M/SL2-Colors

[Ralf.Lange]

Advertisement (gone after registration)

If I photograph the same subject, with the camera's automatic white balance, each with identical settings for the exposure, each time with the same (M) lens, once with the M10 / M10R and once with the SL2, the M image will be from both Lightroom and Capture One developed rather (too) warm and the SL2 image developed rather (too) cold. Leica has therefore opted for a completely different white balance for these two camera systems. The result of the SL2 images is a bit more sterile / digital. Even the "Pro-Profile" for the M10R advertised by Phase One leads to surprisingly unnatural, far too warm colors, at least for my taste. With the profiles, the RAW converters naturally have leeway in interpreting the RAW files, but in the end the results are unsatisfactory in my opinion. It is astonishing that Leica has obviously made a conscious decision against achieving a certain uniformity of the look between the systems. If there are third-party profiles somewhere that provide neutral and more comparable starting points between the systems for your own processing, that would be of great help to me.

[250swb]

34 minutes ago, Ralf.Lange said:

 It is astonishing that Leica has obviously made a conscious decision against achieving a certain uniformity of the look between the systems. 

Leica clearly expected people to put the 'wrong lens' onto the SL but I don't think that means they should design the entire SL system around it. 

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[wda]

Since you are scientifically minded in your analysis, why not produce your own profiles using a ColorChecker Passport? With either camera, you will achieve exactly the colours you desire, consistently.

[jaapv]

1 hour ago, Ralf.Lange said:

If I photograph the same subject, with the camera's automatic white balance, each with identical settings for the exposure, each time with the same (M) lens, once with the M10 / M10R and once with the SL2, the M image will be from both Lightroom and Capture One developed rather (too) warm and the SL2 image developed rather (too) cold. Leica has therefore opted for a completely different white balance for these two camera systems. The result of the SL2 images is a bit more sterile / digital. Even the "Pro-Profile" for the M10R advertised by Phase One leads to surprisingly unnatural, far too warm colors, at least for my taste. With the profiles, the RAW converters naturally have leeway in interpreting the RAW files, but in the end the results are unsatisfactory in my opinion. It is astonishing that Leica has obviously made a conscious decision against achieving a certain uniformity of the look between the systems. If there are third-party profiles somewhere that provide neutral and more comparable starting points between the systems for your own processing, that would be of great help to me.

Automatic colour balance is just a starting point for postprocessing and set upon raw conversion. Raw converters don't just have leeway, raw converters set the colours.
What the camera makes of it is fairly irrelevant, it is the user who determines the outcome.
If you want it to be consistent profile your cameras in your postprocessing program using the X-rite Color Checker Passport as David suggests.  It won't take more than a few minutes per camera.

https://xritephoto.com/colorchecker-passport-photo2

[elmars]

I was one of the Beta-testers of the M10-R firmware. When we started, the white balance was too cool in comparison to other M cameras. Leica chabged it for a consistance look with the M10/M10-P.

[Ralf.Lange]

Thank you for your answers. Yes, every RAW converter converts the RAW files differently, but the camera does a white balance and saves it in the RAW file. And obviously warmer values are stored with a Leica M and cooler values with an SL. I will try it once with profiles I have created myself, but from what I have read about it here in the forum, I got the impression that there are many pitfalls and possible errors here. My aim is not to determine _ the_ "correct" value, but to get comparable results with both camera systems.

[wda]

Advertisement (gone after registration)

There is no definitive 'correct' profile. You fine-tune your own calibrated profile to suit your perception. Save that as your personal profile for one particular camera. Load that when importing new files, and you have a better starting point. Remember, approximately 8% of men suffer from colour blindness. That means that one in twelve viewers on this forum see colours differently.

[jaapv]

No dangers and pitfalls here - it is as easy as falling off a log. Shoot the target under controlled lighting, I use sunlight and halogen, and drag the DNG into the Xrite software. The profile (either single or dual illuminant) will be created automatically and inserted into LR,PS or C1. If you wish, you can tweak as desired and save as a camera-specific preset.
If you do so for both cameras you will have your consistent starting point - provided the automatic colour balance does not get confused by colour casts in your subject matter...

[jaapv]

I think that you are taking the first steps on the learning curve of colour management. I would advise you to read this book .Slightly dated, and some parts can be skipped but the best introduction to the basics and principles that I know of.  Available in Kindle.

