Reading mode: Light Dark

Whitebalance, what is good ?

Go to last post
E

Eleson

Veteran Member
Messages
1,200
Reaction score
49
Location
Gothenburg, SE
This is quote from the D80 test:

"We're never that surprised to see poor automatic white balance in incandescent light but it's a pity that after all these years of innovation and development that manufacturers still can't offer a solution."

I don't dispute or disagree, but it makes me wonder still.

What do we expect from adjustments of white balance?

Do we want a color or object to be reproduced exactly the same,
independent of lightning?

Or do we want them reproduced as we remember seing them when we took the shot? ( this is what I think I prefer... )

Or somewhere in between? or something else?

--
Cheers
Erland
 
The latter, I think. That's the problem with auto white-balance: even if you managed to make one that gets it "perfect" every time, many times it would still look wrong. Perfect auto white balance would render "golden hour" photos very strangely, for example. Not to mention sunsets, candle light, or bonfires.

That said, shooting under incandescent light in AWB does result in pictures that are a dingy kind of orange. I just think it's not exactly easy to make a camera that leaves "golden hour" sunlight alone but corrects incandescent lighting all by itself. Hell, there's got to be something left for the guy (or gal) behind the camera, no?

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
 
Petteri Sulonen said:
I just think it's not exactly easy to make a camera that leaves "golden hour" sunlight
alone but corrects incandescent lighting all by itself.
Click to expand...
I think it would be easy with a sensor of some sort that can detect tungsten light, but then, which manufacturer would spend money on that?
 
Thing is, what kind of sensor could detect the difference between tungsten and golden-hour? Both are continuous spectra "white light," only heavier towards the red end. A tungsten bulb is about 2800 K, a sunset is around 2800 K, and daylight is 5200 K. IOW, "golden hour" would be just about where the tungsten bulb is. There is no objective difference between the two; it's all a matter of perception.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
 
Talking of white balance.... Any tips on how to get good white
balance when shooting through a tinted car window that
can't be opened (e.g. a safari park)??

Bill
 
This is shot that way



I think auto contrast in PS and some saturation is what is done.

--
Cheers
Erland
 
So,
I don't fully grasp how one parameter (color-temp) can be expected
to fully describe the relationship between the two?
Which, in a way is what you are saying.
And also why (most?) raw converters enable adjustment of
color temp and tint.

--
Cheers
Erland
 
--
Cheers
Erland
 
Not exactly. RGB is just one way of representing color (and in fact it's not complete -- there are colors we humans can see that cannot be expressed in any RGB space.)

Color temperature is different, and it's grounded in physics. Basically, if you take an object -- any object -- and heat it up, it emits radiation. When it's relatively cool, the radiation is invisible: infra-red. However, once it heats up to a certain point, it starts to emit visible light -- specifically, it gets "red-hot." Heat it up even more, and the light turns orange, then yellow, then white, then blue. The interesting thing is that the material doesn't matter -- any material will emit the same color of light at any given temperature. This means that when describing the shift of light from red to blue, we can express it with a single number -- the temperature, in Kelvin, of an object that would emit that color of light.

That means that for any light source where the light is produced by heating an object, we can specify the color temperature. Such objects include the Sun and incandescent light bulbs. However, there are other ways of producing light too; for example fluorescence, to which this doesn't apply.

Bottom line: no, color temperature isn't enough to describe any light, or any color, but it is enough to fully describe any light emitted by heating an object. And since both the light from the sun, even filtered through a dusty atmosphere, and the light from a tungsten bulb are this type of light, there really is no difference between the two.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
 
I am (or was when I studied) well aware of the physics behind
color temperature.
But that doesn't describe what is illuminating photo subjects,

due to filtering, reflections and, as you say, light sources that are not uniform in spectra, but rather scattered.

So to me there are two issues here:
1. We cannot fully control the color balance with the tools given to us.
2. If we can, we still have no true right answer to what the
correct balance should be in all situations.

