the thing I am just wondering about is, does full well matter if I combine luminance data (with burned out stars) with rgb data that contains well exposed stars?
Andi,
Others have answered correctly that FWC matters with luminance as much or more than RGB. But to your question about combining with well exposed stars I think there are two answers depending on what you mean:
If you mean that when you create the LRGB combo the well exposed color data will "fill in" the overexposed Luminance data, no it won't. The fully saturated areas of the Luminance data will still be white because the brightness level of all three RGB channels will be 100% or in other words identical. It may be best to think of the Luminance values inherent in the RGB data being replaced by the values from the Luminance master. Any area where the brightness of the luminance exceeds maybe 80% will not hold the color and will be white.
If you mean combining the Luminance values of the RGB data with the Luminance to create a super-luminance, the answer is maybe that could work but it's still not a good idea. Imagine a star that is blown-out 100% in the luminance and in the RGB data was at 50%. Creating a traditional SuperLuminance might combine the data with a weighting of something like .25 for the RGB data and .75 for the Luminance. In this case 100*.75 + 50*.25 = 87.5% of saturation - still over the 80% threshold I mentioned above. But worse, the transition between the saturated and not saturated parts of the star in the luminance data would still be visible, so you would have a hard ring in the stars. You might be able to combine the RGB Luminance data and the Luminance using a star mask that smoothly blended the star cores of the RGB stars into the Luminance data. This would exceedingly difficult to do well.
By far, the best idea is to avoid saturating your stars in the Luminance data.
Kevin
