Hi Graham. Try a wavelet transform on that galaxy. I have ST open in front of me and note the wavelets tool.
Ivo can assist here. If you apply just 5 layers or the equivalent the detail should pop. Also, the green is post deBayer noise. ST has a way of getting rid of that to.
With >3 hours there should be lots of detail to see. You might be surprised whats there.
I'm not sure either Graham. I'm a Pixinsight user. Might be a good question to ask of Ivo on the ST forum.
Big difference all the same. Galaxy colours and some additional detail are coming through. Gradient could do with some work, but all-in-all a big improvement.
Where the heck do you get all these clear skies from to do all this!? You're a machine!
Just responded to your PM re: gradient removal.
Wipe should definitely be able to do the trick (see example).
EDIT: to remove the gradient using ST, just launch Wipe and click 'Do'. Click the 'Fill mask' button if asked - it is often unnecessary to mask out objects and/or sample specific parts of the sky unless they obscure the real sky (buildings, mountains, trees, dust donuts, etc.). You can play with the Aggressiveness parameter if Wipe is being too gentle to your liking. I also bumped up the Wipe's noise/dead pixel filter somewhat as very noisy images and/or JPEG artefacts can 'scare' it into being too gentle. All this should remove the gradient effectively and you''ll be left with an even background. Note that there is a fair bit of chroma (color) noise present. Wipe does not remove this, but you can use the Chroma Noise Filter in the Development module. I further used ST 1.2's upcoming Wavelet Denoising algorithm and used the Isolate preset in the Life module (without a mask).
Rowland is giving some good advice with regards to Wavelets. Wavelets let you manipulate features of different sizes in your image, while leaving others alone. Indeed, they can be really useful to emphasize a galaxy's spiral arms. NGC300 in the JPEG posted is rather small, so it's a bit hard to apply it in there as an example.
Just a quick post on what Wavelets do and how you can use them.
You can think of wavelet manipulation like having a graphic equalizer for images. Just like a graphic equalizer lets you modify or isolate specific frequency bands in a song (boost/cut bass, treble, etc.), Wavelets let you do the same thing with images (and more).
You can think of the larger scale structures (for example spiral arms) in your images as the bass (e.g. the lowest frequency band). You can think of small scale structures (small stars, single pixel blips and noise) as the treble (e.g. the highest frequency band).
If you have a song with a great bassline but you can't hear it very well because it's very faint, you could turn up the lower frequency band a bit.
Similarly, if you have a great image of a galaxy, but you can barely make out the spiral arms, you could turn up the largest wavelet scale a bit.
Wavelets in image processing can be useful for a lot of things, but they can be somewhat of a crude tool. They don't really solve or correspond to a real world problem (e.g. a problem that was caused by your rig, skills or imaging conditions). Just like a graphic equalizer will never allow you to 'boost all violins' or 'cut all backing vocals', wavelets will never allow you to 'boost all galaxies' or 'cut all stars'. There will always be unintended consequences.
Another problem is that, because Wavelets operate on the whole image, boosting one frequency band in your image can totally overpower another, actually making your image worse in some places.
If you combine wavelets with a mask, however, you may get a bit more mileage out them. This way you can manipulate only those places in your image that you think can benefit from using a wavelet operator.
The upcoming 1.2 version of ST will have a revamped wavelet module that is a bit more intelligent with what is being boosted, only boosting a pixel if it actually results in better detail.
Look at those galaxy colours - nice work. Great analogy Ivo. I understand wavelets much more now.
Perhaps, isolate the galaxy from the stars with a star mask and protect with a luminance mask. Star mask + luminance mask, then do the wavelets that way.
First try 5 layers and then 2 layers, that way you get at a greater range of frequencies/scales, without affecting the entire image.
Just to clarify, in Star tools use mask /auto/sat for star mask, and flood fill lighter pixels for luma mask -If I understand you there is a way of having two mask sets at the same time? I.E a mask over the galaxy and a star mask?or do you do them consecutively?
As for my attempts at flattening the background, it is maddening!
Each time I use wipe, (classic luma) and increase the aggressiveness I am getting more gradients/ colour bloom including a fog of green which didn't appear before.
I get to aggressiveness 98% and some of the bloom goes but then so does some background stars!!
I
Any tips?
