Link to: "how-to-get-perfect-skin-colors-with-every-camera"

[ErikKaffehr]

Hi,

This article discusses "hue twists" in Adobe's profiles and gives very simple advice on getting rid of them.

http://petapixel.com/2014/10/01/colorchecker-how-to-get-perfect-skin-colors-with-every-camera/

The short version is that Adobes profiles tweak colours depending on luminance (probably for a reason).

Generating a DCP profile with Adobe DNG Profile editor creates a profile without those tweaks, while Color Checker Passport has it's own tweaks.

Now, I would suggest those tweaks are intended to improve rendition, but it seems that the author of the article finds it much easier to work with the untweaked files.


Here is more about the hue tweaks: http://chromasoft.blogspot.se/2009/02/visualizing-dng-camera-profiles-part-1.html

The enclosed screendump shows how AdobeStandard profile tweaks some colours from the colour checker.

Best regards
Erik

[TonyW]

Thanks for the link Erik.  Have just ordered the Colour Checker Passport so nicely timed 

[Tim Lookingbill]

Thank you, Erik, for posting this.

At first I thought it was just a rehash of already known information about creating custom DNG profiles, but then I noticed Hans was creating a Dual Illuminant profile by individually selecting each CC chart based illuminant "Table: Tungsten, D65" separately in the dropdown menu and clicking "Create Color Table" instead of choosing "Both Color Tables" and then clicking "Create Color Table"...or does that give the same results?

And/Or...By starting out with Adobe Standard (mine as always been ACR 4.4-yellowish orange skin and reds) as the Base Profile which always adds a maroon-ish hue to warm skin colors as he's demonstrated in the image sample which basically forces a huge hue shift correction away from the maroon-ish look in the Look Tables of the new custom Dual Illuminant profile. Is that what's happening?

From what I was reading from Mike Blume's comment linked to in that article, this non-"Both Color Tables" method removes the 3D tables that invoked this hue twist in a custom DNG profile.

Other than that understanding I derived from that article, I don't see any new information.

And something that bothered me about the studio demo shots with the beauty dish is that the desired non-hue twist profile affected a wide swath of the hue range that changed the lipstick color giving an overall plastic look to the skin or a make-up job done by a mortician.

It's something this color grader didn't like and attributed it to something inherent to shooting digital vs film...

http://juanmelara.com.au/blackmagic-cinema-camera-davinci-resolve-colour-grading-breakdown/

..when from my own observations of light on skin tones has a lot more to do with the character and quality of light used.

[digitaldog]

Thanks for the link Erik.  Have just ordered the Colour Checker Passport so nicely timed 

Custom DNG profiles are very useful but I hardly ever seen the need to make more than one per illuminate. So don’t go crazy with ‘one per shot’, I think it’s unnecessary and illustrate that here:



In this 30 minute video, we’ll look into the creation and use of DNG camera profiles in three raw converters. The video covers:

What are DNG camera profiles, how do they differ from ICC camera profiles.
Misconceptions about DNG camera profiles.
Just when, and why do you need to build custom DNG camera profiles?
How to build custom DNG camera profiles using the X-rite Passport software.
The role of various illuminants on camera sensors and DNG camera profiles.
Dual Illuminant DNG camera profiles.
Examples of usage of DNG camera profiles in Lightroom, ACR, and Iridient Developer.

Low Rez (YouTube):
http://youtu.be/_fikTm8XIt4

High Rez (download):
http://www.digitaldog.net/files/DNG%20Camera%20profile%20video.mov

[Tim Lookingbill]

How does one explain what is happening in the shot below of sunset lit clouds between applying two table based profiles?

Is Adobe Standard engineered to control/tame gamut clipping in shots like these? Or are the hue twists tied to luminance meant for something else that might be interpreted as an attempt to find middle ground in defining "accurate" vs "pleasing" color?

[TonyW]

Custom DNG profiles are very useful but I hardly ever seen the need to make more than one per illuminate. So don’t go crazy with ‘one per shot’, I think it’s unnecessary and illustrate that here:...

Thanks Andrew, once again a very informative video and your practical demonstration illustrating why it may not be necessary to go OTT creating profiles struck home and corrected at least one of my misconceptions .

[ErikKaffehr]

Hi,

Thanks for responses everyone!

Personally, I read the articles with some interest. My own experience is limited, I have been using Minoltas and Sonys since 2006. In general I have been quite happy with the Adobe Standard profiles.

In 2013 I added a P45+ to my equipment. On that camera I was less than happy with the Adobe Standard profile. So I have tested different profiling solutions, including Adobe DNG Profile Editor, Color Checker Passport and QPCard. What I have found that I was most satisfied with Adobe DNG Profile Editor.

