Combining HA with RGB:
I was asked what process I use to avoid the dreaded salmon pink shift that often occurs. The key in the process below is that the luminosity I substitute in is not much lighter (here it is actually darker) than what I am replacing. Here is a break down of some of the steps in processing my Jellyfish Nebula.
First, here is the combined RGB stack. It is still linear at this point. I have already done gradient removal (DBE) and some saturation and color balancing. What is being shown has a screen transfer function (STF) applied to stretch it so the results are visible.
I was asked what process I use to avoid the dreaded salmon pink shift that often occurs. The key in the process below is that the luminosity I substitute in is not much lighter (here it is actually darker) than what I am replacing. Here is a break down of some of the steps in processing my Jellyfish Nebula.
First, here is the combined RGB stack. It is still linear at this point. I have already done gradient removal (DBE) and some saturation and color balancing. What is being shown has a screen transfer function (STF) applied to stretch it so the results are visible.
Here is the HA data. Again, I have already done some gradient removal. The image is still linear and is shown with a STF.
Now I do a RGB and HA combine with the linear data in PixInsight using the NBRGBCombination script that is available.
Data remains linear and is shown with a STF.
Data remains linear and is shown with a STF.
Now I redo the color calibration. Data is still linear with STF.
I now brought the linear RGBHA data in StarTools, stretched it, and denoised it, along with some other stuff. Data is no longer linear. I could have stayed in PixInsight for this step.
Next I brought the linear HA data into StarTools. Again, it was stretched; a deconvolution, wave sharpening, and denoising were also done. Data is no longer linear. Again, I could have stayed in PixInsight.
The linear RGBHA data was then brought into Photoshop and converted to LAB color. This allows working with the lightness channel.
The lightness layer of the LAB color converted RGBHA data is then selected.
The HA data that was previously prepared as the luminosity layer is now pasted in place of the lightness layer.
The image is then converted back to RGB. Note the lack of horrible color shifts to salmon pink.
Various other tweaks in terms of levels and such are done resulting in the final image.
The reason the above worked and did not lead to bad color shifts is that the lightness channel was at around the same level or brighter than the luminosity data that I copied in.
The reason the above worked and did not lead to bad color shifts is that the lightness channel was at around the same level or brighter than the luminosity data that I copied in.