My Pixinsight Processing Steps:
Combine the Red, Green and Blue images into a RGB color master using Channel Combination.
Use Graxpert to remove any gradients from the masters.
Color calibrate the RGB.
Improve star shapes in the RGB image with Blur Xterminator.
Sharpen each master using Blur Xterminator. Reduce star sizes if needed.
Clone the RGB image and use the clone for a stars-only image.
Stretch the images.
Apply noise reduction.
Remove stars while only saving the stars from the RGB copy.
Combined RGB image after Channel Combination but before Graxpert or color calibration. The red master was the brightest, so the overall color cast of the image is red. Graxpert and color calibration will fix that.
Removing gradients using Graxpert:
This image was from a very dark sky site, so there aren't many gradients to worry about. I set Graxpert to the highest smoothing level (1.0) and Subtraction. For more complicated gradients I would use a lower smoothing level.
I usually use Subtraction on color images and Division on monochrome, but I'm not sure it matters that much.
The Graxpert_background image shows what was removed.
RGB Color-Calibration
First, I'll run the ImageSolver script on our RGB image. This determines exactly what portion of the sky the image covers. This is needed for the next step.
Next, I use SPCC to color-calibrate the data. It will find known stars and compare their color in the image to their actual color, adjusting the entire image accordingly. When calibration is complete PI produces a chart of the results. Since the star plots are close to the slope of the graphs, with very few outliers, we know this is a successful calibration.
Blur Xterminator
First- RGB stars. Middle- RGB stars with BXT set to corrections only. Last- RGB stars after BXT with sharpen stars set to .30. I didn't shrink the stars much. If I do more, they'll look too small in the final image.
Stretching
Next, we have to permanently stretch (brighten) the images. To do this, I take Pixinsight's temporary stretch and make it permanent using the Histogram Transformation process. I tweak the screen stretch by eye to ensure the backgrounds are similar and nothing is too bright. Areas that are too bright will look washed out when images are combined in Photoshop.
To create small, saturated stars, I stretch the RGB copy using the Star Stretch script.
First- RGB image after stretching with HT. Second- RGB_stars image after stretching using Star Stretch.
Now we have 4 stretched images: Lum, RGB, RGB stars, and Halpha:
Removing noise in the Halpha image using Noise Xterminator.
Final steps in Pixinsight
Apply Noise Xterminator to each image.
Remove stars from Lum, RGB, and Halpha images using Star Xterminator. Don't create separate star images; we won't use them.
Remove the stars from the RGB_stars image using Star Xterminator in unscreen mode. Throw away the starless image.
Extract the Red channel from the RGB image using the imaginatively named Channel Extraction process. We'll use this image when we add the Halpha in Photoshop.
Save as 16-bit TIFFs for Photoshop:
We have 4 starless master files and 1 stars master to save:
Lum
RGB
RGB-red channel
Stars
Halpha
I save each of these as a 16-bit TIFF file so they can be opened in Photoshop. Now we're done with Pixinsight.
Next is Photoshop Processing, Part 1