M10 color rendition compared to M9-M240-SL cameras

[otto.f]

Advertisement (gone after registration)

XMA and ZGL seem to me the M10 captures (on my old iPad mini somewhere around the world)

Edited April 17, 2017 by otto.f

[adan]

I'm waiting to see if anyone else wants to contribute to my little test before revealing all. But in the meantime, a few thoughts on "correct color."

 

@ Mitch: yes, I am on an iPhone from out of the bush and something went wrong editing... I do feel that the differences are caused by the dyes in the Bayer filter and can be made to be so small in real life photography as to be insignificant in anything but product shots. -Where metamerism will rear its ugly head...

 

The Bayer dyes are certainly a factor - but there is a lot that goes on between the electrons sliding off the sensor and the final color in a picture, or even the final digital values per R/G/B pixel recorded in the raw file on the SD card. At a minimum - there are the algorithms used to convert an analog output from a sensor pixel (raw number of electrons) to a digital value (00101101 10010110). And I use the plural "algorithms" advisedly because it is not as though there is a single universally-used "scientifically-correct" equation (an "E=MC²") that all manufacturers use by default.

 

Beyond that, there is the "predetermined" calibration profile, from which most of us have to start. It may be Leica's "embedded" profile, or Adobe's, or CaptureOne's... but we are handed a "black box" that has already made decisions about how magenta or yellow "red" is, or how cyan or yellow "green" is. And as we know, such canned profiles often produce very different colors. On top of which we can add our own adjustments in the calibration pane of our raw developer.

 

All of which has its analogs in creating color films. The film engineers make creative decisions about how their film will reproduce reds, or skin "oranges" or foliage greens, by playing with both the sensitizing dyes and the final image dyes. Which is why every film has its own palette.

________________

 

From a different angle, my "Color Photography 101" instructor impressed on us that there are many ways of defining "correct color."

 

- There is scientifically-correct color: the wavelengths of light produced by the computer monitor, or reflected from a hard-copy's surface, exactly match the wavelengths of light that originally reflected frm the subject. Virtually unachievable in the real world, although the Lippmann Plate process came close, theoretically. If the original color that hit a spot on the plate was 513 nm, the plate would use constructive and destructive interference to ensure that only light of wavelength 513 nm was reflected from the processed plate at that spot. https://en.wikipedia.org/wiki/Lippmann_plate.

 

(BTW, the same effects that produce purples, blues, greens, reds and yellows on the surfaces of coated lenses, from colorless coatings).

 

- There is technically-correct color, which gets as close as possible to scientifically correct as current technology allows. The gross overall color of a photograph of a patch on a Macbeth Colorchecker (or the blue of a dress, or the green of a John Deere Tractor) matches the original, when placed side by side. Remember the internet meme of the gold/white - blue/black dress? https://images-production.global.ssl.fastly.net/uploads/posts/image/52649/the-dress-white-gold-blue-black.jpg

 

- There is visually-correct color: things look to us like the colors we saw or remember. We can identify a car (or person or whatever) in the real world from a color photograph of the car/person/whatever.

 

- there is emotionally-correct color: the colors bring back the experience we felt viewing the original place or event. OR change the colors to convey a different feeling.

 

A color that "pops" is quite likely NOT an accurate rendition of the original scene color. But it may appeal to us more, emotionally.

__________________

 

It is always useful to remember that the "color" we see doesn't really exist until processed by the brain. Something is not "yellow" until we decide, by processing signals from the retina in the visual cortex, that it matches up with our stored neural definition of yellow. Prior to entering the eye, it is just a certain band of wavelengths (or combinations of bands) on the EM spectrum.

 

Who is to say that humans' perception or opinion of color is more trustworthy that, say, bees'? Bees have been around a lot longer (if you favor tradition) and there are billions more of them (if you believe the majority is always right), and their vision covers a wider range of the EM spectrum than what we see as light.

 

http://whyfiles.org/2012/bee-vision/

[Advertisement]

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

[Guest]

Adan - I would guess that XMA and ZGL and TPC are from the M10. 

 

I know that Charles Peterson is working on achieving M9-type color from the M10: he is working on a project that he started with the M9 that he wants to complete with the M10. Would be good to get his input as well. 

