weird color on face shot in the snow
AAABluestocking
Registered Users Posts: 116 Major grins
Why is there this weird color on the face? When I try to increase the contrast in the L - LAB curves for the snow it makes the face even weirder. How do I fix this?
Gallery of other shots that I'm happy with PP.
Gallery of other shots that I'm happy with PP.
My SmugMug Galleries
Learn the various techniques to make all things possible and then choose deliberately which you actually want-rutt
0
Comments
Is this a Raw original or you shot JPEG?
Author "Color Management for Photographers"
http://www.digitaldog.net/
OK good. Lets render the image as best as possible before we complicate things with Lab (which I bet is totally unnecessary). Do all the heavy lifting in the Raw converter.
What converter are you using and have you attempted to alter contrast and other rendering adjustments there?
Author "Color Management for Photographers"
http://www.digitaldog.net/
Jpgs can be opened in Adobe RAW converter in CS4 and when this is done you can try to white balance the image by clicking with the eye dropper on the white shoulder patch of the jacket, or the white lettering on the helmet, or if you believe the snow is white, on the snow. For me, the white lettering on the helmet seemed to work best for the skin tone, so after a but of adjusting to taste, I opened that as a Smart Object, with the intention of opening a second Smart Object for the snow, but found that a second Smart Object was unnecessary. I did some exposure correction in ACR as well.
The skin tones on the chin read about 28C, 45M, and 52Y which sounds in the ball park and looks reasonable. The cheeks are still quite magenta, but I think that is probably how kids cheeks look when skiing.
You can see my edit here and down load it. If you prefer me to take the image down after you have it, I will be happy to do so.
I see there remains a bit of chromatic aberration around the edge of the white on the jacket that could also have been remedied in ARC as well. With the original Raw file, the image can be significantly better in the highlights.
The CMYK values typically expected for Caucasian skin can be found here in Smumgmug's Skin Tone suggestions.
In LAB, you will expect the b channel to be positive on the neck, forehead, or back of the hand, and greater than the positive value in the a channel.
These statements are generally true for people with some sun exposure, and not wearing makeup. They do not necessarily work well with infants or small children in very cold weather, when the magenta will be higher due to their pinker skin.
Here is an excellent thread about skin tones
Moderator of the Technique Forum and Finishing School on Dgrin
Yeah, that is my intention.
I have a CR2 file. I was thinking the double smart object as well. It's explained in my Scott Kelby book...
I've grabbed a copy so it's up to you if you want to keep it there. That's funny I've read that in the context of beauty retouches...never occurred to me to use it on a kid. lol. I've gotten as far as the yellow/magenta but got lost. Now I see I need to save a CMYK curve. That's actually pretty easy!
So it's basically a white-balance problem. Makes sense.
Thanks everybody for your input on this. I was really baffled by this but then DH has put me through the paces trying to fix his pictures [probably should start a 'how I fixed it' thread. lol]. We're all just learning.
-Nancy
Not sure that's even necessary. At least on the JPEG, I see a need for highlight recovery which would be super easy with the Raw data. And Unfortunately, you're using an older version of ACR, newer versions have selective tone and color options (as does Lightroom 2.0). If you ever want to make the Raw available, be fun to take a crack at it with the newer versions.
Author "Color Management for Photographers"
http://www.digitaldog.net/
What to do about it? You could color balance on something you know (or strongly suspect) is neutral. Pathfinder chose the white lettering on the helmet. You could try the white on the jacket. Whatever. But there is a problem with this approach: the snow really is blue, even under sunlight. That blue is reflected and causes a cast in the flesh and other surfaces. So this image has a mixed cast. If you just balance it on one area you may get good flesh, but you will get bad snow. Or visa versa.
Here is another approach. Let's assume we like the colors in the highlights of this image, the snow and sky. But we don't like what these reflected colors do to the midtones, particularly the flesh and rocks. We can use RGB curves to tone down the blue just in the midtones:
I like this better than Pathfinder's version because the snow is more believable. And actually, I'm not so sure the letters on the helmet should be white. Do we want to show them reflecting the blue of the sky (like the snow does) or not? You could shadow and highlight locks from my curve to get some of the blue out of the highlights and shadows as well. It's a matter of taste. Play with it.
BTW, you don't have to use ACR to get the point and click color balance. PS curves have the same feature. The middle eye dropper is for neutrals. Bring up a curves dialog, click on the middle dropper, and click on a spot you want to be neutral in the picture and it will write curves that make that spot neutral. I use both ACR and curves to do this, depending on what happens to be the most convenient.
As for the blown mountains in the distance, convert to lab and use the highlights part of shadow/highlights: Set tonal width to 0 and play with amount to get just a little detail back:
Here's my final edit. My conclusion is that you definitely can have your cake and eat it too!