Taming the Skin Tones of M240

[verendus]

Advertisement (gone after registration)

Yes, the image is on the cooler side as I had to adjust the WB to make the default image acceptable.  Do you find the top image on the cooler side as well?  If I were postprocessing the file using the above settings as import default, I would've probably set the WB around 500K warmer.

[thrid]

Yes but C1 offers "sessions" as well : http://tinyurl.com/l3pz2lk.

 

 

Yes, I am looking in to that. 

 

thanks

[jaapv]

Yes, the image is on the cooler side as I had to adjust the WB to make the default image acceptable.  Do you find the top image on the cooler side as well?  If I were postprocessing the file using the above settings as import default, I would've probably set the WB around 500K warmer.

Personally I find the top image more pleasing. As for skin tones, we should realize that when we see a person in the flesh our brain will filter out unevenness of tint. In fact the camera is not seeing the skin wrong, but more accurately.

The thing to avoid in postprocessing is overcompensating and creating a Barbie skin.

 

To (re)quote an old post on the difficulty of correcting for skin:

 

 

f we consider the anatomy of skin in as far as it is relevant to photographic rendering we find a layered structure. A fatty base layer, opaque yellowish white, then a layer of capillary veins red to blueish, depending on oxygen saturation and with a density depending on the distribution of veins and arteries. Highly IR reflective. Then Melanin particles which are visible light absorbent, UV reflective and providing toning from Finnish nothing to the blackest African skin, covered by a translucent white layer.

This means a large amount of metamerism and a virtually impossible exact rendering.

[Rob Roskopp]

Hello,

As a standard import preset, I start with

Camera Calibration - Red Primary

Hue +8

Saturation -8

and no further adjustments to HSL/Color (I want to reserve this sliders for later individual adjustments). Most common action would be to reduce the magenta saturation...

 

Depending of the situation (especially indoors/artificial light and in shadows) I often have adjust the white balance (from the auto-values), typically reducing the temperature by 300-500K, and increasing the magenta tint slightly(+5).

 

Funny enough, the JPEG previews (created with the same auto-white balance settings) look better (cooler), than the imported DNGs in Lightroom. It almost seems, that the JPEG engine of the camera and Lightroom have a different interpretation of a color temperature.

[algrove]

Yes, the image is on the cooler side as I had to adjust the WB to make the default image acceptable.  Do you find the top image on the cooler side as well?  If I were postprocessing the file using the above settings as import default, I would've probably set the WB around 500K warmer.

You know your subject better than us, but the top image looks way more natural to me on my calibrated monitor. I did not see an answer as to if you are using a retina display.

[verendus]

I have to eat my words.  I am using Apple Cinema Display.  Just out of curiosity, I recalibrated my monitor, and sure enough the calibration has drifted quite a bit.  The images are now much cooler, and the settings I posted seem even cooler and more pale.  I have set the Red Primary settings back to 0 and now using -9 Red, -20 Orange with the split tone settings (this takes care of slight magenta highlights).  I will see if I can repost the image. 

[Herr Barnack]

Advertisement (gone after registration)

When faced with mixed light sources, I use one of these and choose the greycard mode in the White Balance sub-menu http://www.bhphotovideo.com/c/product/375202-REG/Lastolite_LL_LR2050_EZYBalance_Grey_White_Card.html

 

This method never fails to nail the color balance and produce accurate and pleasing skin tones, making for less diddling in Lightrooom.

[algrove]

OP

That's good new. Gladly I used the terms "recently calibrated monitor".

[verendus]

Ok, after recalibrating my monitor, I came to conclusion that the split tone and red primary adjustments were unnecessary.  

So it is now just:

 

Red -9

Orange -20

 

The again is straight OOC, the bottom is with the above adjustments + slight curve adjustment for global contrast.

I take this as a good lesson on the importance of color management.  :-)  Let me know if the latter edit is more to your liking.  I still think the OOC shot is too red and orange and not natural.

 

 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[farnz]

The LAB numbers on the 'white' shirt area directly below her chin are as follows.  I used this as a reference point because its a block colour so there are fewer variants whereas the colour of the skin and teeth contain comparatively much more local variation so it's more difficult to compare apples with apples.  I think that the colour mods will apply equally all over the picture because they're profile (global) modifications.

 

   Original       1st mod.      2nd mod.

 

L      93                92              96     

A       0                 -2               0

B      -5                 -6              -3

 

What this shows first of all is that the L channel is not consistent (93, 92, 96) so the 1st mod is slightly darker and the second is brighter than the original.  RGB colour space will do that because a third of the luminance information is integrated into each channel and it often affects pictures almost imperceptibly.  This inconsistency is not a big deal but luminance can affect our perception of colour in a picture so it's something to take into consideration.

 

The second thing I notice is that the A channel in the original and 2nd mod are both zero, ie neither more red than green, but that the first mod is -2 and therefore greener.  Without a white balance card it's difficult to know what the true colour of the reference shirt is but because it's gone greener in the 1st mod and back to zero in the second suggests that the monitor's calibration was out (as I think has now been established and righted).  The next thing I notice is that in all 3 pictures the B channel is negative so there's a light blue element at play, which might actually be the colour of the shirt itself.  I took samples from the cheek and teeth too and they suggest that the shirt has a blue 'bias' tint because they all rose and fell by similar amounts.

 

The important question is: Are the skin tones in the second mod more realistic than the OOC original or 1st mod?  Her true skin tones in that lighting could fall within a broad range of tints depending on her ethnicity, her tan, the amount of exertion she's recently done but to my eye on my screen it looks truer than the original or the 1st mod.  The LAB numbers show that the major difference between the original and 2nd mod is that the yellow component has increased but the reds and greens have stayed the same, which may be a surprise.  The increase in luminance may be masking this a little though and on my monitor her teeth appear to have a slight green tint, which is supported by the LAB numbers.

 

Pete.