Hi Terry,
In images where there is an important contribution to the nebula from other wavelengths, I always put the H-alpha data in the lightness of the color image. This is done to preserve all the color information.
The main problem we face by putting the H-alpha image as lightness of the color image is that the nebulae are much brighter in the ew lightness than in the original one. This causes a loss of color saturation (typically the red nebulae become pink) and even a complete saturation of the the brightner red nebulae. That's why we should darken the H-alpha image before putting it as the new lightness. The operation to darken the image strongly depends on each particular case; sometimes HDRMT works well, and sometimes a simple midtones or curves adjustment does the job. This process will minimize the loss of color saturation, which will be easily recoverable to the original state of the original color image with a simple curve adjustment in the S channel.
The other problem we face with this technique is the dramatic change of brightness of the stars. How you protect the stars and the reflection nebulae will change from image to image. In my workshop in Munich last year I showed three different examples on how to solve this problem, each one with different star masks. One of the examples was this one:
https://pixinsight.com/workshops/cfa-2014/munich-02.htmlIn this particular case I protected the stars in two ways:
- First I mixed the L component (or a synthetic L from the color image if you don't have L-filtered images) with the H-alpha image. A small amount of the L image will help to recover the dimmer stars, which almost dissapeared in the H-alpha image. For this example I used 20% of L and 80% of H-alpha; this proportion is a good balance between noise transfer from L to H-alpha (considering that usually the H-alpha image is much cleaner) and the effect on the dimmer stars.
- I created a mask to protect the stars. In this case I created the mask from the B channel of the color image because there were also some tiny reflection nebulae. The H-alpha introduction into the lightness of the color image should be regulated by this mask.
For your image, even if I haven't seen it, I would say that you'll need two masks. One that protects the reflection nebulae and another that protects the stars. If your star and reflection nebulae in both masks are bright (so you invert the mask when appplying the lightness introduction), you can mix both masks in PixelMaxh with this equation: Max( mask_stars, mask_nebulae ).
IMPORTANT: All these processes are applied to non-linear images!My plan is to publish by the end of this year an article with a compilation of the best examples of my workshops during the last year. The example above will be one of them.
Hope this helps.
Best regards,
Vicent.
