Lights, white-balance, skin-tone, an experiment that surprised me.

I did a lighting- and white-balance test to see how it affects skin-tones and wrote my findings down and made an example-image.

If anybody is interested in reading it, I posted it on my facebook-page:

https://www.facebook.com/jacoblarsen.248/posts/10207613772208597

I was surprised by learning that different light-sources affect skin-tone differently even if you white-balance to correct for their differences.

Is this a known issue?
Is it some effect of hue-twist caused by the built-in picture-styles?

Anyway, I'm a bit surprised about this, so I thought I'd share my findings and see if anybody has an explanation.
Thanks :)
JEL
 
Oct 16, 2015
345
2
inbuild picture styles affect tonality in general but only if you shoot in JPEG and not in RAW.
WB correction of 8 bit JPEG shot does not work well for many photographers :)
WB correction is not enough for the skin tones to be accuratelly reproduced. it is a good starting point though. for setting accurate skin tonality, I would normaly create X-Rite colour checker passport profile and apply it to the shot in post.
 
Upvote 0

Besisika

How can you stand out, if you do like evrybdy else
Mar 25, 2014
779
215
Montreal
Color checker passport is a good thing to have, yet it causes me trouble when light is flickering especially when its distribution is not constant across the room. I shoot manga convention every year and the center is awful. The passport is totally lost there.
Besides, videographers don't have raw to play with so well balanced skin tone in camera, and not in post, is highly recommended. Not to mention that Premiere Pro CC doesn't accept the passport yet (unless I am not aware of it). I have to go to Davinci when I really have to.

Thanks for the post, I hope more tests have been made. I would benefit from it. I have trouble with skin tone as my eyes don't see the distinction.
Please post, if you happen to play more.
 
Upvote 0
Jan 29, 2011
10,675
6,121
JEL said:
I did a lighting- and white-balance test to see how it affects skin-tones and wrote my findings down and made an example-image.

If anybody is interested in reading it, I posted it on my facebook-page:

https://www.facebook.com/jacoblarsen.248/posts/10207613772208597

I was surprised by learning that different light-sources affect skin-tone differently even if you white-balance to correct for their differences.

Is this a known issue?
Is it some effect of hue-twist caused by the built-in picture-styles?

Anyway, I'm a bit surprised about this, so I thought I'd share my findings and see if anybody has an explanation.
Thanks :)
JEL

It isn't an issue, it is the very nature of light!

Ever wonder why good photo and video lights and computer screens cost more? Did you ever stop to and try to define a colour, in an absolute term? Ever look at a known colour, like a Pantone, in different colour spaces?

All light sources have spectral characteristics with spikes and troughs at different wavelengths, this makes for tonal differences irrespective of WB.

Actual colour science is an incredibly complicated thing and the apparent ease with which we see something, then illuminate it, photograph it, then try and reproduce that digital image either on screen or print is unbelievably convoluted, it's just the bulk of it goes on in the background without us ever being aware of it. Add in the fact that our cameras and our eyes see colour in a completely different way and you will be on the right track.
 
Upvote 0
privatebydesign said:
It isn't an issue, it is the very nature of light!

Ever wonder why good photo and video lights and computer screens cost more? Did you ever stop to and try to define a colour, in an absolute term? Ever look at a known colour, like a Pantone, in different colour spaces?

All light sources have spectral characteristics with spikes and troughs at different wavelengths, this makes for tonal differences irrespective of WB.

Thank you for that explanation :)
I suppose that the 2 monitors might indeed have different spectral characteristics, which would indeed explain the differences even at corrected white-balance. I didn't really think about that part (spikes and holes in light-wavelengths) until you mentioned it :) . They're both LED-monitors, but one is a 6 year old design, the other is from this year.

It just caught me by surprise when I noticed it :)
The monitors are not fluorescent lights, so I never thought along those lines, but makes sense now that you mention it :)
Thanks, I learned something :)
 
Upvote 0
Mar 25, 2011
16,848
1,835
The Kelvin rating of a light source reflects a overall color temperature, but two lights with the same Kelvin Rating can be vastly different in the way images look.

Professional LED lights, the ones costing thousands of dollars, pay careful attention to detail, so the light is the same from unit to unit.

Software / hardware solutions like passport color checker adjust colors at numerous points in order to attempt to overcome this. For skin tones, the more expensive versions have color chips with common skin tones to calibrate on.

Its always been this way. With film, some types were better for skintones, but lighting was still a issue.
 
Upvote 0
Private explained it pretty well. The light spectrum of you're screen will basically be three spikes of colour- one at blue, one at green, one at red. That will cause tonal variations when compared to other light sources that appear the same colour of white. They will also differ between screens depending on the exact tone of r g and b, and their respective intensities.
A cute inforgraphic pulled from RedUser, the whole discussion here: http://www.reduser.net/forum/showthread.php?148460-Dragon-Sensor-Tungsten-or-HMI-(Native-WB-) (discussion has more to do with CMOS than light. This graph will definitely show you why skin tones always look funny under flourescent light)
 

Attachments

  • IMG_2229.PNG
    IMG_2229.PNG
    78.9 KB · Views: 332
Upvote 0

Jack Douglas

CR for the Humour
Apr 10, 2013
6,980
2,602
Alberta, Canada
IglooEater said:
Private explained it pretty well. The light spectrum of you're screen will basically be three spikes of colour- one at blue, one at green, one at red. That will cause tonal variations when compared to other light sources that appear the same colour of white. They will also differ between screens depending on the exact tone of r g and b, and their respective intensities.
A cute inforgraphic pulled from RedUser, the whole discussion here: http://www.reduser.net/forum/showthread.php?148460-Dragon-Sensor-Tungsten-or-HMI-(Native-WB-) (discussion has more to do with CMOS than light. This graph will definitely show you why skin tones always look funny under flourescent light)

Thanks for this - helps me too.

Jack
 
Upvote 0