Lightroom/Camera Raw: Camera Standard profile for A7R III is terribly off

  • 4
  • Problem
  • Updated 10 months ago
  • (Edited)
The Camera Standard profile Adobe has created for the A7R III adds a notable yellow cast to the image (unlike say Neutral or Vivid). Clearly not a match for the in-camera Standard profile when shooting jpg.Dear Adobe, can you please correct this?thanks
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes

Posted 2 years ago

  • 4
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
My CC Passport will arrive next week!

In the meanwhile I will be suing Adobe Standard, I think it is the most accurate in terms of color (I mean closer to jpg) and does not exhibit any color shift under any scenario (which is what we would like Camera Standard to do...). It is quite muted but nothing that cannot be fixed in the Basic Panel that I would use any way.
(Edited)
Photo of Todd Shaner

Todd Shaner, Champion

  • 1869 Posts
  • 626 Reply Likes
Andrew Rodney sid: Yes, the JPEG output's a desired rendering based on proprietary processing inside the camera. The raw can be rendered any way one desires within reason. 
I don't believe Adobe EVER guaranteed that their profiles will match the camera JPEG exactly nor should they. BTW, exactly would be within 1 or less deltaE; every pixel matches.
Andrew, I think you're over-analyzing this issue. I've checked numerous raw + JPEG file pairs shot with my Canon cameras. In all cases the camera JPEG very closely matches Canon's Digital Photo Professional raw converter with 'As Shot settings. They also very closely match LR's Camera Standard rendering of the CR2 raw file with As Shot WB and a very slight adjustment to Exposure (+0.05 to +0.10 EV). The images look virtually identical when doing a critical AB comparison using LR's Loupe with the right or left arrow key to quickly swap the image previews. I can't tell you what the Delta E is, but they are far and away much, much closer than the OP's Sony A7R III Camera standard images.

I rest my case and leave it up to Adobe to decide if it's something worth fixing. That's not my call, or yours, or anybody else's here. Thank you!
Photo of Cameron Rad

Cameron Rad

  • 161 Posts
  • 49 Reply Likes
Seems like it'd be Sony's problem before Adobe's. They should "fix" their RAW Converter so Adobe has a better reference. Now what happens if someone liked the Sony Imaging Edge result and all the Adobe profiles were made to match SOOC JPEG? Would the story be flipped the other way? "ACR Camera Standard doesn't match Sony Imaging Edge Camera Standard with the same RAW".
(Edited)
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
haha I think they should all aim to match the jpg. I don't think people want to match a bad Sony conversion. I am sure Sony will not do anything, converters are not their core product. Does Adobe really need Sony to fix their converter before they can fix their profile? Can't they aim for jpg instead of RAW+CS in the Sony converter?
Photo of Andrew Rodney

Andrew Rodney

  • 938 Posts
  • 192 Reply Likes
>>In all cases the camera JPEG very closely matches 

But you have no deltaE values to tell us how close. You're telling us subjectively they appear close

>> I can't tell you what the Delta E is,

I can! But you have to understand deltaE only compares TWO solid values. An image contains perhaps millions of pixels so we'd need an average dE which again, I can produce. 
You also need to accept the issues surrounding metameric failure for some of those groups of pixels being compared. You cannot fix this with a profile. Neither can Adobe or anyone else. Unless the camera follows the Lurther Ives condition, and I know of no shipping camera that does, this is another issue.  

>>I rest my case and leave it up to Adobe to decide if it's something worth fixing.

You can rest the case, don't hold your breath that the evidence you have will convince Adobe of anything other than to move forward improving all aspects of how it renders colors. You may see newer profiles in the future for all cameras but that in no way implies these profiles are anything but Adobe's continuing work in improving their colors. 
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
Without overcomplicating, we are just hoping Adobe can achieve with Sony profiles the good results they achieved with Canon ones. If we like what we see and we manage to print what we see who cares about deltaE we cannot see (nor measure us human)?
The ultimate user of the pic will judge a picture based in what he sees, never heard anyone complaining that deltaE were too high when looking at a print.
(Edited)
Photo of Jao van de Lagemaat

Jao van de Lagemaat

  • 228 Posts
  • 76 Reply Likes
It is not unlikely that an early version of the camera rendered identical to the raw software but that Sony did a firmware update (wasn’t there one that ate stars?) that changed the rendering and that they did not update their raw software.
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
The firmware v1 should have addressed only a bug in the naming system plus other minor bugs. Would be surprised if they changed the rendering.
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
The CC Passport has arrived...it is going to be a fun weekend... stay tuned :-D
I was hoping to hear something from the Adobe employees that monitor the forum.
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
Guys, I shot the colour checker and created single and dual illuminant profiles with both the Adobe (did not change any settings in the Color Tables and Tone Curve tabs, is that right? What is the base profile I should start from?) and X-Rite software.