Just to clarify, in Star tools use mask /auto/sat for star mask, and flood fill lighter pixels for luma mask -If I understand you there is a way of having two mask sets at the same time? I.E a mask over the galaxy and a star mask?or do you do them consecutively?
I think what Rowland is suggesting is to use Wavelets just on the Galaxy by creating a mask. Additionally he is suggesting (I think - but correct me if I'm wrong Rowland!) to eliminate any stars that could potentially still be selected in the mask.
You're right Graham that in ST it is sufficient to use the 'flood fill lighter pixels' brush in the mask editor to select just the galaxy (no stars should be selected that way, except maybe some stars that sit 'in' the galaxy).
However, if you would like to further remove any stars (EDIT: from the MASK - not from your image!) afterwards (per Rowland's suggestion), you can click Auto, click the 'Stars' preset, set the 'Old Mask' parameter to 'Subtract New From Old' and click Do. What this does is create a star mask like usual, except it subtracts the newly generated star mask from the mask you had before (e.g. it punches holes in your previous mask where stars are detected). Now you can be sure that just the galaxy is selected without any stars.
Quote:
As for my attempts at flattening the background, it is maddening!
Each time I use wipe, (classic luma) and increase the aggressiveness I am getting more gradients/ colour bloom including a fog of green which didn't appear before.
I get to aggressiveness 98% and some of the bloom goes but then so does some background stars!!
I
Any tips?
I'm at a loss why my version of Wipe works fine with your data and your version is misbehaving! I'm assuming the steps in the e-mail I sent didn't help? (e.g launch StarTools, Load image, Crop galaxy, Wipe, set 'Top End Treatment' to 'Classic (Luma Mask)', Do.) I end up with the attached image, no further processing.
In 99% of the cases the default settings 'just work' (TM). So there must be something else going on.
I'm going to send you a link to a 1.2 alpha version - let's see if that solves your problem. It'll have some other perks you will find useful with this particular image (for example a brand new noise suppression algorithm that does wonders for stretching faint signals like these).
I don't mean to remove the stars, but protect them from wavelets. In PI I would create a luminance mask and a star mask, combine and apply to the image and then run wavelets. However, a luma mask may be adequate. Try that first and see what prevails.
I mean that masking the stars avoids wavelets on the stars themselves. A fairly strong mask with feathered edges is needed, otherwise the stars may finish up hole punched or have hard edges. A luma mask over the whole image combined with the star mask is necessary to protect other regions while having control over the galaxy wavelets The effect on the galaxy would otherwise be extreme and result in ring artifacts.
Try a luma mask first. But go with Ivo's directions. It's the principle that I am getting at, of protecting other data while you enhance another part of the image.
I don't mean to remove the stars, but protect them from wavelets. In PI I would create a luminance mask and a star mask, combine and apply to the image and then run wavelets. However, a luma mask may be adequate. Try that first and see what prevails.
I mean that masking the stars avoids wavelets on the stars themselves. A fairly strong mask with feathered edges is needed, otherwise the stars may finish up hole punched or have hard edges. A luma mask over the whole image combined with the star mask is necessary to protect other regions while having control over the galaxy wavelets The effect on the galaxy would otherwise be extreme and result in ring artifacts.
Try a luma mask first. But go with Ivo's directions. It's the principle that I am getting at, of protecting other data while you enhance another part of the image.
Cheers,
Rowland
Great stuff. That's exactly how I understood it Rowland.
Rowland, just out of curiosity, why would you in this case create a luminance mask (e.g. something that 'gradually' masks depending on the brightness of the source or destination pixel) rather than a 'hard' on/off mask for the galaxy?
Is there a reason why you would like darker (or brighter) parts to be less affected by the the Wavelet manipulation?
Is that to protect bright (or dark) areas from clipping and/or ringing artifacts in PI?
Always looking for new techniques/ways of doing things...
Ivo, my steps up till now have been -load image, screen stretch, bin to either 35-39%, crop stacking artifacts then-
develop to high 90's, some of the time I think I over 'develop' and the image from then on cannot be ' wiped' back to a nice neutral uniform background? Is this correct- what I want to do is max the galaxy signal but create uniform bg.
After develop I have tried with either wipe (no partial masks) or with a mask then processed from there.
NB I am using the TIF file from deepsky stacker not the jpeg.
