I am assuming that you are using a OSC camera, not a monochrome camera. The filter basically has pass bands around the Hb, OIII, Ha and SII emission lines. The Hb and OIII lines lie around the edges of the sensitivity ranges of the blue and green pixels. The Ha and SII lines lie in the sensitivity range of the red pixels. My advice would be to simply treat the images as RGB images and process them like that. This is what we do with the Optolong L-eNhance and L-eXtreme filters and that works very well for us.
To my best of knowledge it is impossible to split Hb from OIII and Ha from SII.
There are extraction algorithms for Ha, Hb, OIII and SII, so you can select each one in tab 0 and then load in the data and process that signal. Clear the data, change the algorithm and process again. It may very well be that those that have overlap are pretty much the same, I never actually tested a quad-band so I will find some data to check this. You can compare the above approach with only using "extract Ha" and "extract OIII" to see if that is basically the same. And maybe try the Ha-OIII color algorithm as well to have a quick way to produce a RGB image from it.
There are extraction algorithms for Ha, Hb, OIII and SII, so you can select each one in tab 0 and then load in the data and process that signal. Clear the data, change the algorithm and process again. It may very well be that those that have overlap are pretty much the same, I never actually tested a quad-band so I will find some data to check this. You can compare the above approach with only using "extract Ha" and "extract OIII" to see if that is basically the same. And maybe try the Ha-OIII color algorithm as well to have a quick way to produce a RGB image from it.
That is what I did. I've used that algorithm to extract the following:
-Halpha
-Hbeta
-OIII
-SII
My question is what should I use to combine them (after processing) together?
If you have registered (press that first) and normalized these frames with each other.. save the normalized frames (all the way down in tab 5). Then you can combine them in Combine RGB in tab 9. In there you can load the data and mix the channels in any way you want.
Hi Jerry, how did that work out for you using the Triad Ultra in APP? I'm currently thinking about buying it as well. Right now I have the L-enhance and I'm using the Ha-OIII extract Ha and OIII to create HOO.
Hi Jerry, how did that work out for you using the Triad Ultra in APP? I'm currently thinking about buying it as well. Right now I have the L-enhance and I'm using the Ha-OIII extract Ha and OIII to create HOO.
I am not sure if I am processing it correctly. I was using Optolong L-Enhc and so far I didn't see a big difference that would justify the $1000.00 filter. One thing I want to mention - I am new to this hobby and I am pretty sure I am screwing it up somewhere. Following is the process-workflow I was using:
I was using AstroPixelProcessor to get OIII, Ha, Hb, and SII. Then I took it into PixInsight to post-process it further.
1. In Pixel match I combine all temporally to see what I need to crop out
2. I use Dynamic Crop to create a process - I use it on all to get the same outline
3. Automatic Background Extractor on all -(Correction - Division)
4. In Histogram Transformation I compare Luminocity of all to chose the brightest one as the reference for Linear Fit
5. I use Linear Fit to make them all the same
6. In PixelMath I combine them together to create RGB image - Formula as follows - Red-SII, Green-Ha, Blue-OIII and A- Hb
7. Time for Background Neutralization - I Create a Preview Window without any stars as possible
8. Color Calibration -
9. SCNR to remove Green
10. I use channel extraction (CIE L*a*b*) to extract L (luminance) - To create a luminance mask
11. Using Screen Transfer and Histogram Transformation I make mask non-linear
12. I apply a mask to the image (inverted)
13. To remove some noise I use MultiScale Linear Transformation with 4 layers - RGB.
Layer 1 Threshold 4, Amount 0.50 Iteration 3
Layer 2 (2.000, 0.50, 2)
Layer 3 (1.000, 0.50, 2)
Layer 4 (0.500, 0.50, 1)
14. Remove Mask and I make image non-linear Using Screen Transfer and Histogram Transformation
15. I create another mask using channel extraction (CIE L*a*b*) to extract L (luminance)
16. Using HistogramTransformation I press AutoClip Shadows and AutoClip Highlights to give the mask a little stronger
17. Apply the inverted mask to the image
18. Another noise removal with MultiScale Linear Transformation with 4 layers RGB.
Layer 1 Threshold 3, Amount 0.50 Iteration 2
Layer 2 (2.000, 0.50, 2)
Layer 3 (1.000, 0.50, 1)
Layer 4 (0.500, 0.50, 1)
19. ACDNR - noise removal (the inverted mask is still on)
All the numbers at default value except
StdDev - 2.0
Amount - 0.75
Iteration - 3
20. Remove Mask
21. Using Range Selection I create another inverted mask to protect nebulas or galaxies but work on the background
22. Using MultiScale Linear Transformation with 8 layers and Target selected to Chrominance
Layers 5-8 with Detai Layer checked - set Bias to -1
23. With mask still on I would adjust the darks using Histogram Transformation
24. Remove mask (don't delete) and play with Curves Transformation to work on the overall image
25. Duplicate the image and use HDR Multiscale Transformation on the duplicated image with the following settings
Number of Layers -3 and Number of Iterations -2
26. Create Process - Star Mask with setting
Noise threshold - 0.20
Scale - 6
Large Scale - 0
Small Scale - 0
Compensation - 2
Midtones - 0.01
Everything else - default
27. Adjust star mask with Histogram Transformation to pop the stars
28. In PixelMath Create a new GRAYSCALE image (this is my new mask) with the following Expression- range_mask-star_mask
29. Apply a new mask to the image
30. With MultiScale Linear Transformation with 8 layers Chrominance.
Layer 5- Detail Layer Bias 0.15
Layer 6- Detail Layer Bias 0.15
Layer 7- Detail Layer Bias 0.20
Layer 8- Detail Layer Bias 0.20
31. HDR Multiscale Transformation on the duplicated image with the following settings
Number of Layers - 8 (or 10) and Number of Iterations -1, with a checkmark at TO LIGHTNESS and checkmark at LIGHTNESS MASK
32. Remove Mask
33. Now it is time to play with colors using Curves Transformation
Sorry I can't help with the Pixinsight workflow. I'm sticking with APP and Photoshop. I was just reading on CN that you can't really differentiate SII from Ha when using an OSC since both would get mapped to the red channel. Maybe only benefit is get much cleaner signals since we are looking at 4nm and 5nm band pass versus 18nm and 12nm for for OIII and Ha respectively. OK, well, it looks like I'll either go mono or go with something cheaper. Good luck.
