Workflows for STC Astro DUO Filter

My Andromeda data has some bad frames which gave these squiggly lines. Doesn't selecting 90% remove these?

Here is a sample crop screenshot. Data uploaded 

Not sure it would necessarily remove those, the quality stacking looks at the stars mainly and SNR. You can try, but usually these lines are caused by hot pixels, which you want to remove before stacking by a good dark and/or bad pixel map. MAD sigma rejection might also help.

So I started processing your duo-filter images. The darks you uploaded I didn't use as these weren't the raw files (tifs actually). Do you have bias as well? Would be nice to have darks and bias in raw format as that's always the best place to start out with.

Anyway, I went on with first extracting the H-alpha and stacking and processing those first. After which I re-did it, but this time selecting to extract O-III, APP then asks to re-normalize the subs and you select to do so. In both cases I opted to stack 50 out of the 63 frames and I stacked with the quality weights, 1st degree LNC 3 iterations, MBB 5% and MAD sigma clipping 2,5 with 2 iterations. The resulting stacks were combined in the "combine RGB" tool, by adding H-alpha to H-alpha, O-III to O-III and then again O-III to green. The combined image I calibrated the background of, did star color calibration, cropped it a bit, saturated and sharpened and this is what I got out of that in the end;

There are some artefacts which look like hot-pixels that weren't removed. Probably as I didn't calibrate the subs (and they seem to be the same lines you have in Andromeda, leading me to think the calibration data might have an issue). But the main issue is solved it seems.

@vincent-mod

Thanks!

SO after you finished integration with first Ha set, did you remove the calibrated frames and then redo, because app was asking me whether I want to overwrite yes/no

It will be great to have a small PDF with screenshots for commonly used workflows and settings. Will really help us users.

No it doesn't as when you change to "extract OIII" in the 0-tab, the resulting files will have "HAOIII-HA" or "HAOIII-OIII" in the filename. I'm now suspecting you didn't change the setting in the 0-tab..

A PDF is being made for the entire program by Mabula. When it's released I don't know myself.

@vincent-mod

I did so. But still my Ha integ came out red instead of black and white. 

I will try this again and document the specs in a single PDF and share.

I DID IT!!!

Here is the final result. Attaching PDF with steps

I downloaded Pixinsight Trial too and gave it a whirl (Half Hearted I must say). Did not  get to actual processing.

Don't feel the need. Am buying APP. Just wondering if any deals or coupon codes come now and then so I can snag a discount 😀

That said @vincent-mod and other developers/mods, will it help if I keep sharing Duo Filter data with you for improving the tool. I am not doing it for freebies, just want  the program to get better :).

Would love to have a one step dummy mode where you load lights darks etc., tell tool what type of image, and then press one button to go through all the steps. After all, 90% of all processing is anyways done with default options. Maybe give some more options to tweak before clicking do everything button!

Hey @tsk1979,

Thanks for choosing APP! As far as I know there aren't deals or coupon codes, sorry.

A 1 button does all is kind of the idea behind APP. Compared to PI it already has that for many steps and Mabula is always going to try to make that as easy as possible whilst maintaining top quality. Unfortunately it's not really possible to have literally everything behind a button as data varies an enormous amount between setups and require some tweaking. It could be an idea to have like case scenarios with preset settings, and ofcourse saving the settings would be nice and Mabula has that on his to-do list.

Hi. Jumping on the topic - tried that very same workflow, and have an issue with it. Did all the steps until integration of the Halpha layer. But when I switch back to 0. LOAD, and change the algorithm to H alpha - OIII extract OIII, APP does not ask me to renormalize.

If I normalize, then integrate at that point, I have in fact the exact same signal as in my H alpha integration...

There must be something I'm doing wrong, but nothing I try seems to do the trick. Any idea?

Thanks. Adrien

You don't have to load the subs again, just go back to the 0-tab and switch to the other algorithm.

