Hello everyone.
After 2 months of cloudy nights I finally caught a night when I can take my telescope outside. For my next target I decided to image M 33. I bought a UV/IR cut filter and have been using that for the first night of collecting data.
I also recently purchased the Askar Colormagic D1 and D2 filter set. If I image the same target with the same framing, but using the D1 filter (Ha + OIII), can I later add the D1 data to the Luminance data I gathered?
For context I own a ZWO ASI585MC Pro OSC camera.
For processing I use Pixinsight and Affinity Photo 2.
Thanks in advance to everyone who takes the time to respond.
Adding Dual-Narrowband data to an OSC image
Well Written Respectful
Hi Rene,
Of course you can do that.
Very quickly explained: in my case, I first split the HO filter data in Ha & Oiii data using an script called DBXtract, then you need to do the continuum subtraction of that images using RGB data (in this case the Red signal for the Ha and the Green signal for the Oiii), convert the result to non-linear data (as well as the RGB data) and finally I merge the three signals (RGB + Ha continuum-subtracted + Oiii continuum-subtracted) using an script called CombineRGBandNarrowband.
Hope this helps,
Juan Carlos
Helpful
Juan Carlos Bárcenas · Sep 19, 2025 at 08:10 AM
Hi Rene,
Of course you can do that.
Very quickly explained: in my case, I first split the HO filter data in Ha & Oiii data using an script called DBXtract, then you need to do the continuum subtraction of that images using RGB data (in this case the Red signal for the Ha and the Green signal for the Oiii), convert the result to non-linear data (as well as the RGB data) and finally I merge the three signals (RGB + Ha continuum-subtracted + Oiii continuum-subtracted) using an script called CombineRGBandNarrowband.
Hope this helps,
Juan Carlos
Thank you for the reply, Juan.
I will give that method a try. Do you do anything with the stars? Or should I just use the stars taken with the UV/IR filter?
Clear Skies.
Well Written Respectful
Straight-through addition (average) of the color calibrated RGB data and the color calibrated NB data using PixelMath. Continuum-subtraction, assuming that it means anything, doesn't work very well for OSC (and even more so the IMX585) because of cross-pollution of the Ha (R) signal into the G Band (to a smaller degree the B band too) as well as the reverse.
Rene Matjanec · Sep 19, 2025, 08:18 AM
Juan Carlos Bárcenas · Sep 19, 2025 at 08:10 AM
Hi Rene,
Of course you can do that.
Very quickly explained: in my case, I first split the HO filter data in Ha & Oiii data using an script called DBXtract, then you need to do the continuum subtraction of that images using RGB data (in this case the Red signal for the Ha and the Green signal for the Oiii), convert the result to non-linear data (as well as the RGB data) and finally I merge the three signals (RGB + Ha continuum-subtracted + Oiii continuum-subtracted) using an script called CombineRGBandNarrowband.
Hope this helps,
Juan Carlos
Thank you for the reply, Juan.
I will give that method a try. Do you do anything with the stars? Or should I just use the stars taken with the UV/IR filter?
Clear Skies.
I always use the stars taken with broadband filters so their natural colors are preserved.
Juan Carlos
andrea tasselli · Sep 19, 2025, 08:57 AM
Straight-through addition (average) of the color calibrated RGB data and the color calibrated NB data using PixelMath. Continuum-subtraction, assuming that it means anything, doesn't work very well for OSC (and even more so the IMX585) because of cross-pollution of the Ha (R) signal into the G Band (to a smaller degree the B band too) as well as the reverse.
The Ha/Oiii contamination has nothing to do with continuum subtraction though. You can still subtract the continuum signal just fine. You’re right, OSC data will be cross-polluted, but you can extract the narrowband with pixelmath or scripts. I think it might be technically impossible or very difficult to get a perfect extraction though you can get a really close approximation which is fine for “pretty picture” purposes.
Well Written Helpful Insightful Respectful
I’ve since switched to mono but here’s a workflow that should suit your needs :
In WBPP, make sure your lights from different filters are marked as such. This means you end up with 2 different but already aligned master light files : one broadband, one dual band.
For both, split your RGB channels. Then, from the split dual band data, synthetically reconstruct the Ha and OIII signal (or SII if you have it) using Pixel Math, and discard the original RGB channels of the dual band image. Keep in mind that this is an imperfect reconstruction and will never defeat mono cams in terms of data fidelity :
Ha = 0.8*R + 0.15*G + 0.05*B
OIII = 0.95*G + 0.45*B
S = 0.75*R + 0.15*G + 0.05*B
You now have 5 images : broadband RGB and Ha and OIII (or some other dual band combination depending on the filter). Process and stretch each channel individually, then combine Ha with R and OIII with G and B again using Pixel Math (or some other combination if you don’t just want straight HOO) :
R_new = ~((~R)*(~Ha))
G_new = ~((~G)*(~OIII))
B_new = ~((~B)*(~OIII))
And finally, use LRGB combination to create the final result. You can still stretch individual channels within your final result to tweak the balance between your colors if it’s not to your liking.
Helpful
Patrick El Hage · Sep 19, 2025 at 03:03 PM
I’ve since switched to mono but here’s a workflow that should suit your needs :
In WBPP, make sure your lights from different filters are marked as such. This means you end up with 2 different but already aligned master light files : one broadband, one dual band.
For both, split your RGB channels. Then, from the split dual band data, synthetically reconstruct the Ha and OIII signal (or SII if you have it) using Pixel Math, and discard the original RGB channels of the dual band image. Keep in mind that this is an imperfect reconstruction and will never defeat mono cams in terms of data fidelity :
Ha = 0.8*R + 0.15*G + 0.05*B
OIII = 0.95*G + 0.45*B
S = 0.75*R + 0.15*G + 0.05*B
You now have 5 images : broadband RGB and Ha and OIII (or some other dual band combination depending on the filter). Process and stretch each channel individually, then combine Ha with R and OIII with G and B again using Pixel Math (or some other combination if you don’t just want straight HOO) :
R_new = ~((~R)*(~Ha))
G_new = ~((~G)*(~OIII))
B_new = ~((~B)*(~OIII))
And finally, use LRGB combination to create the final result. You can still stretch individual channels within your final result to tweak the balance between your colors if it’s not to your liking.
Thank you for the detailed workflow.
I will definitely give this a try. I saw something similar on youtube while I was researching this further and they showed some good looking results (obviously not as good as mono).
Maybe I will end up using just the UV/IR filter data, but I want to give this a try.
Clear Skies.
Well Written Respectful