White balance

[Roger Hicks]

First, 'white' isn't a reflectivity standard: it's a light standard. And there are many varieties of 'white' light, from International White Light SA (2,848K, the sensitometric standard) to International White Light SD (7,500K, overcast American daylight).

Second, we all know that a sheet of 'white' paper looks 'white' under low-wattage tungsten light (typically 2600K) or blue sky light (10,000-20,000K).

Third, very few things reflect all wavelengths equally. The most usual 'white' lab standard for high, uniform, diffuse reflectivity is (or used to be) magnesium carbonate block: not grey cards, white paper or anything else.

Fourth, how is the picture being reproduced? On a self-luminous screen? Of what sort? On paper? To be viewed under what lighting? Remember: 'white' in a picture is not the original light: it is a reconstruction of that light. The same goes for colour: it is a reconstruction, using phosphors, LEDs, dyes, inks...

Is it not possible, therefore, that those who get excessively excited about white balance are chasing a will o' the wisp, and with the wrong tools at that?

There is no such thing as a 'correct' exposure, but a 'perfect' exposure is the one that conveys exactly what you intended. Surely the same is true of white balance. It can never be 'correct', but it can be 'perfect' if it conveys exactly the effect you want.

Any thoughts?

Cheers,

R.

[aad]

I shift white balance until the picture looks the way I like it. Most times I just leave it on auto, though.

[Joe AC]

When I shoot digital, I shoot just about everything set to cloudy. For some reasons this seems to give me bright vibrant colours similar to that of slide film. Maybe it's silly but I just look at different white balance setting just as I look at different colour film. They all render colour, but they all do it differently.

Cheers

Joe

[pgk]

Photography is first and foremost a practical application of technology. Whilst it is possible to be completely pedantic and define precisely how every step of the photographic process should be carried out (and this might be required in highly specific scientific or technical applications), at the end of the day its mostly about producing an image which satisfies. So white balance, like most other parameters is finally a subjective parameter and should IMHO be adjusted to suit the image rather than anything else.

[Pablito]

well, you can tell when someone gets it wrong!

[dave lackey]

+1 Roger... I still think we are in the dark ages of digital. The day that I can pick up a high-end digital camera (with just a lens, shutter speed, aperture and little else, and shoot it (without all the LCD, menus, white balance, etc.) like a film camera, then, I think we will have arrived at some sense of sanity.

Dear Santa, are ya listening?

[willie_901]

When a scene has mixed light sources, i.e. strong tungsten sources on the right wall and sun light streaming in from the left, there is no correct light balance. The same goes for deep shadows outdoors compared to regions of direct sunshine. An average WB setting ends up being wrong for both areas.

All you can do is modify the white balance for each region. This is relatively easy to do with digital images and difficult (using gels during printing) with film.

[Gumby]

Originally Posted by Roger Hicks

First, 'white' isn't a reflectivity standard: it's a light standard.

Calcium Carbonate is the "white" used to calibrate spectrum photometers - the ones used for scientific reflectivity measurement. Or maybe mymemory is fading (it's been a long time since I've been in a lab)... maybe it was Magnesium Oxide.

[f6andBthere]

Originally Posted by dave lackey

+1 Roger... I still think we are in the dark ages of digital. The day that I can pick up a high-end digital camera (with just a lens, shutter speed, aperture and little else, and shoot it (without all the LCD, (turn it off) menus,(don't look) white balance, (set it to cloudy as said above) etc.) like a film camera, then, I think we will have arrived at some sense of sanity.

Dear Santa, are ya listening?

Santa doesn't give a **** generally but it might be worth your while to whisper "M9!" in his ear next time you're sitting on his knee at the local shopping centre.

[Rob-F]

"Second, we all know that a sheet of 'white' paper looks 'white' under low-wattage tungsten light (typically 2600K) or blue sky light (10,000-20,000K)."

Yes, very true. My color photography professor called this "color constancy." He said that the eye and brain adjust our interpretation of an object's color based on the color of the light striking it. The paper in front of me right now looks white even though I have the incandescent lights dimmed down to a warm yellowy color.

[Rob-F]

Originally Posted by Joe AC

When I shoot digital, I shoot just about everything set to cloudy. For some reasons this seems to give me bright vibrant colours similar to that of slide film. Maybe it's silly but I just look at different white balance setting just as I look at different colour film. They all render colour, but they all do it differently.

Cheers

Joe

Yes I think that can be a good idea. At times when in doubt I have set to cloudy, and at times to shade. It's like erring on the safe side.

[wolves3012]

Originally Posted by Gumby

Calcium Carbonate is the "white" used to calibrate spectrum photometers - the ones used for scientific reflectivity measurement. Or maybe mymemory is fading (it's been a long time since I've been in a lab)... maybe it was Magnesium Oxide.

Often, spectrophotometers are calibrated to a white tile. The tile isn't an "absolute white", not least because nothing is but it does have a known deviation from true white. In other words, if my white tile is known to have 99.813% reflectance at wavelength X, I can correct that back to a genuine 100% but non-existent white tile for wavelength X. Ditto for all the calibration-point wavelengths.

