Stuck on OSC + Dual Band Processing

There are several ways to go about this but what you are essentially trying to do is a NBLRGB combo. Heres a high level overview on one way to do it without pixel math. There are some great scripts to help you accompish this with OSC. I would grab these if you dont already have them - Pixinsight toolbox from https://www.ideviceapps.de/PixInsight/Utilities/. Cosmic photons tools for image blend script https://www.cosmicphotons.com/pi-scripts/imageblend/ The continuum subtraction script from seti astro pixinsight scripts https://raw.githubusercontent.com/setiastro/pixinsight-updates/main/ and the pixinsight script DBXtract https://dbxtract.astrocitas.com/

If you have never added scripts, you should be able to find a quick youtube tutorial on how to manage your repository and add the URLs.

I would use your broadband image as your base image, do your normal linear things. Extract the RGB channels and set aside. Then stretch your broadband image - set aside.

Use the DBXtract script on your duo narrowband linear stack this will get you a fairly isolated Ha and OIII channel to process like a mono channel. Ideally you would then use the continuum subtraction script on your NB data (will utilize those RGB images you had set aside) - this will spit out non linear Ha and OIII data.

Then explore the Combine Ha with RGB or combine NB with RGB scripts in the toolbox to blend the data. Alternatively, you could use the imageblend script to blend the two images combined with masks - something along the lines of screening data with a color mask applied.

Use the selective color correction script and or curves to dial in your final color adjustments.

Now that you know the names of the scripts, you should be able to find the assortment of youtube tutorials to get you started.

Have fun!

What Scott described works very good for me

Tony Gondola · Apr 13, 2026 at 03:57 PM

PI is wonderful but this might be a place where a layer based photo editor might be a better tool. Something like Affinity might be well worth learning as things like this are very simple to do.

Agreed, masking and layer based editing has a huge advantage over PI in this stage of processing.

I like the idea on the blended layering approach, do you guys know of a layer based software solution like Photoshop that allows 32-bit FP data? It would be great to be able to work in 32-bit FP space like .xisf files rather than truncate to 16 bit tiff.

Lucas Jacobson · Apr 15, 2026, 08:33 PM

I like the idea on the blended layering approach, do you guys know of a layer based software solution like Photoshop that allows 32-bit FP data? It would be great to be able to work in 32-bit FP space like .xisf files rather than truncate to 16 bit tiff.

I dont personally, have never really strayed from PS. But, I often leave my luminance or extract luminance in PI and then only work on chrominance if I need layered editing in PS - bring that back to PI and layer it as color over my preserved 32b luminance data in image blend if that makes sense.

I watched seti astro present SASPro in the most recent TIAC youtube presenation - wow! It has a lot more tools than I realized. He mentioned non destructive layered editing but didnt have time to demo it - I suspect it would be the go to to work in 32-bit if needed.

64 fixed point would be great too. Just something to keep from stacking round off errors in the processing. I try to wait to dither to 16-bit until the last moment it makes sense. Most of the time it’s when I do the final save as .tif/.jpeg.