Well, I'm at a complete loss here. Any help is welcomed right now. Here's what I have/do, following the steps Mabula mentioned:

1: shot my first STC Duo Narrowband HA/OIII images with a ASI294mc-Pro. 10 images of M27
2: [tab 1] load lights
3: [tab 0] select Ha-OIII extract Ha (leave all default, i.e. pattern = supported)
4: [tab 2] calibrate using defaults
5: [tab 3] analyse stars
6: [tab 4] register using defaults
7: [tab 5] normalise using defaults
8: [tab 6] integrate. Everything default except outlier rejection --> LN Sigma clip (I used dithering when taking the images. Learned that this setting works best for that)
9: back to [tab 0] select Ha-OIII extract OIII
10: straight to [tab 5] normalise using defaults
11: [tab 6] integrate, same settings as at step 8

Now I have 2 files in my working directory. One for Ha and one for OIII. So far, so good I guess. Both images look exactly the same though! I think something already went wrong here but ok...

12: Go to [tab 9] and select combine RGB
13: load both files, select the first as Ha and the second as OIII filter
14: Slide the red slider for Ha to 100% and green and blue for OIII to 100%

The result: A black and white image, no color. Nothing like what I see here on the topic. 
I also followed the steps in the pdf shared here, same result. 

What am I doing wrong here? 

Hi Arjan,

Problems will always arise when doing new things, no worries we will get to the bottom of it. 🙂 For the quickest answer, you can upload like 10 lights, maybe calibration masters as well, to our APP server. Use login and password: appuser

Please create a folder with your name on it first. I'll then have a look at the process and see if I can create a proper result and tell you about it.

@vincent-mod

I will. thanks. Surely I'm doing something wrong, but that's the best way to learn, right 😉
It's going to be 10 lights only, no darks etc. 'cause this was a first test with the duo narrowband filter. I've got 10 color-lights, too but those can wait. 

You can add those too, no problem, I can use them to immediately go for the full picture.

@vincent-mod

okidoki. Done. Don't look at quality and stuff. This was 'first light' for the ASI294MC-pro with both filters. So the files starting with 'NB...' are with the STC duo narrowband and the files starting with 'L...' are the lights with a Baader IR/UV cut filter. 

Tnx again! Hope the workflow will generate some results and I love to understand what's happening.

No worries, I always only look at the problem, I'm not here to judge the actual quality. 😉 I'll download them and have an answer by tomorrow.

So the problem is because the debayering is not taking place. APP doesn't recognize the camera as having a Bayer filter on the sensor. So to solve this you select "Force CFA" in tab-0 and then go through the steps again. Then you will get the following result (I didn't try to make it nicer or anything, but it solved the issue):

I'll not say anything about them not being in focus. 😉 But you've got some nice signal here!

@vincent-mod

Holy moly! Thanks, I'll give it a go myself. I played with the [tab 0] and selected RGGB, so at least I was on the right track. sort of 🙂
At least the Ha and OIII data is there, so that crazy expensive filter actually does something 😉

As for focus, I found out I can do a 4x4 binning in APT, exposure to 0.1 and use my bahtinov mask. It's just so different from DSLR but I'll get there...

Another lesson learned. I'll let you know if I can get the same result.

got it! Same workflow as described, but changing to Force CFA did the trick.
Now to figure out how to combine the Ha/OIII with the color-lights... Luckily I'm going on a holiday in 3 days, plenty of time to play around with it!

seems I had the same issue in fact / Force CFA not ticked... Thanks a lot for the support.

Adrien

Hi All,

I have an Optolong L-Enhance which is also a dual band for OSC similar to the STC.

I have followed the workflow and I have extracted the Ha and OIII. In "Combine RGB", I have tried allocating Haextract to Ha and OIIIextract to OIII, also tried allocating Ha to Red and OIII to green and blue. I am confused as my trials bring at best a pale result, at worst all green, all red or similar terrible results.

Could someone please explain how you play with the sliders? What is the workflow to produce a good image when combining those narrowband extracts ? 

Thanks!

