Color theory - why are the whites going blue?

[Pete B]

I moved the middle sliders of each of the RGB levels around to get this, but when channels are blown, there's only so much you can do.
Pete

 

[Sparrow]

... or use Fuji superia 400 and do AutoColour in Photoshop?

 

[Sparrow]

It wouldn't help if the root cause is dissimilar light sources or UV phosphorescence. Both impact film as readily as digital and film is more limited in the degree of correction that can be applied in post, digital or wet darkroom.

... I'm not sure that's true

 

[brusby]

The OOF highlights on the corrected image look weird to me.

Here's an attempt. In ColorPerfect's TouchUp function, I clicked the neckstrap on the left side to set the colour balance. Then back in CS5 using the Levels's grey eyedropper I clicked the light grey part of the "bend" of the frame of the glasses.

More could certainly be done, for instance the skin is still a bit too red for my taste but, it's a start.

cheers
Philip

Phillipus, you seem to have gotten the essence of what I was trying to achieve in skin tones without the blue tint to the whites. But that leaves the obvious question of how we're getting similar skin tones but drastically different whites when we're both using a global approach to correcting color. I'm suspecting maybe I screwed up when I made my Colorchecker profile. I was shooting so fast . . . But thanks!

 

[Pete B]

I desaturated cyan in your second file here:
Pete

 

[brusby]

In LR4 I hit Auto WB and got a pretty good result but still a bit too blue. The t-shirt is white in the light and pale blue in the shadows. Temp= +50, Tint= +14.

Next I grabbed the Temp slider and increased it to +85. This gives nice warm skin tones while the t-shirt is a warm-white. Looking good.

Added Vibrance +25 and Saturation +11 to give a stronger colour to his face. Boosted Clarity to +15, and then pulled back the Highlights to -19 as they were looking a little pale after boosting Clarity.

Vignetting -35.

The only remaining problem is the blown t-shirt. I'd use an adjustment brush to bring it back within the exposure range (Edit: just done. Also desaturated the white t-shirt at the same time. But forgot to do the same for the neck strap 🙁).

Pulled back exposure to -0.10 for a more conservative look.

That now looks pretty good to me. What do you think?

Very interesting Lynn. I like how you're able to keep the whites white and the skin tones are in the ballpark for where I was trying to put them. I think we see things a little differently particularly regarding contrast and how we like faces rendered, as I think you may like things just a bit brighter and more contrasty than I do. But I appreciate your approach and your ideas.

Hmmm, why are only my whites going blue? Starting to suspect my color profile.

 

[brusby]

I desaturated cyan in your second file here:
Pete

View attachment 98294

Aaaaahhh, so maybe my profile or my adjustments left too much cyan which was most obvious in the whites, but not so apparent in the skin color.

Now we're getting somewhere.

Thanks!

 

[brusby]

It could be UV Phosphorescence, UV light causing the chemicals in the white fabric to glow in the blue to violet portion of the visible spectrum. This is common when fabrics have been treated with "optical brighteners" like those in many laundry detergents.

I did a quickie adjust in LR4 using the White Balance sampler and then sliding the Highlights slider all the way dark. The WB sampler initially produces what seamed a clean white, but when I dialed down the highlights a cyanish tinge returned. Killing the saturation for the Aqua and Blue in HSL gave this result:

Dwig, I think you may have the explanation. Great detective work! Thanks!

 

[brusby]

The T-shirt will have been washed some regular laundry detergent containing optical brighteners (which down-convert UV to blue to counter the natural yellow tinge of cotton). When underexposed, it IS slightly blue. What is more, there is no other good white reference point in that image either - the reflection in the glass is a stop or two above white, the light outside the window seems to be warmer than the light on the subject (flash, or a south view vs. lit by north window issue), and eyes aren't really white either until they have been post-processed.

Yes, I'm starting to think this is at least part of the explanation.

 

[brusby]

I'm coming to the conclusion there are two contributing causes. First, as so astutely pointed out above, was an existing blue tint to certain white fabrics from optical brighteners, and second, too much cyan left in my image which disproportionately affected those whites which were already too blue.

Thanks everyone!