StarTools

Hi Ivo,
Just thought I’d post in this forum rather than IIS
I’ve had mixed results with OSC narrowband processing via the Compose Module
First things first here’s my procedure below to get started and load the data in Compose
Please confirm if it’s correct and if so I will process one of my latest images ( give me a day or so ) and show you my results for comment and evaluation

Startools Compose Module version 1.7 ( Duoband Data using a DSLR or OSC camera )

1/ Open Startools
2/ Select Compose at the top
3/ Luminance, Color change to L+Synthetic L From R (2xG)B,(GB)(GB) (Bi Color from OSC DSLR )
4/ Color Channel Interpolation set to On
5/ Select Luminance at top and load your Duoband data set by selecting file and press Open
6/ Select Red / SII at top and load your Duoband data set by selecting file and press Open
7/ Select Green / Ha at top and load your Duoband data set by selecting file and press Open
8/ Select Blue / OIII at top and load your Duoband data set by selecting file and press Open
9/ Once all channels are loaded press Keep at top
10/ Type of Data set press Linear
11/ Now start processing your image as normal , you will be processing in mono until you hit the Color module , then you have many presets to choose from including Bi colour, HOO , SHO etc...

PS: I’m well aware that a OSC with L Extreme filter can never match the performance and results of using a Mono camera and separate NB filters for true narrowband images

Await your reply
Thanks
Martin

Hi Martin,

It would be helpful if you could explain or show what you mean by "mixed results".
How was the outcome , what did you expect it to be like..

about your method, not sure if you should load the same file in all slots in Compose,
maybe being an L-eXtrem, so Ha and OIII perhaps only Luminance and R + B ?

in any case, did you try to do your workflow starting by simply "Open" button, instead
of "compose" ?

I haven't got the filter yet myself, but had one L-eNhance borrowed from a friend to play with
This example is the result of 30x5min stack, processed straight from "Open" button, and then
in "Colour" module, click the "Bi-color" preset, and tweak your saturation levels as you wish. Not agreat photo, and still noisy, but as the processing goes.. not sure what would the benefit be
doing the same from "Compose" in this case.

Regards

Carles.

Much appreciated Martin! I'm looking forward to the notes/images.

The Compose workflow looks about right, though you probably shouldn't load the dataset as luminance as well. A properly weighted synthetic luminance is already created (weighted as R + 2xG + B). Loading any dataset as "real" luminance will interpret that dataset as the basic sum of all channels (R + G + B) - really a mono dataset is expected here. The synthetic luminance is then added to that "real" luminance, so this will give an incorrect weighting.

The L-Extreme should definitely be able to get fairly close to single-filters, as it really closely matches a combined Ha and O-III filter (obviously results will also depend on how strong emissions are in each band for a particular object; not all objects are suitable, and strong O-III is much more elusive)
Of course, only a quarter of your sensor's pixels is collecting Ha (and half for O-III, though blue may pick up some O-III as well), so you will likely want to bin to 50% X/Y at least. That said, your OSC may have a bump in response to Ha in all channels (not uncommon), which helps Ha sensitivity in the other channels.

These duoband filters are such great tools!

Ivo,
Thanks for replying and detailed information
I’ve just re processed my data on NGC 6357 Lobster Nebula identically using 2 options -
1/ OSC Linear data set , Open and Bi colour preset in colour module
2/ Compose , L+ Synthetic L from R (2x G) B, R (GB) (GB) Bi Color from OSC DSLR and Bi colour preset in colour module , loading data set just in R,G, B ( nothing in Luminance slot )
Captured 4 hours ( 3 minute subs ) using my 2600MC and Optolong L Extreme filter under Bortle 3 skies , good sky conditions

As you can see from the two images the Compose image is significantly better with improved colour contrast, but it does not look like a narrowband image as such. There is no distinct bi colour in the image as far as I can see or maybe I’m asking to much from my OSC in producing distinct narrowband contrast and depth.Also Stars are green in both images , which could be rectified by using a star mask.Don’t get me wrong I’m very happy with the image detail and depth , it’s just not what I was expecting in narrowband processing. Mono may have to be my next venture forward
I didn’t tweak the sliders in Colour module , just hit the Bi Colour preset button
Your comments would be appreciated
Thanks always
Cheers
Martin

Startrek wrote: ↑Thu Jul 08, 2021 5:29 am As you can see from the two images the Compose image is significantly better with improved colour contrast, but it does not look like a narrowband image as such. There is no distinct bi colour in the image as far as I can see or maybe I’m asking to much from my OSC in producing distinct narrowband contrast and depth. Also Stars are green in both images , which could be rectified by using a star mask.