Philippe

@phfenain

Hi Philippe,

I am by no means an expert on this, but I did find that this next workflow works for me:
You seem to be able to extract Ha and OIII and combine them in RGB-combine. So far, so good. What I do next is select the Ha-image for H-Alpha and the OIII for oxygen. Then, slide the R(ed) to 100% for Ha and G(reen) and (B)lue both to 100% for OIII. Click Calculate. I started playing with the sliders afterwards, but if you keep a green haze, try PixInsight to remove that (I am currently on a trial-license and found that it works great, it's just a bit difficult).

Hope this helps a bit. See the screenshot as an example, it shows the sliders for both files. Note that if I leave Green at 0%, the image turns all pinkish or whatever. No details anymore. So perhaps you can try this.

Arjan

@arjan-kievits

Thanks! That helped me a lot. Still need to investigate and understand the effect of the sliders.

Posted by: @mabula-admin

Hi @daniel-tackley & @dav78,

Method 2) is preferred because you will be able to get much more and different colors in the bi-color composite.

It is quite stratightforward to accomplish what you want to do:

1) you will need the mono (for lum if you need/want is as well) and the extract algorithms

2) start with the Ha-OIII extract Ha (usually more signal so easier registration), process the data up until the Ha integration.

3) don't completely restart now

4) set the other debayer algortihm, for instance the Ha-OIII extract OIII,

5) now go to 5) normalize to normalize the OIII data. (no need to perform registration again, and this ensures the Ha and OIII integrations will be registered to each other)

6) Integrate the OIII data ,

The resulting integrations will be registered to each other thus are ready to load into the RGB Combine tool 😉

Let me know if this is clear and if you can follow this flow 😉

Kind regards,

Mabula

 

 

Hi,

I have processed NGC7000/IC5070 with data aquired with an Optolong L-enhance filter (similar to STC DUO) and followed the method Mabula describes earlier in this thread. Loaded Ha to the Ha channel (x1,000 bg 0,8 - 100%R 50%G 0%B) in Combine RGB, OIII to the OIII channel (x5,000 - 0%R 50%G 100%B). I am totally blown away from the result! The only thing which could be improved is the transition between the blue areas and the orange/red areas, especially in "the wall" region. If you zoom to 100% the color channels have a prominent dotted structure. Maybe there is some solution to reduce this effect further?

Cheers,

Chris

It's something that could use a bit of optimization maybe (from APP). If you could share the data, Mabula can use it to improve upon it. Did you do HSL selective color on it?

@vincent-mod

Hi Vincent,

thanks! I did not use HSL selective so far - did all my postprocessing in Photoshop. I will try to find a way to provide the data...

Chris

You can upload it to our server if you want, login/password is appuser
Just create a folder with your name and put it in there. Would be awesome. 🙂

Posted by: @mabula-admin

1) you will need the mono (for lum if you need/want is as well) and the extract algorithms

2) start with the Ha-OIII extract Ha (usually more signal so easier registration), process the data up until the Ha integration.

3) don't completely restart now

4) set the other debayer algortihm, for instance the Ha-OIII extract OIII,

5) now go to 5) normalize to normalize the OIII data. (no need to perform registration again, and this ensures the Ha and OIII integrations will be registered to each other)

6) Integrate the OIII data ,

The resulting integrations will be registered to each other thus are ready to load into the RGB Combine tool 😉

This has delivered very good results, and is in fact simple to learn.

Now the problem I have is how to add to this mix an unfiltered (full RGB) set from a 2nd camera (different geometry / pixels). I have succeeded doing that using multi-session processing (session 1: Ha-Oiii, session 2: RGB), but have failed to do that using the workflow discussed here ( session 1 / channel 1: extract-Ha, session 1 / channel 2: extract-Oiii, session 2 / channel 1: RGB). In the latter case no matter how I try going through the all process (at the combine RGB I keep getting the " The images aren't compatible! please make sure that the data has same dimensions and bit depth " error).

Any hints on how to do multi-session together with separate Ha and Oiii extracts from one of the sessions?

--Simon

@chris_16

Would love to know how you produced the hubble palette in Photoshop with the ha and oiii files from APP using the Optolong L-enhance filter - I have the same one with a ZWO 294MC pro.