A palliative approach to detail extraction

Hi,

This is an interesting project for at least two reasons; first because it is almost exactly the opposite of what I was trying to do and second it is a way to get gobs of detail from an image that looks like there is not much.
You can access all these files here:
http://picasaweb.google.com/ronviers/Detail
If you want you can skip right to the end and load file, ViVslinearNormalvivid.jpg, here:
http://picasaweb.google.com/ronviers/Detail/photo#4989463967 360548882 In order to see the amount of detail I am talking about you might want to download the file.
Or if you want to follow along start with, Normal.jpg, here: http://picasaweb.google.com/ronviers/Detail/photo#4989464890 986135570 Normal is a picture of the original unmodified raw file just as it was shot.
The first step was to make a value channel by copying the entire image and pasting it into a channel – then invert.
The next step was to use that channel make a new layer called VALUEinvert which you can see here:
http://picasaweb.google.com/ronviers/Detail/photo#4989465484 683706386 Then that layer was duplicated and shrank (shrunk?) by one pixel on each side. This was done by placing it above the just created layer called VALUEinvert layer and setting the blend mode to
‘difference’, this makes it easy to use the transform control to shrink one side at a time – you have to be very careful not to go more than one pixel or it will mess it up. After the shrinking, the blend mode was reset to ‘Vivid’ and the layer named VALUEshrink. Which you can see here:
http://picasaweb.google.com/ronviers/Detail/photo#4989466083 981918226 Then these layers are grouped together and a normal layer placed below that group with the group’s blend mode set to ‘Linear’. You can see the results of this blend here:
http://picasaweb.google.com/ronviers/Detail/photo#4989466988 741263378 It is a good idea to group these layers together so they do not get mixed up with other things.
Finally you can see the end result here:
http://picasaweb.google.com/ronviers/Detail/photo#4989463967 360548882 Notice how the high key background now has so much detail. I had to go back to see if it was really there.
If this gets a little confusing you can see the work area here: http://picasaweb.google.com/ronviers/Detail/photo#4989462383 700738066 You can create a grow group if you like with the only difference being that instead of shrinking the value channel by one pixel on a side you grow it by one pixel.
Seemingly the only difference between the grow group and the shrink group it that the light and dark side of the edges reverses sides. One thing I learned by doing this is that there are primary edges in a scene with a single light source. Of course there are edges throughout a scene but if you really want to see how dominant the primary edges are then do the grow group and compare them. This would make a nice assignment for a beginning Photoshop class.
Btw, the original idea here was that I would put a VALUEinvertGrow mask on a VALUEinvertShrink layer and blend that back to a Normal layer in order to isolate the yarn fuzz – needless to say that didn’t work – an ironic attempt to avoid the cumbersome ‘Extract’ filter. Please comment.
Good luck,
Ron