White balance through warm lenses on digital cameras

[raphael]

Advertisement (gone after registration)

Some lenses are said to be warmer than other lenses. For instance, the Zeiss ZM line produce warmer images than my Leica or Voigtlander lenses.

I'm wondering whether and how warmer lenses affect the auto white balance on digital Leicas.

If they have an effect, I assume that

• a warmer lens results in a shot with lower Kelvin
• a cooler lens results in higher Kelvin.

That would imply that there wouldn't be much of a difference between warmer and cooler lenses on digital Leicas (regarding the warmth in the picture).

However, I can see a difference in the warmth between a ZM lens and a Leica/Voigtlander lens.
Does that mean warm lenses do not affect white balance? If so, why not? I thought the auto white balance is determined through the lens.

Currently I cannot clarify this on my own, so I'm asking you 😊

Edited February 6, 2020 by raphael

[Advertisement]

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

[a.noctilux]

AWB is what it is 😉 not perfect.

For years, I see the same behavior on different lenses of different periods, from 1930 no coating, for some of my lenses.

In my multiple M, those "AWBed" with different lenses are never "same coloring as expected", I learn to live with that as I use only DNG

which can be "later PP manually corrected in my taste".

 

Otherwise, I appreciate this choice from Leica as the lens character is fully retained.

[raphael]

Excuse me, maybe I wasn't so clear with my words. I'm not talking about AWB as a flaw in a digital M.

I was asking, how is it possible that a warmer or cooler lens still can be recognized in the final image when the AWB is working as a counterpart?

Edited February 6, 2020 by raphael

[evikne]

Not really an answer, but more a supplementary comment:

I have an old Noctilux that constantly goes towards a green tint. The AWB doesn't compensate enough for it, so I always have to move the green–magenta slider further towards magenta. To get the best result, I use an ExpoDisc or a WhiBal white balance card, but even with these tools the colors are never fully correct, obviously because this cannot be fixed with WB alone. Sometimes I think this is a charming feature of this lens, other times I convert to B&W.

[otto.f]

I can only think of one possibility and that is that the color temperature measurement and calculation from the camera uses a part of the color spectrum to determine an AWB. It might be that in calculating the AWB one cannot linearly extrapolate the measurement over the whole color range, because reality is seldom linear, only within certain boundaries and contexts. And to my knowledge computers  treat reality as if it were linear.

[pgk]

White balance will be affected by all sorts of factors such as illuminant or mix of illuminants (colour temperature is not as easy to define as simply as a kelivin reading plus tint, because many illuminants do not provide spectral outputs which can be defined in such a straightforward way), predominant colour of subject matter, and so on. And probably the 'colour' of the lens (or at least any wavelengths its absorbs). All that said, if you are shooting RAW files it may well be possible to adjust to taste or 'correct' depending on what you are thrying to do. I would suggest that the colour of a lens has a fairly minimal effect in all honesty.

[a.noctilux]

Advertisement (gone after registration)

😀 Otto and Paul got it here.

I've never thought of no linearity of "M computer".

So why AWB in mixed lighting can be so "slough" and not consistent in my many digital M.

I think that the cold/hot/green/rose - color transmitting of lens has minimal impact when we use DNG, anyway.

[farnz]

3 hours ago, raphael said:

Does that mean warm lenses do not affect white balance? If so, why not? I thought the auto white balance is determined through the lens.

No because white balance is processed from input from the tiny round window above the lens and to the left of the red dot or screw so the light doesn't travel through the lens.

Pete.

Edited February 6, 2020 by farnz

[raphael]

22 minutes ago, farnz said:

No because white balance is processed from input from the tiny round window above the lens and to the left of the red dot or screw so the light doesn't travel through the lens.

Pete.

Pete, that’s an interesting point and would explain A LOT!

But are you sure about that point?

In the manual the tiny round window is declared as a brightness sensor, not as a color balance sensor.

I read somewhere that it’s used for determining the used aperture on the lens (by comparing light in front of the body with light through the lens).

Edited February 6, 2020 by raphael

[pedaes]

1 minute ago, raphael said:

Pete, that’s an interesting point and would explain A LOT!

But are you sure with that point? In the manual the tiny round window is declared as a brightness sensor, not as a color balance sensor.

I think it measures brightness, not colour temperature . I also don't think it is linked to sensor and only adjusts camera fearures (eg frameline brightness) . But, I could be wrong and happy to learn if it does more.

[raphael]

2 minutes ago, pedaes said:

I also don't think it is linked to sensor and only adjusts camera fearures (eg frameline brightness) .

Keep in mind that also the Leica M8 has that brightness sensor, while the M8 lacks any fancy stuff like illuminated frame lines.

[raphael]

According to Sean Reid the little round brightness sensor is for estimating the aperture.

If my M10 wasn’t in Wetzlar at this time, I’d do some tests right now 😤 Like blocking or irritating the sensor etc. 😅

[pedaes]

21 minutes ago, raphael said:

Keep in mind that also the Leica M8 has that brightness sensor, while the M8 lacks any fancy stuff like illuminated frame lines.

I said eg - how about rear screen. As I said, open to correction.

Sean Reid is usually correct, so that too.

Edited February 6, 2020 by pedaes

[farnz]

2 hours ago, raphael said:

But are you sure about that point?

Hmm, not quite as sure as I was. 🤔

I agree that Leica calls it the Brightness Sensor with two functions: 1) to provide a differential versus the exposure value measured through the lens to estimate the aperture value; and 2) in the M240 and M10 to adjust the brightness of the electronic frame lines.