Here are the target shots I used to build the profiles, in case you want to check whether the set up was right
https://www.dropbox.com/s/5gdgwm778iy4650/2018_01_untitled_001.DNG?dl=0
https://www.dropbox.com/s/xzinsxxjil9wb6i/2018_01_untitled_004.DNG?dl=0

Here are the profiles
Adobe
https://www.dropbox.com/s/odniyeec69snrh1/DPE%20A7R%20III.dcp?dl=0
https://www.dropbox.com/s/0r90eqwbvjeqvkl/DPE%20Dual%20A7R%20III.dcp?dl=0
X-Rite
https://www.dropbox.com/s/0jrko90bt4y5y48/X-Rite%20A7R%20III.dcp?dl=0
https://www.dropbox.com/s/c8xgji7ynnqcxso/X-Rite%20Dual%20A7R%20III.dcp?dl=0

For convenience I am putting here the link to my images (raw and jpg) from the previous posts

1
https://www.dropbox.com/s/mpgmgxstlpcfixx/DSC01411.JPG?dl=0
https://www.dropbox.com/s/fsmt09hr3pvouyv/2018_01_Ghana_637.dng?dl=0
2
https://www.dropbox.com/s/1n9ftiaqqcwk9u7/DSC01695.JPG?dl=0
https://www.dropbox.com/s/v252om2fq3c4l7f/2018_01_Ghana_891.dng?dl=0
3
https://www.dropbox.com/s/6eq3stnmfn723wr/DSC02422.JPG?dl=0
https://www.dropbox.com/s/lbcpuppcxna77tu/DSC02422.ARW?dl=0
4
https://www.dropbox.com/s/ulhb8xfrggi5xea/DSC02421.JPG?dl=0
https://www.dropbox.com/s/vcrbxiwsnwgbb70/DSC02421.ARW?dl=0

I was surprised to see that using the same target pictures Adobe and X-Rite produce different colours (look at the blanket in n2!), who is right and who is wrong? If one color is accurate then the other is not and if you spend £80 for the CCP to get accurate colors this does not make you happy...

In certain instances (shot n. 3 but obviously not n.1), Camera Standard seems a better match to the jpg.

No yellow issue with the X-Rite and Adobe created profiles.

Overall I think I prefer the X-Rite profiles, they are more pleasing (castle shot, compare the color of the shirt of boy running, or compare color of the car in n.4)  but not sure they are more accurate (if I see that car again want to check with my eyes if it is blue or violet!).  

What is your opinion?
Thanks!
(Edited)
Photo of Jao van de Lagemaat

Jao van de Lagemaat

  • 228 Posts
  • 76 Reply Likes
I'll repeat a previous post here: "no profile will fix metameric failure of the camera".

The profiles you create using the color checkers are only valid in the exact same lighting conditions as you shot the color checker. If the light spectrum changes in any way, it is no longer valid. 
Photo of Todd Shaner

Todd Shaner, Champion

  • 1869 Posts
  • 626 Reply Likes
Your procedure is correct as you've described, but you can open both the CCPP Daylight and Tungsten DNG files in DPE at the same time. There's no need to create separate Daylight or Tungsten camera profiles. A single dual-illuminant profile will provide the best results under those lighting conditions and "mixed" Daylight + Tungsten lighting.

1) Select the Tungsten file and assign the target camera profile in the 'Base Profile' selector. In the Chart panel align the four color dots, in step 3 select '2850K Only' and create the 1st Color Table.

As you've observed the Base Profile selector is fixed with the 'ColorChecker' setting when you now select the Daylight DNG file. I believe this is a bug and using Edit> Clear Color Adjustments or Clear ALL Adjustments and in the Chart panel does not change it. Normally when you open a DNG file in DPE the embedded WB setting and camera profile is automatically loaded and appears in the Base Profile selector. You'll need to manually select the target Base Profile.

2) Select the Daylight file and assign the target camera profile in the 'Base Profile' selector. In the Chart panel align the four color dots, set Daylight '6500K Only' in step 3 and create the 2nd Color Table.

3) Use CTRL + E to export the camera profile to the User Camera Profiles folder giving it a unique descriptive name. Optionally you also save the recipe to another folder for future use. I suggest doing this so you can go back later and add changes using the Tone Curve or Color Matrices panel without having to run the ColorChecker again.

