Curvemeister Forum

Just working on this photo of Faneuil Hall in Boston. I've set the highlight on the lightest bit of the clock tower, and shadow in the darkest bit of the porch in the lower right, and the neutral on the white wall of the clock tower.

In Lab and HSB modes it looks about right; in RGB and wgCMYK modes it looks, well, green! I'm working in 16 bit mode, but it does it in 8 bit too. If I change from Adobe 1998 to sRGB it behaves properly.

This is with CM 3.0.5.

I've put the full size JPEG here: http://www.derekfountain.org/misc/img_2007.jpg

Hey Derek,

Can you go back into the RGB version and nudge the neutral a bit and see if it is the sample size causing the issue?
You might have some yellow/orange pixels in the sample skewing the neutral.  Just want to make sure....

Thanks!
Greg

Can you go back into the RGB version and nudge the neutral a bit and see if it is the sample size causing the issue?
You might have some yellow/orange pixels in the sample skewing the neutral.  Just want to make sure....

No, it's trickier than that. I was just playing moving the markers around a little and noticed that in RGB mode I can get it looking mostly OK. If I then switch to Lab mode it still looks OK, but when I then switch back to RGB mode without moving any of the markers it goes green again. Most odd.

Besides, shouldn't the results pretty much match for all modes? If there were some rogue pixels in the sample point, wouldn't they have largely the same effect for each mode?

They should,  but often when I have seen this in class the problem was sample not program. 
On the otherhand switching between modes should not effect it unless you have more than 1 neutral set in RGB. 
Lab has to choose one.  Doesn't sound like the case here.  This one might have to have the Meisters touch applied...

I'll try it out on another image....but my results were pretty consistant

Greg

In this case the neutral is set to a relatively light part of the image.  Lab mode only uses one neutral, so the colors are more stable.  In RGB mode the shadow and highlight are treated as neutrals, which makes three neutrals total.  The neutral is close to the red highlight, and this has the effect of adding an unnatural bend to the red curve, pushing the whole image toward green.

What I don't have a handle on is why the image looked OK in RGB mode, then went green on returning to RGB from Lab.  I'll see if I can reproduce this.

What I don't have a handle on is why the image looked OK in RGB mode, then went green on returning to RGB from Lab.  I'll see if I can reproduce this.

Hmmm, I only saw that once, last night when I was pretty tired. I think it happened, but given a sensible explanation for the effect I saw, I can't help wondering if I imagined it... If you can't reproduce it, or see why it should have happened, that might be the explanation. :-[

I doubt that you imagined it.  I've seen this happen myself, in earlier versions, but have not been able to document the the steps needed to recreate the problem.

um...nice icon....

travel seems to be good for you...

What can I say?  Spinach was in season