 

Sorry, in a post above I called you "Andy Petite" instead of Andy Piper — the former was a classmate many moons ago.

_______________

Alone in Bangkok essay on BURN Magazine

Edited April 17, 2017 by Guest

[erudolph]

The colour rendition of the M9, M (Typ 240), M10, and the SL is something I will cover in a future issue of LFI. Stay tuned …

Is this issue out yet?

[bcdano]

My guess would be that the files XMA / ZGL / TPC are from the M9. 

[elmars]

Yes, part 1 in the current issue.

[otto.f]

Advertisement (gone after registration)

XMA and ZGL seem to me the M10 captures (on my old iPad mini somewhere around the world)

and AFD

[jonoslack]

First pair - two shots, each made with one of the cameras. Identical processing in Adobe Camera Raw, which is to say, none, except for the basic profile used (which is fixed). The M10 image used the embedded "Leica M10" profile (camera FW 1.0.2.0). The M9 image used the "Embedded" M9 profile (camera FW 1.196). All the other calibration hue/saturation settings, and all exposure settings - for this test - were zeroed out. WB for both pictures is the same - 5350 K, zero tint green/magenta.

 

 

FTCcolor.jpgXMAcolor.jpg

HI There

Just a thought - shouldn't you be using the LR Standard profile for the M10 rather than the embedded profile? It's certainly what both Adobe and Leica intend (and it's very different - much more subdued and less saturated)

[adan]

HI There

Just a thought - shouldn't you be using the LR Standard profile for the M10 rather than the embedded profile? It's certainly what both Adobe and Leica intend (and it's very different - much more subdued and less saturated)

 

Since I have no interest in "renting" Photoshop via the Cloud (and eschew LR - I need full Photoshop), and thus stick with the last DVD version (CS6), my Camera Raw is now frozen at v. 8.3. Can't get new camera updates from Adobe, that I know of. Which is fine, so long as Leica sticks with the .dng format.

 

Besides which, "much more subdued and less saturated" is exactly NOT what M10 users trying to match M9 colors want. We want the rich, saturated CCD M9 colors, as close as possible.

 

And I'm puzzled by the idea that Leica would do a lot of work producing their own embedded profile - and then intend people to ignore it and use Adobe's. Necessarily. I've never found Adobe's idea of good color to match mine - and since my colors win national awards, I'll stick with my own opinion.

______________

 

On that subject, in the process of making these comparisons, I've come to the conclusion (always subject to adjustment) that Leica's M10 embedded profile is the best "stock" manufacturer's camera profile I've ever run across. It already does most of the things I tweak for in a profile. I'm probably going to remove all my own calibration adjustments, and just use the out-of-the-box "Leica M10" profile as my default setting for a while. (Except under yellow indoor light - that will always need tweaking to avoid magenta reds and purple blues once the yellow cast has been removed by white-balancing).

 

I would say the biggest color drawback in processing M10 files is the horrible stock WB. "As shot" comes out pink as H*ll, and "Auto" is almost always fooled one way or another.

 

I've picked a "standard sunlight" WB of 4900 with zero tint for the moment, as my "all images" default, which I can tweak as desired per image or lighting. Below is a sample of why I do that (and with the default, unadjusted "Leica M10" calibration).

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

To get the closest match to M9 colors (or even better, in the case of oranges, as olgierdc pointed out), I simply ramp up the overall saturation in raw developing, to about +20 (+10 was my default M9 setting). And/or, if desired, add contrast via either black and white points, or the general contrast control. Depends on how much a given picture needs the full extra DR of the M10, which I will happily sacrifice - when appropriate - for punchier color.

_________________

 

as to the coded examples, the reality is:

 

M10 pictures - FTC, AFD, and QFT

Edited April 18, 2017 by adan

[olgierdc]

I agree in 100% with Adan.

As a starting point I would choose FLASH (in Lightroom) because it has a temperature of 5500K and has no tint as opposed to DAYLIGHT with a magenta tint +10.

Blue sky looks better without tint. For me 4900K i little to warm. Everything depends on where we live (angle of sunlight).

Edited April 18, 2017 by olgierdc

[cp995]

...

 

as to the coded examples, the reality is:

 

M10 pictures - FTC, AFD, and QFT

 

The reality I hoped it ist not ...

