As Rick pointed out, flatness should not matter after colour calibration is done, but there are still issues to consider even then, particularly when using LRGB.
Thanks very much Goeff! I see what you mean about the narrow peaks. I've subsequently tried to stretch a bit more but I'm paying a very high price with noise. It increases dramatically. I've already pushed noise suppression as far as I'm able to. As for the black point, do you think I've set it too high? I like to leave some room and not have things too dark. My monitor might be misleading me.
Peter
Peter, have you tried masking the background while stretching? That should help with the noise. I don't think the black point is too high, it may not be high enough.
Wanted to compare against Allans work. With this I took the white balance of all of the stars in frame instead of the galaxy. Had to remove a bit of a green cast afterwards.
Wanted to compare against Allans work. With this I took the white balance of all of the stars in frame instead of the galaxy. Had to remove a bit of a green cast afterwards.
Hi Colin,
Well, to my eyes your version looks better than mine. Excellent. I'll give this a try. Would you mind telling me just how you went about this?
I think balancing against the stars is pretty valid and the result is quite nice.
Peter, have you tried masking the background while stretching? That should help with the noise. I don't think the black point is too high, it may not be high enough.
Honestly, no because I saw too much noise in the galaxy too. I'll try again. I assume more stretch would attempt to reveal fainter details...exactly where things get pretty noisy. Or, do you mean something else?
The black point.... I didn't express myself properly. You would like a darker background, correct?
Honestly, no because I saw too much noise in the galaxy too. I'll try again. I assume more stretch would attempt to reveal fainter details...exactly where things get pretty noisy. Or, do you mean something else?
The black point.... I didn't express myself properly. You would like a darker background, correct?
Thanks!
Peter
It's a bit hard to make predictions here because I don't have the fits file.The histogram looked a bit narrow to me, which was why I suggested the stretch, it may not be the right thing to do. I also think that stretching may brighten up the core too much, so you could try HDRMultiscaletransform to reduce the dynamic range. Don't get me wrong now. I think the image is very good as it is. I was just suggesting some of the things I would try--no guarantee that it will improve things.
Re black point. Slide to the right to nearer the base of the histograms. It may chop out some of the background noise or it may make the background too dark--experiment!
...see? told ya 
and have been guilty of showing it too strongly at times, not as much these days probably but you ask the IIS old timers about Mike and his bloody blueness 
. 
If I go with aesthetics alone I recon combining your result with ReproAl's might give the best of both worlds 
