Not getting the results I'd like with LHaRGB

[BrendanC]

Here's what I came up with after following Ivo's log, plus sharp, HDR etc. I think I still need a couple more goes at this, but definitely this is much closer to what I had in mind.

Honeyview_lhargb.jpg (55.87 KiB) Viewed 5570 times

I didn't really know what I was doing with the Visual Spectrum Narrowband Accent Compositing, but I can see now how you reduced the noise with the auto dev, and then I was surprised at how much you increased the Gamma by, but I can see that this was key to bringing out the Ha.

Also, I never would have considered using the Legacy setting for the Color. I'll have to play around with that a bit, and see if I can get a similar result but without using Legacy.

Two quick questions:
* I realise the Ha is noisy, and this is because it was shot with a variety of exposure times (because I'm still testing), in a full Moon, close to the horizon. But, is it very much noisier than you would expect? Would the fix be to use shorter exposures under those conditions and/or more exposure time?
* For the Wipe module, is it 'normal' to have to go up to 94%? It's very much higher than I used with the DSLR. I'm happy if this is within normal bounds but it seems very high, and I just want to make sure my acquisition is up to snuff before I start trying to compensate during processing.

Anyway, thanks everyone again.

Cheers, Brendan

[admin]

I realise the Ha is noisy, and this is because it was shot with a variety of exposure times (because I'm still testing),

Never, ever mix exposure times! I suspect it's also one of the reasons why @Mike in Rancho had trouble keeping the stars from being affected; normally stars in don't show up much in Ha and therefore will not get accentuated.

* For the Wipe module, is it 'normal' to have to go up to 94%? It's very much higher than I used with the DSLR. I'm happy if this is within normal bounds but it seems very high, and I just want to make sure my acquisition is up to snuff before I start trying to compensate during processing.

It's not usually needed or normal, but that's why I said;

Parameter [Gradient Aggressiveness] set to [94 %] (too lazy to crop away the stacking artifacts again )

E..g. I was too lazy to go back to the Crop module and crop away the stacking artefact, and instead lazily opted to ameliorate it by using a more aggressive Wipe.
You can away with this if the detail in your image is confined to small areas, see docs;

Gradient modelling and subtraction

Wipe discerns gradient from real detail by estimating undulation frequency. In a nut shell, real detail tends to change rapidly from pixel to pixel, whereas gradients do not. The 'Aggressiveness' specifies the undulation threshold, whereby higher 'Aggressiveness' settings latch on to ever faster undulating gradients. At high 'Aggressiveness' settings, be mindful of Wipe not 'wiping' away any medium to larger scale nebulosity. To Wipe, larger scale nebulosity and a strong undulating gradients can look like the same thing. If you are worried about Wipe removing any larger scale nebulosity, you can designate an area off-limits to its gradient detection algorithm, by means of a mask that masks out that specific area. See the 'Sample revocation' section for more details.

Hope this helps.

[BrendanC]

Thank you again Ivo, this is really helpful.

Re the Ha exposures, strangely I noticed that the shortest and longest were best after testing. Then, like a fool, I decided just to throw them all into the mix. I might split them out and see what effect that has.

Re the Wipe amount, well @Mike in Rancho very kindly helped me out with this image previously, but also used a 94% wipe even though he'd cropped enough. I've since discovered I need to use 94% wipe for other LRGB images taken with this kit too (130PDS with ASI1600MM-Cool, ZWO 1,25" LRGB and Baader HSO filters). I'm still very much learning how to get the best out of the camera, having upgraded from a modded DSLR, and I'm just concerned that there's still something amiss with my acquisition. So, if you're saying 94% is unusually high, and I'm finding I need to keep using 94%, then what might be the fix?

[Mike in Rancho]

Interesting!

Brendan if you create a new, non-Frankenstein Ha file, it would be fun to try to process this again and see if the undo buffer can be dispensed with. No red-eye stars!

