I'd like to offer up a processing tip for achieving better balance with narrowband channels for SHO or custom blend imaging. I use pixinsight, so the steps I describe are PI based, but the concept may be applicable to other processing suites.
------------------
One of the challenges with narrowband SHO imaging is that for most targets the signals for each of these channels is vastly different. HA is typically the dominant signal followed by Oiii and Sii. When combining these images into a color image you need to make sure the channels are fairly well balanced bringing out the fainter Oiii and Sii signals becomes a serious challenge. In the past I've used Histogram Transformation or Linear Fit to balance out the channels, but I'm never very happy with the result.
For my LRGB imaging I've been using Photometric Color Calibration (PCC) tool for a while now and really love the even color balance that it achieves with DSO and stars alike. Generally only a slight green color cast remains, which is pretty easy to neutralize.
To make it more challenging for me I tend to do custom pixelmath blends for my narrowband images, instead of a straight SHO where Sii is mapped to Red, Ha is mapped to Green and Oiii is mapped to Blue. Until recently I've simply linear fit the Ha and Sii data to Oiii, otherwise the Oiii is just drowned out. Even matching Ha and Sii to Oiii still results in a fairly dilute Oiii channel, so I find it's quite a challenge to get the blues to shine.
Here are all three channels with a screen stretch. HA is by far the dominant channel here.
Simply mapping the channels SHO to RGB results in a terrible red color cast.
A few nights ago I collected 4.5 hours of data split evenly among SHO filters. It's a fairly short integration despite being from an Epsilon, but to make matters worse I collected it during the full moon. In my skies... you dont have a choice. If it's clear... you image! So the Oiii was even more washed out compared to the dominant Ha than normal. I didnt like the Linear Fit method I've been doing, so I decided to expirament with PCC. In PCC you can assign bandwidth and bandpass to specific color channels, so after combining SHO as an RGB I ran PCC with these parameters. I was really excited to see a wonderfully balanced SHO image!
Running PCC on the above image resulted in a very balanced SHO image:
If you are like me and like to do custom blends with pixelmath, here is where the tip comes in. Simply split those channels back up and now you have your separate SHO channels, but they are properly balanced. Now you dont have the undesirable effects of Linear Fit or curves, etc... You just have perfectly balanced SHO channels to work with.
In my case I like to do a custom Hubble Blend of:
R=.4*Ha + .6*Sii
G=.4*Oiii + .3*Ha + .3*Sii
B=Oiii
Notice how nicely the Oiii shows up. It's not just a whitewashed blue tint, it has a really nice color to start with:
Upon combining I had a really wonderful base result to work with that was free of the negative effects the other methods I have tried seem to produce.
So give PCC a try for your narrowband imaging, it's not just a broadband balancing tool!
The final result was produced without any crazy gymnastics. I removed the stars and processed the nebula and stars separately. Recombined. Did some color saturation, contrast and noise reduction and thats it!
Full resolution version can be found here:
Please feel free to add your own processing tips. I would love to hear how you manage your channels so Oiii isnt lost in the sea of Ha.
------------------
One of the challenges with narrowband SHO imaging is that for most targets the signals for each of these channels is vastly different. HA is typically the dominant signal followed by Oiii and Sii. When combining these images into a color image you need to make sure the channels are fairly well balanced bringing out the fainter Oiii and Sii signals becomes a serious challenge. In the past I've used Histogram Transformation or Linear Fit to balance out the channels, but I'm never very happy with the result.
For my LRGB imaging I've been using Photometric Color Calibration (PCC) tool for a while now and really love the even color balance that it achieves with DSO and stars alike. Generally only a slight green color cast remains, which is pretty easy to neutralize.
To make it more challenging for me I tend to do custom pixelmath blends for my narrowband images, instead of a straight SHO where Sii is mapped to Red, Ha is mapped to Green and Oiii is mapped to Blue. Until recently I've simply linear fit the Ha and Sii data to Oiii, otherwise the Oiii is just drowned out. Even matching Ha and Sii to Oiii still results in a fairly dilute Oiii channel, so I find it's quite a challenge to get the blues to shine.
Here are all three channels with a screen stretch. HA is by far the dominant channel here.
Simply mapping the channels SHO to RGB results in a terrible red color cast.
A few nights ago I collected 4.5 hours of data split evenly among SHO filters. It's a fairly short integration despite being from an Epsilon, but to make matters worse I collected it during the full moon. In my skies... you dont have a choice. If it's clear... you image! So the Oiii was even more washed out compared to the dominant Ha than normal. I didnt like the Linear Fit method I've been doing, so I decided to expirament with PCC. In PCC you can assign bandwidth and bandpass to specific color channels, so after combining SHO as an RGB I ran PCC with these parameters. I was really excited to see a wonderfully balanced SHO image!
Running PCC on the above image resulted in a very balanced SHO image:
If you are like me and like to do custom blends with pixelmath, here is where the tip comes in. Simply split those channels back up and now you have your separate SHO channels, but they are properly balanced. Now you dont have the undesirable effects of Linear Fit or curves, etc... You just have perfectly balanced SHO channels to work with.
In my case I like to do a custom Hubble Blend of:
R=.4*Ha + .6*Sii
G=.4*Oiii + .3*Ha + .3*Sii
B=Oiii
Notice how nicely the Oiii shows up. It's not just a whitewashed blue tint, it has a really nice color to start with:
Upon combining I had a really wonderful base result to work with that was free of the negative effects the other methods I have tried seem to produce.
So give PCC a try for your narrowband imaging, it's not just a broadband balancing tool!
The final result was produced without any crazy gymnastics. I removed the stars and processed the nebula and stars separately. Recombined. Did some color saturation, contrast and noise reduction and thats it!
Full resolution version can be found here:
Please feel free to add your own processing tips. I would love to hear how you manage your channels so Oiii isnt lost in the sea of Ha.