Thanks a lot for your work and explanations!

Edited April 18, 2017 by cp995

[Guest]

I haven't reached any conclusions yet because I've been too busy to shoot much in the last three weeks, but so far I like starting with the LEICA M10 profile. Of course, the subsequent WB, BP, WP or contrast or clarity adjustments can overpower the choice of camera profile.

 

Don't laugh, but what I've been doing for daylight shots is, neutralizing the WB if there is a gray that I want in the picture, and then applying from the VSCO FILM 06 Push and Pull package either the [L - Portra 800⁺¹ -] or the [L - Portra 400⁺¹ - ], and then using that as the starting point to make adjustments. (The VSCO presets apply the LEICA M10 profile).

 

The picture below uses the  [L - Portra 800⁺¹ -] profile; it was shot at ISO 1600 to test using the DR Summicron near focusing range with the EVF 020. This is at the closest focus of the lens, 19 inches (47.8 cm), giving the minimum field of 7x10 inches (17.78 cm x 24.5 cm) — a 1:7.4 magnification ratio — not a macro lens but nevertheless useful.

_______________

Alone in Bangkok essay on BURN Magazine

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited April 18, 2017 by Guest

[jonoslack]

 

Besides which, "much more subdued and less saturated" is exactly NOT what M10 users trying to match M9 colors want. We want the rich, saturated CCD M9 colors, as close as possible.

 

 

HI Andy

Of course - I didn't mean 'you should' but 'should you?'

Personally I shoot daylight WB outside, and a custom WB inside, and I quite agree about the dreadful pinkness of the Adobe AWB . . . does anyone use it?

But I'm not hunting for M9-ness, I'm looking for good skin tones and a gentle starting point - so I like the LR profile . . but like Sean Reid I do remove the lens and moire corrections first. 

 

Colour is such a personal thing!

[Chaemono]

Love what Andy does with the colors on the M10. But with the SL I sometimes prefer the LR profile because of the subdued and less saturated colors. I actually always prefer it with the Summilux-SL where that cooler look seems to work particularly well. Skin tones are nice, too.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[adan]

...and since Chaemono raised the subject, a word on the COLOR of Leica lenses, and how that may affect one's white-balancing with the M10 (or any Leica). A test I was already doing today anyway.

 

Below are samples of a gray card, shot with 6 different Leica lenses from various eras, with the M10.

 

Procedure - I opened the 1956 Summicron image first, and white-balanced that with the raw developer's sampling eyedropper - its gray is numerically neutral at RGB 111 ± 2.

 

Then I opened each of the other images with "same as previous" settings - so that they got the exact same WB (which was 7350K and tint of +3 (pink)). Not necessarily "correct," just identical for all 6 samples. Lighting was outdoors, cross-lit sunlight under blue sky with scattered clouds - same for all shots.

 

One or another - or none - may look "neutral gray" to you, but I was mostly interested in the relative colors - which lenses are more pink (or less green) or more blue (or less yellow), etc. when the WB is identical. Which means the "ideal" daylight white balance for you will depend on which era(s) of lenses you use - and care must be taken choosing the lens to shoot one's calibration pictures.

 

It somewhat confirms my long-held (but never really tested) observation that Dr. Mandler's Leitz-Canada lens designs tend to run slightly green/yellower than the modern Leica Germany lenses (or conversely, the modern Leica lenses are more pinkish), and that earlier Wetzlar lenses tended a bit blue. And that my beloved 1980 21 Elmarit is the greenest of all. Interestingly, the 135 TE - supposedly a Mandler design but never made in Canada - does seem to "group" with his other designs (21 1982 and 35 1980) even though 13 years earlier.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

As a side note - here is an interesting little table from Leica reviewer and commentator Erwin Puts, published in a .pdf monograph titled Leica M-Lenses: Their Soul and Secrets. It is from 2002, and thus does not include Leica's most recent new lenses (where he refers to the 28 Elmarits, the "current" one is the one before the ASPH, and the "previous" one is the 1980 Mandler Canadian design).

 

The chart lists the blue, green and red values for each of the listed lenses, normalized for the blue value at zero. The values give the ISO Color Contribution Index for each lens, where the standard for "neutral" is 0/5/4. The color markings around the edge are my additions.

