|
The trick is to not attempt to match your RGB data in one hit. Would suggest blending Ha data into the red filtered data as previously mentioned, but only a small amount to start with. Then layer the Ha data as a luminance blend mode over the [Ha+R]GB blend at 20% to 50%. Flatten this. You've now created a 'super' RGB that will better accommodate the contrast of the Ha data. Tweak the saturation of the 'super' RGB. Given it a fair boost, then layer the Ha data again but at a higher blend. You can keep doing this process until you feel the RGB data has matched the Ha highlights. This how you come up with LLRGB composites, its just re-layering the luminance to match the increasing strength of the RGB.
Remember - RGB is an additive colour space. 255:255:255 is completely white and as such you can't get any more colour into to it - no matter how hard you try! Shades of grey on the other hand will accept colour. Based on this you may find that it is near impossible to blend 100% of the Ha data if its too 'contrasty'. If this is the case, reprocess the Ha data and don't stretch it too far. Grey highlights are good, white are bad. Of course once you've got the colour you're after, you can then flatten and give a final tweak to the curves to make the highlights pop a little - but do it when the image is flattened, not on the luminance layer alone otherwise you'll knock your colour out which you've worked so hard to maintain.
Cheers
|