IC410. Erratum. Adding HA to LUM actually works just fine.
Posted: Sat Mar 13, 2021 12:24 am
I have to fess up. It turns out that I made a pigs ear of adding the HA signal to the RGB luminance in my last post and that is the reason it apparently hadn't worked all that well.
Actually..processed properly ...it looks to me to be the best way to make use of HA data to add nebula detail to RGB without distorting the colours. Anyway to set the record straight and in case folk do use PI (or equivalent steps in other software) and are interested in this technique.
So same data as before.
Skywatcher 200 PDS Newtonian, Baader flattener, F 5.0, CEM70 mount with PDS guiding via AS 120 MM camera and 9 x 50 finder. Pegasus Cube focuser
ASI 294 MM PRO camera, -10C, Optolong 7 nm HA filter
ASI 294 MC PRO camera, - 10C, using an Astronomik UHC filter
Sharpcap livestack captures filtered on FWHM and Brightness, Master darks (33-40 frames) and flats (100 frames) captured in Sharpcap.
124 x 70s (145 min) of RGB at gain 124 and 38 x 3 min (108 min) of HA at gain 124 captured by SC and preprocessed in PI
Same processing steps as before in PI but, crucially, carried out using the pre-stretched non-linear images
1) Process the linear RGB and HA images as normal (fix background (DBE), shrink star morphology, remove green (SCNR), Colour calibration. noise reduction etc.
2) Then stretch both about equivalently using Histogram transformation
3) Then extract the RGB luminance channel and use the Linear function to stretch it to be equivalent to the HA image (now call it say 'RGBLS')
4) Then use the Pixmath max function to create a new luminance image using the function max (HA, RGBLS)
5) Then add the new luminance channel to the RGB image using the LRGB process
Finally curves and anything else to tidy up further
The crucial difference was to apply the process 1) to 5) to non-linear rather than linear images. It also helped to tick the 'reduce chrominance noise' button in LRGB. It just goes to show how important it is to get the processing as well as the capture right!
TimH
Actually..processed properly ...it looks to me to be the best way to make use of HA data to add nebula detail to RGB without distorting the colours. Anyway to set the record straight and in case folk do use PI (or equivalent steps in other software) and are interested in this technique.
So same data as before.
Skywatcher 200 PDS Newtonian, Baader flattener, F 5.0, CEM70 mount with PDS guiding via AS 120 MM camera and 9 x 50 finder. Pegasus Cube focuser
ASI 294 MM PRO camera, -10C, Optolong 7 nm HA filter
ASI 294 MC PRO camera, - 10C, using an Astronomik UHC filter
Sharpcap livestack captures filtered on FWHM and Brightness, Master darks (33-40 frames) and flats (100 frames) captured in Sharpcap.
124 x 70s (145 min) of RGB at gain 124 and 38 x 3 min (108 min) of HA at gain 124 captured by SC and preprocessed in PI
Same processing steps as before in PI but, crucially, carried out using the pre-stretched non-linear images
1) Process the linear RGB and HA images as normal (fix background (DBE), shrink star morphology, remove green (SCNR), Colour calibration. noise reduction etc.
2) Then stretch both about equivalently using Histogram transformation
3) Then extract the RGB luminance channel and use the Linear function to stretch it to be equivalent to the HA image (now call it say 'RGBLS')
4) Then use the Pixmath max function to create a new luminance image using the function max (HA, RGBLS)
5) Then add the new luminance channel to the RGB image using the LRGB process
Finally curves and anything else to tidy up further
The crucial difference was to apply the process 1) to 5) to non-linear rather than linear images. It also helped to tick the 'reduce chrominance noise' button in LRGB. It just goes to show how important it is to get the processing as well as the capture right!
TimH