I clearly recall opinion from years ago that the M8's round window also measured the colour temperature of the scene but thinking about it now if a Visoflex III was used then it would block the brightness sensor window and the M8 wouldn't be able to read the colour temperature and calculate white balance.

Pete.

[adan]

On the original topic: Yes, the setting of WB in postprocessing can, to some extent, negate the natural "warmth" or "coolness" of a lens. Thus it is up to the photographer whether and how they will do their white balancing, to retain the particular color-chracter of a lens for which they paid good money.

But not entirely. Because once a lens has transmitted colors differentially (emphasizing warm or cool), the balance between different colors (which is not the same as shifting all colors together with WB) is permanently changed.

Take the extreme example, and put a red filter on a color digital camera (or a camera containing color film for that matter; in either case not really different that using a warm lens, just more extreme). Now, try to recover correct greens and blues from the resulting file. Not possible, because the red color of the filter/glass has distorted the relationship between red, green and blue.

The green-cyan color of the Mandler lenses from around 1980 (as evikne notes with his Noctilux f/1) actually removes some red from the scene - selectively. From neutrals, and reds - but not as much from actual blues (sky) and greens (grass). If you then try to make a global WB adjustment to restore red everywhere, you may get the reds back to their original color - but you are adding an unnatural amount of red to the greens and blues. A distortion.

Something similar happens if you shoot a white sheet of paper next to a yellow sheet of paper under yellow light (e.g. low-wattage indoor bulbs). It will be difficult to fully restore the difference between the original yellow and white sheets.

Similarly, faces (a pale shade of orange) shot under yellow light do not change hue as much as neutrals - and once blue is added back everywhere in setting WB, faces tend to turn too red/magenta (excess blue + red original = magenta). Thus separate tweaking of red to be more orange/less magenta is needed, via a special "yellow light" calibration profile, or by fiddling with the HSL settings per color (red and orange).

Once the transmission spectrum of a photo situation (a lens tint, a filter color, the color of the light itself) has reproduced colors differentially - that is toothpaste that often cannot be put back in the tube.

Example below. Picture made under indoor fluorescent tubes. At left, default or "Daylight" WB - whole picture is yellow-greenish. Center, simply white-balanced for the gray wall (top) - face becomes excessively sunburnt-pink. The blue needed to fix the overall WB, added to the face's less-changed red/orange, pushes it too far towards magenta. Right, face (reds) corrected to the natural tanned complexion with a "yellow-light" calibration profile that shifts reds (and only reds) back to the accurate skin-tone pale orange (without messing up the overall WB).

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!

This is one reason I got "stuck on" the 1980s Mandler lenses, back when I used purplish Velvia, but also for digital color. Their slight green tint (a different transmission spectrum) balances red, green and blue differently - and in a selective way that cannot be reproduced (or negated) by a simplistic WB of "degrees K ± a tint".

[adan]

However, I will add another point regarding "warm" vs. "cool" lenses.

Sometimes, it has nothing to do with the actual lens color.

We humans see a spectrum from violet to red, but as we know, what we actually see is proportions of blue, green and red.

Under normal circumstances, half (or more) of what we see are actually the warm colors - green, yellow, orange, red, magenta. We don't encounter a lot of blue (except for the sky*) or violets/purples.

Therefore if a lens is contrasty and saturated (or if we dial up the contrast or saturation in processing) - we will notice the saturation as an increased amount of "warmer" colors, on average. The picture looks warmer simply because the emphasized colors are warmer.

Consider the picture below, containing the visual and photograpahic primaries (blue, red, and green - and mixes)

Does one version look cooler than the other?

The only thing I changed in processing them was the global contrast (and thus saturation)

And - hmmm - Zeiss lenses are know for their - contrast and clarity.

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!

(*of note, blue is one of the last colors to be defined as a distinct color, as cultures evolve: https://www.sapiens.org/language/color-perception/)

[raphael]

Really good points, adan.

Your second part would also explain why many popular review sites claim the ZM line has pronounced blue and green tones, while others including myself find it to be warmer. At the end it may the higher saturation that we all mean.

Can someone make a side-by-side shot with a Mandler lens and a ZM lens (with their digital M)? 🙏

[250swb]

I tend to agree with farnz that the round window also reads the colour temperature.

But digital photography isn’t like balancing a warm/cool lens with a particular film and you make do with whatever the outcome is. Like Andy has shown photographers now need to have their own opinion if they want to translate the scene accurately and/or override the camera. But do it by setting your camera to AWB and have your final say in post processing.

[Advertisement]

Hallo,

Look here: Leica M

[jaapv]

13 hours ago, raphael said:

Keep in mind that also the Leica M8 has that brightness sensor, while the M8 lacks any fancy stuff like illuminated frame lines.

The M8 used the window for 

1. Estimating aperture

2. Setting preflash intensity in the M-TTL system.

I would be very surprised if that were different on the M10, with an additional function of frameline brightness.
I rather doubt whether it measures AWB, though. Other types and brands of camera, without such a window, behave the same as the M series cameras.

The only camera that I know of that had an AWB window was the Digiklux2.

It is very easily tested, though. Just hold a colour filter in front of the window. If colour balance were measured, AWB will be all over the place doing so.

 

[J.Nordvik]

Three 50mm lenses: 5cm Elmar, Jupiter 8, Zeiss Planar ZM. WB set to Daylight in LR. Auto Tone.

 

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!