[Guest Nowhereman]

3 minutes ago, jaapv said:

... I would advise you to read this book ...

Great recommendation: I was afraid that you would recommend Margulis' book, which I would call "paint by the numbers". 😀 

Years ago, Bruce Fraser had a long discussion with Dan Margulis on their respective views. I much preferred Bruce's approach. 
________________________
Frog Leaping photobook

[jaapv]

I find Dan's books equally valuable - but on another level of comprehensibility 

[Ralf.Lange]

Thanks for your help. The “problem” probably has little to do with color profiles, but primarily with white balance. I had bought a Colorchecker card and just created a profile for the M10R and a profile for the SL2. It was really very easy. Regardless of whether these profiles are better or worse than those of Lightroom or Capture One (they do not differ very much), they do not solve my main problem: the very significant difference, how the M10R and the SL2 carry out white balance, write the corresponding values in the RAW file, which are then read by Lightroom or Capture One and lead to completely different results (not between Lightroom and Capture One, the differences are rather small, but between the cameras). Regardless of the profile used, whether Adobe Standard or Adobe Color or a profile created by myself, the SL2 and the M10R are always clearly very different in color. If I then do an automatic white balance or a white balance using a gray card within Lightroom or Capture One, the results are mostly close to each other. So my "problem" remains: Leica has apparently tuned the two cameras completely differently and the only option is evidently to match the colors by means of a subsequent white balance or subsequent matching of the colors when taking pictures with the M10R and the SL2, for example, in a coherent project.

[jaapv]

In that case set both cameras to "daylight" colour balance and keep them there. (or Tungsten when applicable)

[elmars]

Jaap gave one option. I always shoot "cloudy" like in the old film days so I can judge the light atmosphere afterwards.

The other option is to correct the white balance manually in the cooler direction. It will always be the same value. You will have to try out the right one.

[wda]

Use your greycard in a reference shot so that you can remove colour contamination with the eyedropper. Forget about WB settings, then. Just correct for a batch at a time. Easy! If you prefer your pictures warmer or cooler, just make the adjustment when tweaking your personal profiles. Where colour is important, I always take a shot with a reference greycard picture to aid post-production.

[Ralf.Lange]

David, Elmars (vom Niederrhein :-)), Jaap: thank you!

 

Ralf

www.osx.li

[farnz]

On 5/4/2021 at 10:45 AM, Nowhereman said:

Great recommendation: I was afraid that you would recommend Margulis' book, which I would call "paint by the numbers".

That's a surprising view to me since Dan Margulis's book is about working in LAB colour space as opposed to RGB colour space.  With RGB each channel is divided into 255 levels whereas in LAB each channel is divided into plus and minus 100 levels.  Personally I find working with + or - 100 much simpler than fractions of 255 but I accept that YMMV.

 

Certainly, becoming familiar with working in LAB colour space takes some effort and perseverance but it yields significant rewards, superior results, and ease of use not only in colour manipulation but also in sharpening, dodging, burning, noise reduction, and tonal manipulation because luminance information is completely separate from colour information and therefore altering one doesn't affect the other as it inevitably will in RGB.

Pete.

[Guest Nowhereman]

2 hours ago, farnz said:

That's a surprising view to me since Dan Margulis's book is about working in LAB colour space as opposed to RGB colour space.  With RGB each channel is divided into 255 levels whereas in LAB each channel is divided into plus and minus 100 levels.  Personally I find working with + or - 100 much simpler than fractions of 255 but I accept that YMMV.

Certainly, becoming familiar with working in LAB colour space takes some effort and perseverance but it yields significant rewards, superior results, and ease of use not only in colour manipulation but also in sharpening, dodging, burning, noise reduction, and tonal manipulation because luminance information is completely separate from colour information and therefore altering one doesn't affect the other as it inevitably will in RGB...

Pete - Yup, that's at the heart of it. Bruce Fraser (for RGB)  and Dan Margulis (for LAB) had a long discussion on these issues around 2008. Have a look at a post by Andrew Rodney, which you can see by scrolling down in this page. Rodney has a long quote from Fraser, that is worth reading in its entirety, beginning with: Let me make it clear that I'm not adamantly opposed to Lab workflows. If they work for you, that's great, and you should continue to use them. My concern is that Lab has been oversold, and that naive users attribute to it an objective correctness that it does not deserve...and ending with Rodney's statement: It's [LAB] been way over sold, especially if you're working with Raw files in a good Raw processor. Using those tools should (and I would love to see the Lab proponents prove otherwise) diminish the need for Lab to about nill. Its a non intuitive space to work in numerically. There's data loss and time lost going from space to space. It brings little if anything to the party that a good RGB working space can accomplish after proper use of RGB data in good Raw processor. 