That is on the P45+. On the Sony Alpha SLT 99 I also have tested colours were very good with both the Adobe Standard Profile and the profile by Adobe DNG Profile Editor, so on the Alpha SLT 99 I stayed with the Adobe Standard Profile.

On the P45+, Capture One using linear curve was best (In terms of DeltaE* CIEDE 2000), but I stayed with the DNG Profile Editor generated profile in LR5, as I much prefer LR to Capture one in all respects except suppression of colour aliasing. But this is a personal thing.

Finding are here: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/79-p45-colour-rendition?start=3

Whole article on the issue is here:
http://echophoto.dnsalias.net/ekr/index.php/photoarticles/79-p45-colour-rendition?showall=1

Best regards
Erik

[ErikKaffehr]

Hi,

My primary guess is that the tweaks represent a way of achieving pleasant colour.

Would you be kind to donate some raw files, we may be more able to do some experiments trying to answer your questions.

I may try to dig up some dramatic sky images of my own and do some checks, but I don't have that many.

Best regards
Erik

How does one explain what is happening in the shot below of sunset lit clouds between applying two table based profiles?

Is Adobe Standard engineered to control/tame gamut clipping in shots like these? Or are the hue twists tied to luminance meant for something else that might be interpreted as an attempt to find middle ground in defining "accurate" vs "pleasing" color?

[Tim Lookingbill]

Hi,

My primary guess is that the tweaks represent a way of achieving pleasant colour.

Would you be kind to donate some raw files, we may be more able to do some experiments trying to answer your questions.

I may try to dig up some dramatic sky images of my own and do some checks, but I don't have that many.

Best regards
Erik

Erik, I think your previous response about the P45+ and after rereading Hans' article focusing on his quote...

Every camera is different: every lens, every sensor, every filter.

...pretty much answered my question along with figuring out (and still not sure I have, yet) what Hans was actually illustrating in the article was simply how to create a custom DNG profile that doesn't have a Look Table that inflicts a hue twist when adjusting Recovery/Exposure according to Sandy's article. What makes me not sure what problem Hans is is trying to solve comes from my going through my shots of daylight lit skin as opposed to sunset colors and find I don't get any hue twists when applying a custom DNG or Adobe Standard. See the example image below where even when I applied Recovery or Exposure adjusts does not twist the hues. Switching to Adobe Standard vs Custom DNG profile just changes the overall hues at the outset. If I make any luminance adjusts I get no hue twists just darker/lighter versions of the two different color renderings seen in the image below.

So I don't believe it is image specific (maybe camera sensor?) thus I don't see how examining my Raw file is going to come up with any answers. Besides I really don't feel like opening a DropBox account at this time unless you can give specifics on how and what you can derive from examining a Raw file.

Anyway I'm now thinking the more pertinent question that should be answered is whether Adobe Standard is engineered by Adobe specifically for each camera brand/model or is it a one size fits all profile. A more revealing experiment would be to have two different cameras record in Raw the same scene in a controlled lighting environment and apply Adobe Standard to each within the same version of ACR/LR and see whether there is a difference. If there is, then you'ld have to determine if the differences are caused by the two sensors/lens combo or whether there are tags within Adobe Standard that are triggered to tweak the hues according to the camera brand/model.

[Tim Lookingbill]

And to rule out contrast influenced edits I've provided Default ACR 6.7 settings with As Shot WB.

One with a custom Dual illuminant profile and Recovery applied...

the second with Adobe Standard and Recovery applied to show I get no hue twist.

So what Hans is really illustrating is the hue twists he gets with his specific camera which is probably why he included the quote that every camera/sensor/lens combo is different.

[ErikKaffehr]

Hi Tim,

I looked at some different images using:

- Adobe Standard Profile
- Adobe Standrad Untisted
- A DNG Editor generated dual illuminant profile

There we some differences but they were very subtle, I have also checked with some dramatic clouds images and the changes were still very subtle.

The profiles from the DNG Profile Editor and the Adobe Standard Profile on my Sony SLT99 are very close. So I use the Adobe Standard Profile.

On the P45+ I was never happy with the Adobe Standard Profile, so I generated an own one.

Best regards
Erik

And to rule out contrast influenced edits I've provided Default ACR 6.7 settings with As Shot WB.

One with a custom Dual illuminant profile and Recovery applied...

the second with Adobe Standard and Recovery applied to show I get no hue twist.

So what Hans is really illustrating is the hue twists he gets with his specific camera which is probably why he included the quote that every camera/sensor/lens combo is different.