Class 2 submission: T Gorge

[mikemeister_admin]

My first submittal looked a little green.  I went back to the orginal to reduce the green.

Walt Smith

[ggroess]

It is still pretty green...

Try setting a neutral on some of the rapids foam.  The water is pretty dirty and the whiteish areas make a decent neutral.
Greg

[derekfountain]

My first submittal looked a little green.  I went back to the orginal to reduce the green.

Walt, you appear to have used a lizard tail type curve on one of the LAB colour channels. That doesn't work. The lizard tail is good for increasing or decreasing lightness or colour at the highlight or shadow extremes of a curve. LAB works differently, making the lizard tail much less relevant in the a and b channels.

The centre point of the a/b channels is neutral - grey - and the colour saturation increases as you move away from that centre point. However, in normal images it's only about the middle third of those a/b channels that represent the image's colour information. If you open up a colourful image in CM and move the mouse pointer over it watching the floating ball on the a/b curves you'll notice it rarely if ever moves towards the ends of the curve. The ends of those colour curves represent "theoretical" colours that the LAB colourspace can show according to its mathematical definition, but which don't appear in normal images. Thus applying a lizard tail to your a channel doesn't actually affect much of your image, and so it's still pretty green.

As a further experiment, open up an image then apply a curve to the a or b channel like the one I've attached. Then hit the compare button. Can you see any difference between the image with and without the curve applied?

[ggroess]

The easiest way for Walt to reduce the Green I believe is to select a new neutral. 

That being said, creating a lizard tail in this case does not address the overall color shift and in fact Walt's does not even effect the image; more telling is the control point he has on the A channel that pulls the center of the curve to the right of zero and into the green. 

Eariler last week we corrected the blue shift out of the kite fliers by making a modified version of a lizard tail to flatten a part of the curve and then move the neutral. But as you pointed out it was in the center 3rd of the grid.

Your points are all well taken in that everyone should to look at the "worm" on the color channels to see where you need to adjust the curves to effect the image.  This is where the "mark" command comes in very handy.  I will often hover the mouse over the area of the image effected by the shift and then, Right click on the image and select Mark.  This places a "marker" on the curve in all channels.  It does not add a control point just a placeholder.  I find it highly useful for correcting the "exact" range on the curve where the color lives.

BTW, This is exactly the kind of discussion we hope for and strongly encourage everyone to bring up. 
Derek Please continue to bring these forward.  I think the best solutions come out of these....

THANKS!!!

Greg

[mikemeister_admin]

Nice explanation.My boxes are set up for 0,64,128 and the only change comes when the curve sweeps the inner four boxes.GregM

[mikemeister_admin]

Here is my submission for the river shot.  What do y'all think?



  -- Russ 

[ggroess]

Nice work Russ...

If you feel like it you can push the saturation a bit more but you have a nice image right now...

Greg