ACR for Portraits (A Solution)

Hi All,

I am having big problems using ACR to produce good skin-tones in portrait images, even calibrated using a script they are way off.

I think I may have found a solution and would like your opinions, the image below was processed as follows:-

Left hand Image:-

Image loaded into ACR all settings set to zero and linear curve applied, WB of mid grey square of colourchecker chart. loaded and assigned my ICC profile. In essence I applied an ICC profile to a linear image.

Right Hand Image:-

Image loaded into ACR all settings set to defaults, WB of mid grey square of colourchecker chart.

Observations:-

Skin tones look more natural on left hand images, they look to red/orange on right hand image. Also note background is neutral in left hand image and has red cast to ACR default image.

I am looking for more raw studio shot of people if anybody can help, pls mail to info AT huelight.com

HTTP response: NOT FOUND | MIME changed to 'image/png'

I'm not sure I understand what you mean by "assigned my ICC profile"? What profile?
In the right image, the third gray patch is the only one that's anywhere near neutral...


What does the image look like if you set the "proper" values in the "calibration" tab?

What does the image look like from DPP? (picture style: Neutral, WB on grey field, all other settings "0")

If I understand correctly, Harrymilker created a custom profile setting in ACR by shooting the color checker with his camera under controlled lighting conditions and custom white balance, then adjusting the color checker picture in ACR to get it dead (i.e. gray is gray, black is black, etc...) on so that the calibration tab in ACR can be used for that camera to get photos back to the right balance since most cameras are off a bit and need some tweaking in a controlled environment. This means now that all picture he can run through this calibration profile in ACR and be closer to a good PP start than not having a profile in ACR for his camera.

IO

If the OP hadn't stated "even calibrated using a script they are way off", I'd think that as well