Graham
Ivo. It is a case of protection. Wavelets are quite destructive without a mask, although the new Median wavelet transform tool is much better and pretty much ring free - still I would use a mask keeping other structures intact.
If you have a look at PI, you will notice that most of the native tools are wavelet based, and very effective for noise reduction and sharpening linear and non-linear images.
Not sure how that applies to ST, but wavelets are wavelets.
Ivo, my steps up till now have been -load image, screen stretch, bin to either 35-39%, crop stacking artifacts then-
develop to high 90's, some of the time I think I over 'develop' and the image from then on cannot be ' wiped' back to a nice neutral uniform background? Is this correct- what I want to do is max the galaxy signal but create uniform bg.
After develop I have tried with either wipe (no partial masks) or with a mask then processed from there.
NB I am using the TIF file from deepsky stacker not the jpeg.
Graham
You're doing all the right things Graham - no user error here as far as I can tell. I just sent you a link to a 1.2 alpha version. See if that makes any difference and see if you can reproduce the image I attached in my previous post. May I ask what version your are using currently (bottom left of the main/home screen)?
I know what you mean by 'over developing'. In the presence of a lot of light pollution it may feel that way - lots of stretching is needed to see any sort of signal that is otherwise drowned out. The Normalization filter should help a little (but it is conservative in order to not destroy color balance information/ratios).
Wipe, however, should save the day and my apologies for the grief it's been giving you.
Ivo. It is a case of protection. Wavelets are quite destructive without a mask, although the new Median wavelet transform tool is much better and pretty much ring free - still I would use a mask keeping other structures intact.
Understood. I was just wondering if there was any reason you choose smooth luminance mask (causing darker/brighter areas to respond differently) over an on/off type mask.
Quote:
If you have a look at PI, you will notice that most of the native tools are wavelet based, and very effective for noise reduction and sharpening linear and non-linear images.
Hehe, yeah I can't help but think they've gone a bit wavelet crazy I mean that in a 'to a man with a hammer every problem looks like a nail' sort of way.
Quote:
Not sure how that applies to ST, but wavelets are wavelets.
To be honest, Wavelets as found in PI and Registax are one of those tools that have always bothered me. They can be immensely unpredictable if you don't know what you're doing (and let's be honest most people have trouble grasping the math behind them and I can't blame them!). The end result is often an unnatural looking image with a dynamic range swinging all over the place; appropriate in some places in the image, destroying contrast in others.
This is because, by default, Wavelets have no regard for detail in an image They don't know where, in an image, to ease off or where to ramp it up. Bad luck if the flute and violin occupy the same frequency band - they all get the same treatment. The flute may sound great now but the violin's subtle overtones are drowned out by the comparatively louder prime note. In image terms the large spiral arms stand out more now, but the finer details within them have become harder to see (contrast has been reduced).
The bottomline is, "scale" alone is a poor way of isolating features in an image (random noise excepted!). If you have other, more intelligent means of identifying and isolating the features that you're after, then use those (smarter algorithms or user generated masks). If that is not a possibility or you can live with sub-optimal results, then sure, use wavelets.
Rowland, your example of using a mask and only then using wavelets is a very appropriate way of reining them in somewhat.
You can tell I'm passionate about this and I've indeed been quite reluctant to add 'PI-style' wavelets to ST for aforementioned reasons; they are very user-unfriendly, unpredictable and usually are a poor tool to solve your problem (bringing out detail).
ST uses a more up-to-date set of algorithms for multi-scale aware dynamic range optimization (the Optimize module) where it actually takes into account whether an individual pixel basis stands to gain from brightening or darkening across an infinite number of scales/bands. It determines the best-case (consensus) brightness level for a pixel based on all possible scenarios rather than picking just a single scale. The result is a more natural looking image where dynamic range is optimised across all scales. Best of all, the computer does all the hard work and the user doesn't have to tinker around with masks and abstract/nondescript sliders.
In musical terms, it automatically adjust the equalizer whenever the violin is playing to capture every nuance of the instrument, quickly changing to another optimal setting for the flute. Should both be playing at the same time, it picks the 'best compromise' setting that still enhance both without ruining the performance of either.
I'm assuming the steps in the e-mail I sent didn't help? (e.g launch StarTools, Load image, Crop galaxy, Wipe, set 'Top End Treatment' to 'Classic (Luma Mask)', Do.) I end up with the attached image, no further processing.