I may be using Wipe as too much of a sledgehammer, but, well you know my data isn't the greatest. But even with other data, except for truly clean data, I often end up at or near 95% for broadband. On my own stuff, well I'm still trying to get this Newt figured out as far as light leaks, stray light, and flats. Also, this being galaxies-on-a-star-field season, I can sometimes pump it up to 97-99%, depending on just how bad my gradients are. I know I am not picking up IFN , so not too worried there, and most of the time the galaxies themselves "seem" unaffected, even without a protective mask.

For narrowband and/or nebula, of course I have to back Wipe off from anything like that so as to not erase the good stuff.

[BrendanC]

Cool - here's the non-Frankenstein Ha file, just made up of the 105s exposures: https://1drv.ms/u/s!AqovBuVZMwj3k4gErlm ... A?e=6fbEna

Interesting that you also need to go that far with the Wipe. I never needed to with the old DSLR but can totally see that the ASI1600 is more sensitive. I'm in a Bortle 4 zone but surrounded by houses, so it's not an ideal location. I actually moved the scope's location the other day to hide it from the most intrusive light so let's see if that makes a difference.

Here's where I'm up to with the Bode's and Cigar galaxy. I'm finding that the Legacy colour is definitely better for combining Ha with a galaxy, at least with my data. I've also decided RGB stars are better without the Luminance, which is what I've used in the background for this version.

Honeyview_bodes.jpg (36.94 KiB) Viewed 5539 times

[Mike in Rancho]

Nice job, Brendan.

I did try your new Ha file, but I'm not sure I ended up in a hugely better place than the more combo Ha version, as the outer stellar profiles were still turning red on me. Even trying out different parameters on both of the NBAccent screens, I couldn't find a way to eliminate that without also eliminating the little Ha blobs that I wanted to show in Bodes, and the bursting structures about M82.

Guess I need to learn some more or experiment further. Absent that, it would have to be the undo buffer to revert the stars.

[admin]

With regards to the Wipe's Aggressiveness setting, for this particular dataset you do not require to vary the default 75% at all, except that it is necessary to engage some sort of vignetting compensation (the Vignetting default works just fine), as there is a visible discrepancy in brightening towards the corners.

E.g. after a crop, you should be able to get a good result by clicking the Vignetting preset and increasing the Dark Anomaly Filter a little (3 px was enough for me).

In general, always remember that Wipe's Autodev will relentlessly show you anything that it can see is "wrong". Even if that signal is just 0.000000001% above the noise floor. So, for example, as far as the coloring goes, this is just fine;

StarTools_2829.jpg (382.3 KiB) Viewed 5501 times

(ignore the German interface - working on multi-language support right now )

Of note is also that, if using an alternative luminance signal (e.g. importing L, or using Ha as L, etc.) that does not correspond to RGB coloring (or does not correspond well), then the default "LRGB Emulation Method" (which is Straight CIELab Luminance Retention) may not work very well. That is because it now (erroneously) uses a different luminance that is decoupled from the inherent RGB luminance for the CIELab space calcuations (e.g. the calculations assume a pixel's brightness is closely related to its RGB color values, in order to recover/transpose the original colors). When this is not the case, things may go wrong and it is better to use a different LRGB Method Emulation. The Legacy preset does exactly that (it switches to "RGB Ratio, CIELab Luminance Retention", hence why you are getting better results using that preset. E.g from the Color module docs;

When processing a complex composite that carries a luminance signal that is substantially decoupled from the chrominance signal (for example importing H-alpha as luminance and a visual spectrum dataset as red, green and blue via the Compose module), then the 'RGB Ratio, CIELab Luminance Retention' will typically do a superior job accommodating the greater disparities in luminance and how this affect final colouring.

Hope this helps!

[Mike in Rancho]

Good tips, Ivo.

I've continued to play with this data. Good practice at least, and though I've been messing it up a lot, I'm getting a bit more comfortable with using the NBAccent in Galaxy mode.

I had to restart a couple times after I realized I was putting the filter exposures in wrong when following my prior log, or your log. The logs seem to write the channels out as LBGR, but of course the controls in front of us are always normal LRGB. I wonder how many times I've messed that up.