Thank you Martin,

The important (and convenient!) thing about using the Compose module in conjunction with the Color module, is the following;

The red/green/blue bias sliders map directly and linearly to the strength of the bands. In other words, you linearly control how strong you wish to make the coloured emissions in each band poke through.

The nature of narrowband imaging, is such that, by default, Ha will be overwhelming (as we can see in your textbook HOO image). Its signal is almost always extremely strong and abundant, swamping any other bands. Other bands such as O-III? Not so much, except in the stars, which are always far more visible in the O-III band. A default "color balanced" HOO rendering therefore tends to juxtapose strong O-III stars with Ha detail.

As alluded to, the colouring of your image is precisely the expected colour for a HOO mapping rendered by the Color module by default. That does not, however, mean you are stuck with that Ha to O-III ratio.

To dramatically change the rendition to include more O-III, all you need to do is increase the blue (or green - they both control O's contribution under a HOO mapping) bias. The unique/important thing is that this linearly controls the strength of the O signal purely for the colouring.

The importance of the precious linearity, is that it makes sure that the image stays "true to life" in the sense that the O-III to Ha ratios remain consistent across the image, regardless of brightness. In other words, you may be manipulating the O-III to Ha ratio (to boost O-III prevalence), but you are doing it for the entire image in the same way, therefore allowing the viewer to get a true and accurate sense of the relative strength of emissions across the image.

As said, the cyan stars are part and parcel of a HOO narrowband rendition (in a SHO rendition, they are purple instead for the same reasons cited; stars barely show up in Ha, but show up much more in the other bands).

Indeed, there are many ways to manipulate the appearance of those cyan stars (Layer module with a star mask to desaturate them, Color Taming in Shrink module, etc). For the purpose of deciding on your final coloring, you can/should ignore the stars for now - their (some would say; obnoxious) colouring is a known issue.

One note of caution however; there are plenty of "tutorials" out there by people that are entirely oblivious to what the colouring in a narrowband image actually means/does. The point of a narrowband mapping, is to let the viewer distinguish which emissions occur, where. Aforementioned "tutorials" usually will happily process the individual channels differently, in a non-linear fashion, to create dramatic colour changes, inventing new hues in the process, and completely destroying any useful information that the colouring may impart. The resulting images have 0 bearing on reality and, while pretty, are not of documentary value, falling firmly on the "art" side of the spectrum.

Do not think, however, that staying true to reality therefore must consign NB images to be limited or boring. Linearly throttling emission contribution, combined with more adventurous channel remapping (beyond HOO) can yield absolutely spectacular shots. All while being of true documentary value.

If none of this helps, then if you'd like to share any datasets with me, I'd be happy to have a look at what can be done in the Color module, etc.

Clear skies!

Ivo,
Thanks for your comprehensive reply in explaining narrowband processing and colour in general
The Compose process certainly produced a superior outcome which I’m happy about.
I’ve been enjoying using V1.7 for about 5 months now and still finding out many useful features
I noticed V1.8 is in development phase , any idea when a stable version is likely be released ?
That’s again for you help, you are always accomodating
Clear Skies
Martin
( aka Startrek )

Great to hear it was helpful Martin.

Startrek wrote: ↑Thu Jul 08, 2021 7:39 am I noticed V1.8 is in development phase , any idea when a stable version is likely be released ?

I'm aiming for a stable version towards the end of the year, though I'm hoping to release a public alpha soon.
If you're doing narroband, the new Narrowband Accent module is a very welcome addition.
It makes the most difficult use case for narrowband (adding it to visual spectrum data for the purpose of accentuating features) a lot easier, while also allowing you to do more exotic stuff with ease, like LHO composites. The L-Extreme filter is a perfect companion for this module!

That’s again for you help, you are always accomodating

Any time! I think we're all passionate about this hobby/field, so it's my pleasure.

Hi Ivo,
Bad weather at the moment so took the opportunity to reprocess some good data I captured on M16 Eagle Nebula a week or so ago using my ZWOASI2600MC and Optolong L Extreme filter. Used Compose with L + Synthetic from R (2xG) B (GB) (GB) Bi Color from OSC DSLR
Loaded data set into R,G and B then selected Linear
Processed in Synthetic Luminance mono until I hit the colour module , created a good star mask , inverted, and used the following Matrix -
Image 1/ SHO 40SII+60Ha,40Ha+60 OIII, 100 OIII
Image 2/ SHO 60SII + 40 Ha, 30SII + 40OII , 100 OIII

Just tweaked the red and green bias up , blue bias up a bit and bright and dark saturation up a bit

Very happy with the resultant OSC narrowband Bi Color images

Thanks again for your help
Cheers
Martin

Exquisite!
Beautiful to look at and genuinely informative at the same time.

Ivo,
To kind sir
Kudos to you for your this brilliant software
Keep developing it ( with as many User requests as possible )
Clear Skies
Martin