For more details download the PDF tutorial here:

https://www.adobe.com/content/dam/acom/en/products/photoshop/pdfs/cs6/DNGProfile_EditorDocumentation...

Profiles so created seem to be more pleasing than the ones starting from Adobe Standard. What do you guys think?
IMHO-The DPE dual-illuminant camera profile with Camera Standard base profile appears the best and has the most accurate colors. Good job!
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
Yes, maybe there is a misunderstanding, I did not create separate Daylight and Tungsten, I was talking about the procedure to create dual illuminant (but if you mean the "pairs" I have linked, one is daylight and the other is Dual Illuminant).

What I did is exactly what you described. With one difference, you say:

"As you've observed the Base Profile selector is fixed with the 'ColorChecker' setting when you now select the Daylight DNG file. I believe this is a bug and using Edit> Clear Color Adjustments or Clear ALL Adjustments and in the Chart panel does not change it."

In my case ColorChecker is not fixed in the base profile menu, I just replace it with the Camera Standard option when building the second table.

Anyway I tried to build another one leaving base profile on "ColorChecker" instead of CS and to me looks identical?

Have a look
https://www.dropbox.com/s/cmqo3z5wzvv3rlo/DPE%20CS%20II%20A7R%20III.dcp?dl=0

I agree with you that the result is pretty good, no yellow cast and good color accuracy. I think X-Rite Dual and this modified CS will be my go-to options to start with. X-Rite if I want something more punchy, mod CS if I want something more subtle. 
I have noticed that X-Rite Dual does a terrible job (yellow cast) on the colorcheker shot in tungsten light (file ending "...001"). Any idea?

What about tone curve, is "Camera Raw default" an improvement over "Base Profile"?


BTW I am still puzzled that X-Rite and DPE will render colours so differently in the very same target pic.


PS My CC Passport is very clunky, very stiff, feels like I have to brake it to open it, same with yours?
 
(Edited)
Photo of Todd Shaner

Todd Shaner, Champion

  • 1869 Posts
  • 626 Reply Likes
What I meant is that you don't need to create a separate 'Daylight' OR 'Tungsten' camera profile. The dual-illuminant Daylight/Tungsten camera profile "automatically" switches between the two tables based on the LR or ACR WB setting. Scroll down to Adobe Engineer 'madmanchan's' reply here:

http://forum.luminous-landscape.com/index.php?topic=43733.0

Anyway I tried to build another one leaving base profile on "ColorChecker" instead of CS and to me looks identical?
Good to know it doesn't matter. I checked both of your profiles (CS and CS II) are they look identical when applied to the Tungsten and Daylight CCPP image files.

What about tone curve, is "Camera Raw default" an improvement over "Base Profile"?
In the DPE Tone Curve panel check 'Show Base Tone Curve' and compare it to the 'Camera Raw Default,' which is what Adobe Standard uses. You can create a dual-iluminant DPE Camera Standard profile with the 'Camera Raw Default' setting. It will have lower contrast, which may be useful with some images.

BTW I am still puzzled that X-Rite and DPE will render colours so differently in the very same target pic.
I see the exact same X-Rite camera profile rendering with all my cameras in the Tungsten CCPP test shots. The X-Rite camera profiles are not twisted, which may be causing the difference with lower color temperature lighting. That's why I said, "When things don't look quite right" I use the Adobe Standard profile. It seems to work better with non-Daylight lighting (Tungsten, Fluorescent, Mercury Vapor, Purple UV). Also consider the Camera Faithful & Neutral profile to tone-down high-contrast shots. That's why it's also useful to create a custom Adobe Standard Legacy profile with a moderate 'S' Tone curve and Color Matrices Saturation boost. The DPE Tutorial has examples for both.

PS My CC Passport is very clunky, very stiff, feels like I have to brake it to open it, same with yours?
Same here with both of my CCPPs. Open carefully and try not to touch the chart patches. ;.)
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
Thanks Todd, very helpful.

Ok, I understand what you mean. I created a profile for Daylight and one Dual Illuminant out of curiosity, to compare results. While in the case of X-Rite there is no difference when rendering the daylight shot, in the case of the Adobe profiles I do see a difference in the blues, quite subtle. Your experience is different?