 

And indeed, the Mandler c. 1980 designs (green or yellow M's) run high in green and/or low in red, except for the Noctilux f/1, which is high in both and thus "yellower" than the standard. While the recent Solms designs tend (just slightly) to match red and green values more closely. The APO-Telyt 135, 50 Summicron (still current) and last pre-ASPH 28mm Elmarit nail the neutral "standard" exactly. (FWIW)

 

 

As for me - I now have confirmed that if I want my 90 Summarit to "match" my 21 Elmarit for color, I need to give those pictures a shot of green tint in the WB, and the old 50 DR needs an extra 200°K of yellow, plus some green.

[01af]

... but they both felt essentially that, if you fixed the colors you wanted (say, skin tones), other colors would go out of whack.

Yes ... if you suck at fixing the colours.

 

A digital camera's colour rendition can be anything you want. All you need is a proper camera profile. Granted—the profiles that come with digital Leicas are rotten. But it's no rocket science—albeit somewhat tedious—to make a better one.

[dsalamena]

Talking of colours and white balance, I've noticed that changing the camera profile in Lightroom from Adobe standard to M10 the values of "as shot" white balance change as well. Usually they tend to be cooler with the M10 profile and wormer with Adobe standard.

I was quite surprised as I thought that "as shot" white balance values would be always the same and not be affected by camera profiles. 

Is that normal?

[adan]

Yes ... if you suck at fixing the colours.

 

A digital camera's colour rendition can be anything you want....

 

Not always. There existed NO color profile that would correct the purple blacks and grays of the M8 due to the IR contamination. No way to fix a gray tuxedo that turned purple, without screwing up the (actually purple) tuxedo right beside it.

 

The M240 red problem was analogous - it stained some colors (like the dark architectural-green light pole in the M10 crop below - correctly rendered) so badly that that color became gray or brown.

 

Outside of tinting the white balance heavily green (and making everything in the picture too green), there exists no hue or saturation slider that can turn a brown or gray light pole back to the correct original green color. Gray has no hue or saturation to adjust. And if you try to swing a brown hue (a dark red or yellow or orange) back to green, you make every other red/orange/yellow in the picture too green (as well as doing weird things to other hues).

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Chaemono]

It somewhat confirms my long-held (but never really tested) observation that Dr. Mandler's Leitz-Canada lens designs tend to run slightly green/yellower than the modern Leica Germany lenses (or conversely, the modern Leica lenses are more pinkish),...

"pinkish" - nicely put, Andy. Apo 50 comes to mind. Will try to do some comparisons with the SL50 on the SL this weekend. 

Edited April 21, 2017 by Chaemono

[Advertisement]

Hallo,

Look here: Leica M

[adan]

Talking of colours and white balance, I've noticed that changing the camera profile in Lightroom from Adobe standard to M10 the values of "as shot" white balance change as well. Usually they tend to be cooler with the M10 profile and wormer with Adobe standard.

I was quite surprised as I thought that "as shot" white balance values would be always the same and not be affected by camera profiles. 

Is that normal?

 

First, there are firmware/software mavens who will know a lot more about how camera profiles are really constructed, and whether they may contain tweaks as to what "white" really is (it can be anywhere from 5000°K to 6500°K) as well as adjustments to saturated colors. But I bet they are a lot more complex, with more moving parts, than the simple seven sliders LR or Adobe camera raw give us.

 

But consider the following - software engineer A is constructing a profile for the M10, and in his opinion (or by the standards of his company, and/or the known preferences of its customers) the reds should be more orangey red than straight from the camera, and greens should be more yellow-green than straight from the camera. Whereas firmware engineer L is constructing an embedded profile for the M10, and in her opinion (or by the standards of her company, and/or the known preferences of its customers) the reds are just fine as is, and the greens should be bluer (more cyan).

 

Obviously, for a given, fixed WB, pictures processed with profile A will have more net yellow (anything at all containing even a hint of red or green - 2/3rds of the color palette), and the same shot processed with profile L will show less net yellow. Which to me would translate as looking warmer (A) or cooler (L) even if the actual white setting (say, 5500°K, tint +3 magenta "as shot") and the yellowness/blueness of neutral grays/blacks/whites, is identical.

Edited April 22, 2017 by adan