[wolves3012]

Firstly, let me say that I'm not a colour expert but one who spent many years in the technical side of colour instrumentation and colour reproduction, so I have an above-average understanding of the real and human-interpreted aspects of colour.

I agree that there's no absolute and achievable correct white-balance. It could be argued that "correct" means that the final photograph (be it on paper, LCD screen or whatever) generates the exact same colours as the original scene. The technical difficulties in achieving that are not trivial, however, nor are they justified.

The human eye/brain interprets colours based on surrounding references (as well as your mood and the incident light level and quality) so anything that achieves the relative colours in the same proportion as the original should also be indistiguishable. Again, not a trivial task. Another facet of human vision is that most of it is actually monochrome and the colour is largely synthesised by the brain. I think this boils down to what looks right, to the photographers intentions, is right. Few photographs, if any, are accurate records of reality anyway, even if in some subtle way.

[j j]

Excessively excited by white balance?

[Ben Z]

Originally Posted by Roger Hicks

Second, we all know that a sheet of 'white' paper looks 'white' under low-wattage tungsten light (typically 2600K) or blue sky light (10,000-20,000K).

Our brains also have an AWB of sorts, and not everyone has the same acuity of color perception. My profession has me matching subtle color shades on a daily basis, and many times I'm working at fixing what appears to me a glaring mis-match, long after the patient says it looks perfect to him/her.

[Bike Tourist]

Back in the day (film technology) we had "indoor" and "outdoor" film.

I'm happy enough with "auto" white balance most of the time. If not, I can alter it in PS.

[j j]

Yep. AWB in combination with RAW gives you every option (auto, all presets and the cool/warm & tone sliders). Nothing to get excited about, just something to ignore or use as you wish.

[Rob-F]

Originally Posted by Gumby

Calcium Carbonate is the "white" used to calibrate spectrum photometers - the ones used for scientific reflectivity measurement. Or maybe mymemory is fading (it's been a long time since I've been in a lab)... maybe it was Magnesium Oxide.

Originally Posted by wolves3012

Often, spectrophotometers are calibrated to a white tile. The tile isn't an "absolute white", not least because nothing is but it does have a known deviation from true white. In other words, if my white tile is known to have 99.813% reflectance at wavelength X, I can correct that back to a genuine 100% but non-existent white tile for wavelength X. Ditto for all the calibration-point wavelengths.

This sounds like a fine way to establish a repeatable standard for instruments. However, I don't carry around any calcium carbonate or white tiles when I'm out shooting. So I don't think it helps with Roger's original point, as I interpret it: how do we know, as photographers, when we have judged the correct white balance in our photos?

[Sparrow]

Kodak grey card works when one has time ... sorry Roger I know you disagree.

In reality perceptive white balance is what matters ...

[JohnTF]

Roger, didn't one of the old Forum cronies invent some sort of filter that, though made originally to calibrate meters, is now sold to establish white balance? Expodisc?

Nice guy as I recall, though they were much cheaper then and looked like a bit of overhead lighting screen mounted in a filter holder, was used either as an incident metering aid, or to establish white balance?

Anyone using one?

Regards, John

[Roger Hicks]

Originally Posted by Sparrow

Kodak grey card works when one has time ... sorry Roger I know you disagree.

Dear Stewart,

No, I don't disagree at all. But so does a sheet of white paper. The big differences (apart from price) are first, that colour variations tend to be clearer with white paper than with grey cards (Kodak cards are fine but some 'grey cards' do have colour casts) and second that you can take WB readings in weaker light off paper with an 80-90% reflectance than you can off a grey card.

@ John: Yes, the ExpoDisc is back on sale, offered by the grandson of the late inventor, who meant it to be used for incident light metering with TTL meters: I have both the old and new versions. It's very good but once again it's no better than a sheet of white paper. There's also a Brno lens-cap with a white bobble on it, and some rather nasty Chinese knock-offs of the ExpoDisc where you can actually see the colour casts on the disc.

Cheers,

R.

[JohnTF]

I seem to recall in the old days we sometimes shot a photo of a gray card on the first frame for the printer to set the color.

I noticed when I started, rather late, with RA4 printing, that a lot of color film seemed to evolve toward very similar settings, when my friends with labs prior to that seemed to rely on a lot of "channels" which were set up for particular films.

They used to do a lot of testing daily to get the machines synced to the testing material they bought from Kodak.

Process became a lot more "bullet" proof and easier to dial in correction by the time I started wet color --

John

[wolves3012]

Originally Posted by Rob-F

This sounds like a fine way to establish a repeatable standard for instruments. However, I don't carry around any calcium carbonate or white tiles when I'm out shooting. So I don't think it helps with Roger's original point, as I interpret it: how do we know, as photographers, when we have judged the correct white balance in our photos?

You missed my point, which I made in another post. I wasn't suggesting carrying around a calibration tile.

Short answer: you can't, there's no absolute. If you get the result you wanted, it's correct within the constraints of reality and creativity.