Regarding the tone curve tab, I guess the profiles I created starting from Adobe Standard as base profile (and not Camera Standard) already have the Camera Raw Default curve incorporated even though the tone curve setting was on "Base Profile" because as you say that is what adobe standard uses so I already have them.
A third possibility would be to use CS base profile + Camera Raw default tone curve. But at that point probably will look identical to AS base + base tone curve?  
Photo of Todd Shaner

Todd Shaner, Champion

  • 1869 Posts
  • 626 Reply Likes
While in the case of X-Rite there is no difference when rendering the daylight shot, in the case of the Adobe profiles I do see a difference in the blues, quite subtle.
This is as expected and explained by Adobe Engineer at the link I provided.

Dual Illuminant Profile Characteristics

Interpolates between the two tables based on your white balance setting. Specific method is inverse correlated color temperature. See DNG Specification if you wish to see the details & mathematics.

Improved color reproduction over a wider range of scenes. Degree of improvement depends on the camera model. If the scene lighting conditions (e.g., office fluorescent) varies greatly from the two illuminants used to build the profile (e.g., Solux bulb and natural cloudy daylight) then all bets are off. You will get something "ok" but results will be better under the unusual lighting condition by building a profile for that condition.

It's not the shooting temperature that matters, but the spectrum of the illumination. This is more complex and not easily measurable. General advice: if you tend to photograph under daylight, not worth building profile for each flavor of daylight, regardless of the color temp. There will be minor variations between ~D50 and ~D75 lighting (roughly 5000 K and 7500 K CCT) but this tends to get eliminated once WB is considered. If you photograph under unusual artifical lighting often, consider building a profile for that lighting.

A third possibility would be to use CS base profile + Camera Raw default tone curve. But at that point probably will look identical to AS base + base tone curve
No. The CS profile has different color matrices values than AS, which have higher color saturation.
Photo of Cameron Rad

Cameron Rad

  • 161 Posts
  • 49 Reply Likes
I'm surprised this topic is still going...

Here's a generic dual illuminant profile I made. It's not meant to match camera standard or anything and I can't promise it'll work right for you or your specific camera 
 http://cloud.cameronrad.com/p8BV/download/RAD%20-%20Generic.zip
(Edited)
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
Well, just playing with the profiles and comparing differences...

Thanks, for the profile, what software did you use to build it?
Photo of Cameron Rad

Cameron Rad

  • 161 Posts
  • 49 Reply Likes
Lumariver Profile Designer
Photo of Todd Shaner

Todd Shaner, Champion

  • 1869 Posts
  • 626 Reply Likes
I've just discovered the root cause of the Sony A7R III camera profile rendering issues:

https://forums.adobe.com/message/10263842#10263842

It appears Adobe used a pre-production camera to create the Sony A7R III camera profiles. Apparently Sony made available a number of these cameras for test purposes with firmware v0.01. Images shot with these cameras look fine with the Adobe created camera profiles:

http://www.photographyblog.com/previews/sony_a7r_iii_photos

Production cameras with firmware version 1.0 or the later 1.0.1 version exhibit the rendering issues being discussed here.
(Edited)
Photo of Mario Adario

Mario Adario

  • 85 Posts
  • 2 Reply Likes
Interesting. Hopefully Adobe will fix it once for all. Seems Eric Chan is on it.
Photo of Chris Hase

Chris Hase

  • 2 Posts
  • 0 Reply Likes
Could someone please clarify if this color rendering issue has been fixed in a software patch yet, as currently thinking of purchasing Lightroom...?
Thanks,
Photo of Johan Elzenga

Johan Elzenga, Champion

  • 2836 Posts
  • 1211 Reply Likes
This is just one particular profile that is/was off. Other profiles for the A7III are fine. I have this camera and I love the Adobe Landscape colors. No reason to base your decision on one profile.
Photo of Chris Hase

Chris Hase

  • 2 Posts
  • 0 Reply Likes
ok thanks for that, I'll be using the a7R III so I'll have to give it think i spose.
Photo of Guilherme Escosteguy

Guilherme Escosteguy

  • 1 Post
  • 0 Reply Likes
I ́m still struggling with this issue.
My images present a yellow/greenish tint in the middle to dark tones, specialy visible in skin tones.
Also the color transitions in dark areas of the skin present an abrupt color change in certain tones oversaturating the yellow to orange tones and crating very strange artifacts in the blue chanel. The highlights and small specular whites are never clean of yellows making the hair look awfull...
Mosto of the time people I photograph looks like a Simposon character and trying to solve this with camera calibration just mess with any other color in the scene.
Adobe profiles are terrible, but the Camera matching ones are way worse.

In Capture One I tried other camera profiles and had better results with Canon 1Dx profile anda with Adobe DNG colorspace with bumped saturation adjustments.