9) The way it acts I am guessing that when over 50% of the color of the channel would be affected then it masks. To soften the transistion hold down the alt/option key when grabbing the trangle thingy - you will split the triange and will be able to select a gradual transistion from where the masking of the channel starts and where it ends. Here is the layers flipped with the blue channel selected, blue (right hand) triangles spit...
In an effort to understand how to think 'blend-if' so that I can recognize when and how to use it I have chosen to first play with images that use it badly. I know this sounds wierd, but the easy images are not teaching me to think 'blend-if' - so I tried difficult photos. I will state what the exercise taught me as I go along.
Here are the two images I have combined into one image on two layers, the bottle is the top/active layer and is smaller.
I did two test blends, both with the grey sliders. After moving the slider I moved the top layer around and watched the effect as it moved. Included in each blend is two images with just the top layer moved - same blend.
This acts just like masking the layer, the top layer floats on top of the bottom layer. Notice that the same areas show through as the top layer is moved.
For these I darkened the bottle to make it easier to see.
This acts like the underlaying image, where it punches through, is floating on top of the top layer, Notice that the bottom image is whole and unaffeected where there is not a top image and that the drain area shows through from the bottom image, no matter where the top layer is placed.
Stuff I learned from this:
-Use the top slider as another way to mask the top layer
-Use the bottom slider for objects in the lower image that you want to appear in front of the top layer
-In Photoshop I am used to thinking of how layers interact with each other. The blend-if sliders do not compare the top layer to the lower image, they don't care. What they are looking-at/thinking-about is the channel content of the top/active layer for the top slider and the channel content of the lower image for the bottom slider. They use that info to mask the top/active layer.
-Again, it is not the difference in the layers, but the content with-in the layer or the lower image that is important here. If there is something clearly different in brightness or color between the object you want to seperate and the rest of the image in the lower image, then blend-if will work to 'punch' it through. If there is something clearly different in brightness or color between the object you want to seperate and the rest of the image in the active layer, then blend-if will work to mask it.
In Conclusion, this is a powerful way of masking 'on the fly' without having to create selections which can be used along with layer masks. Thank you Ed for giving me the motivation to figure this out!