Methods for channel blending usually involve destructive methods such as apply Image or Channel Mixer. But there is another method that is nondestructive, and after running a handful of tests I'm scratching my head and wondering why this method does not seem to enjoy wide support.
Anyway, for those unfamiliar with this, here's how it works.
Let's say you have a dirty blue channel, and you think blending in part of the green channel might fix it.
1. Go into Channel palette. Select and copy the green channel. Go to the layer palette. Paste. Now you have the green channel's "grayscale" sitting on top of the Background layer of your image. Activate the Blending Options menu for the "grayscale" layer. Under Advanced Blending|Channels, uncheck the boxes for "R" and "G." Click okay.
2. What you see on your monitor is the image with the grayscale of the green channel substituting for the grayscale of the blue channel. And as you lower the opacity you blend in more of the blue channel and less of the green channel, until you get it to where you want it.
Now, maybe you want to see how the L channel works here instead. You take it from a second copy of your picture, put it in there, and set its advanced channel blending to "B" only. Now you click back and forth between the L layer and the G layer to see which one works better. And because it's too light, you add an adjustment layer, set its advanced blending to "B" only, and reduce lightness. All in a nondestructive way. And of course, you can also use blending modes and layer masks with this.
(Curator's note: Does anyone recognise this? Phosphor didn't write it, but he doesn't remember who deserves the credit. If you know who wrote it, please place a note in the Lounge)
From Andrew Bokelman 18 June 2004
In the Photoshop Lounge Resource Repository is a copy of the beginning of a message I posted about Nondestructive Channel Blending, and at the bottom it asks who originally posted the message. I posted it on 1/13/03.
But I only developed the scenario to explain it. I didn't think of the idea. And in the rest of my original method I said: "BTW, I didn't discover this. It was just suggested to me by Witkacy over in the featurerequest group, and attributed to Dan Margulis."
The thing is, I tried to find where Dan Margulies wrote about this, and found no mention. When I mentioned this to Witkacy, he said that maybe he read it some place else. So we are not sure who first wrote about this, but I would not be surprised to find it was Dan Margulies.