The addition of the wipe vignetting control was indeed the key here. I should have looked at the luminance frame closer and realized that. Still, in galaxy season I think we have some more leeway here, perhaps. There have been times where those remnant color anomalies across the field still seem to end up altering the final image coloring for me. So I've usually tried to flatten out the color frame also when in wipe, which can often lead to some higher aggressiveness in broadband. But I'll experiment more.

I've seen the note about RGB Ratio when the L and C are decoupled such as HaRGB, but here the main composite is a normal LRGB, with known exposures even. So in this case using Ha only in NBAccent, is it because the accent file will "adopt" the emulation method that resulted from Color, or be affected by the results from that scheme (even though in NBAccent we do have power controls for L and Color affect)?

Anyway I ran my last attempt through using the 75/75 Wipe with Vignetting and then full process. It still took me a lot of tests to get a poke-through that seemed a good balance and didn't turn the core of M81 into a tomato. I used cancel a lot to "restretch" the Ha. In the end though I couldn't avoid red stars, so resorted to the undo layer.

It also may have brought some more noise into the image. I thought about running another SS, perhaps Isolate, afterwards, but just bumped the denoise some instead. I wonder if SS might be best off after NBAccent overall?

I'm not sure it looks terribly different than my prior version, though it seemed that changing the detail pixels lower gave me a better look to the accenting of M82, and didn't bleed about as much.

BrendanC NewHa-LRGB M81-82 ST8 1B.jpg (403.58 KiB) Viewed 5485 times

[Stefan B]

I'm reading this thread with great interest, thanks to everyone!

Still, in galaxy season I think we have some more leeway here, perhaps. There have been times where those remnant color anomalies across the field still seem to end up altering the final image coloring for me.

Just wanted to add that I experienced the same, especially with frame filling nebulae. A flat color field in Wipe is extremely helpful when entering the Color module. Higher aggressiveness and higher DAF values (and to some extent gradient falloff) did help me in these cases.

It still took me a lot of tests to get a poke-through that seemed a good balance and didn't turn the core of M81 into a tomato.

That's a huge issue IMHO. In my M81/M82 I unceremoniously masked M81's core and used the undo layering, since otherwise it would have looked ridiculous. Certainly not documentary...

I read about the process of continuum subtraction for cases like this. There, you subtract the red signal from the Ha signal so you aim to retain all those HII regions in the arms but get rid of the signal in the Ha line which comes from continuum emission instead of line emission. As far as I understood you use stretched data for this. But the amount of stretching isn't objective, especially when gamma etc. is manipulated (or some parts of DSOs are highlighted or tamed by HDR in (L)RGB but not in Ha).

@admin Ivo, is there a possibility in ST to do something similar to continuum subtraction? Or to avoid overly red galaxy cores?

I thought about running another SS, perhaps Isolate, afterwards, but just bumped the denoise some instead. I wonder if SS might be best off after NBAccent overall?

I got used to do exactly that. I know it's not recommended but super structures are introduced by NBAccent and these get nicely highlighted by the SuperStructure module.

Regards
Stefan

[BrendanC]

Hi all,

Looks like I've kicked off quite a debate! And thanks again Mike for looking at my data.

I have to say I removed tge data yesterday because I didn't realise this thread was continuing, so I just uploaded it again in case anyone else wants to have a look: https://1drv.ms/u/s!AqovBuVZMwj3k4ofrQG ... w?e=b97sgj

Meanwhile, some questions!
@Mike in Rancho , would you mind sharing your log, or extracts of it, to show how you got your image? (Oh, and btw, I've done exactly the same as you in the past with regards to the log writing out the channels as LBGR!)
@admin - great suggestion about vignetting in Wipe, but in that case, does this mean my subs are displaying excess vignetting?
@Stefan B - when you say 'That's a huge issue', are you talking about my data specifically, or just generally?

As this is my first attempt at LHaRGB it's been really useful to go through this exercise - trying to iron out any problems by getting one image right rather than jump from object to object. The other aspect of all this is that I'm still familiarising myself with the new camera (ASI1600MM-Cool) and so I'm trying to understand whether any issues are down to acquisition or processing, hardware or software. So, if anyone could let me know any way in which my data could be improved, and whereabouts in this chain I could improve it, that would be wonderful.

Thanks everyone again.

Cheers, Brendan