Way back, in 1998, Bruce picked up on a post I had made in the CompuServe Photo forum about problems that I had trying to print B&W without color casts on the huge and hugely expensive Fujifilm Pictrography 4000 laser printer. When asked Fujifilm whether they could make a B&W profile for this printer they simply said that this was a color printer. Bruce was interested in the issues with making such a profile and offered to try to do it: quite an offer to do it for free, for his business was consulting on issues of this type. Over six months we exchanged emails and phone calls, and I sent prints to Bruce by courier. In the process, Bruce found some issues with ColorSync, on which he then corresponded with Adobe; as well as issues with some of the software that we used, although I don't recall what it was, and had me speak to the developers. The upshot was that the color cast issues could not be solved with a profile: whatever Bruce did turned out to be image specific, working on one image but still producing color casts on other images. The solution had to be for Fujifilm to build in a B&W capacity into the printer firmware.

Incidentally, I had hung a few prints from this printer on a wall in Bangkok, and two moths later they had turned green in the tropical heat and humidity — I didn't need to ask about a Wilhelm archival rating! Luckily, I managed to sell the printer, to someone who needed it for printing proofs and was not concerned with archival quality, for a couple of hundred dollars less than I bought it for.

Bruce Fraser was an extremely nice person and, sadly, died of lung cancer at a relatively young age.
________________________
Frog Leaping photobook

Edited May 9, 2021 by Nowhereman

[Adam Bonn]

It's not really Leica's job to pander to the RAW companies...

Leica's base profile (eg "embedded" in LR/ACR) follow's adobe's DNG spec, but only so far as to provide two color matrices.

These color matrices exist only to map XYZ to the camera using values for StdA ("D2800" sic) and D65. (Note that's xyz2camera not camera too XYZ)

The full adobe spec for DNG utilises the aforementioned color matrices, forward matrices, two Hue Sat Value (HSD) mapping tables (one for each illuminant) and a LUT which can add many things (hue twists, tone curves, editing robustness etc)

Oh and here's the rub - the color matrices adobe uses and the ones Leica use are not the same.

So if for example, your M records RAW data with the WB multipliers 0.3456, 1, 0.4867 (I made that up) then the different color matrices used by Leica / Adobe will produce different WB values!

Capture One meanwhile uses ICC profiles. I don't really understand DNG (at all in fact) and I know even less about ICC.

ICC typically uses the D50 colour space.

When you use an X-Rite CC you create a new set of color matrices for your specific camera.

When you use adobe's DNG Profile creator you create a new set of HSD tables.

In the Adobe DNG spec, the colour you get in LR/ACR is basically a product of the forward matrices and the HSDs.

None of the profile creation tools (bar perhaps dcamprof) creates forward matrices for you.

It sounds to me... that what you'd need to do is profile your color checker with one camera, then port that calibration over to the other camera. (instructions in the support documentation for DNG profile editor)

I have no experience of adobe premier (adobe user here but pre-CC), but I believe that this app has a facility to match to cameras colours via generating a new LUT, which is also perhaps what you want.

It's generally considered impossible to match two different cameras exactly.

As well as WB, DNG profiles etc you have other hardware constraints such as CFA, filter stack, cover glass, and even power management/stability.

The classic example here is the M240 and M262 = identical embedded profile data, very similar sensor and hardware - yet different colours.

If using a WhiBal (or whatever) and/or a set WB value in camera gets you closer to what you want then that's probably your easiest answer.

 

 

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[wda]

13 minutes ago, Adam Bonn said:

......

If using a WhiBal (or whatever) and/or a set WB value in camera gets you closer to what you want then that's probably your easiest answer. Far better, is to produce your personal profiles, for each camera, using an X-rite ColorChecker.

 

 

I have been using WhiBal cards for years, with good results. However, they seem to have been unobtainable for a long time, unfortunately.

Edited May 19, 2021 by wda