--
Cheers
Erland
 
Petteri Sulonen said:
Thing is, what kind of sensor could detect the difference between
tungsten and golden-hour? Both are continuous spectra "white
light," only heavier towards the red end. A tungsten bulb is about
2800 K, a sunset is around 2800 K, and daylight is 5200 K. IOW,
"golden hour" would be just about where the tungsten bulb is. There
is no objective difference between the two; it's all a matter of
perception.
Click to expand...
Then a single "Indoor/Outdoor" button should resolve the ambiguity. Perhaps the "Print" button could be called into service.

Cameras need an ergonomic face lift. Photographers used to have two variables: aperture and shutter speed. Now we need to adjust aperture, shutter speed, ISO and white balance on a shot-by-shot basis. The existing control paradigm isn't making it anymore.

Maybe we'll have to wait for EVIL cameras, so we can get rid of all the anachronisms in one fell swoop. (To use an anachronistic phrase.)

Wayne
 
I don't often get a chance to pick on Wayne...
Wayne Larmon said:
Photographers used to have two variables: aperture and shutter
speed. Now we need to adjust aperture, shutter speed, ISO
and white balance on a shot-by-shot basis.
Click to expand...
No, we don't have to. We didn't in the old days. The ISO and white balance was set by the film, and we shot an entire roll with the same "settings". Heck, for 25 years I shot virtually everything at ISO 400 and daylight white balance.

Digital ALLOWS us to change those things from shot to shot, giving us a freedom that we didn't have with film. But it doesn't MAKE us change those things.
 
The OP is wondering about Mr. Askeys test on a specific camera. I admit right now that why a sensor cannot be set to take proper WB is a mystery to me. Many here can provide answers, I cannot.

What I can say is that all of my Olympus P&S cameras, and I've had four of them, all had better auto WB than my two expensive Nikon DSLR's. Since I shoot Raw, this is not as serious a problem as it could be - Nontheless, why a $300 camera had better auto WB that one that costs $5000 has always been another mystery... :)

Dave
 
Doug Pardee said:
I don't often get a chance to pick on Wayne...
Click to expand...
Probably blind luck. I go off half cocked all too often. :-(
Doug Pardee said:
Wayne Larmon said:
Photographers used to have two variables: aperture and shutter
speed. Now we need to adjust aperture, shutter speed, ISO
and white balance on a shot-by-shot basis.
Click to expand...
No, we don't have to. We didn't in the old days. The ISO and white
balance was set by the film, and we shot an entire roll with the
same "settings". Heck, for 25 years I shot virtually everything at
ISO 400 and daylight white balance.

Digital ALLOWS us to change those things from shot to shot, giving
us a freedom that we didn't have with film. But it doesn't MAKE us
change those things.
Click to expand...
I guess a poor choice of words on my part. I meant that when photographers used film, that ISO and WB was chosen when the film was chosen, and wasn't a decision that is made on a shot-by-shot basis (other than by swapping bodies that are loaded with different film stocks.) The ergonomic design of DSLRs haven't been updated to reflect the change. Aperture and shutter have dedicated control dials. Not so for ISO and WB. If the light is changing enough such that aperture and shutter speed need to be changed quickly, then ISO and WB also might need to be changed just as quickly. The current ISO and WB parameters should also appear in the viewfinder, for the same reason. (I think ISO is being addressed with newer cameras' viewfinders. Not sure about WB.)

Wayne
 
Eleson said:
I am (or was when I studied) well aware of the physics behind
color temperature.
But that doesn't describe what is illuminating photo subjects,
due to filtering, reflections and, as you say, light sources that
are not uniform in spectra, but rather scattered.
Click to expand...
That's true, but both sunlight and incandescent light are continuous spectra. IOW, they're identical.

However, white balancing actually uses two axes -- blue/red and green/magenta. The former is covered by color temperature, and the latter corrects hue casts. Together they're adequate for correcting almost all light sources, including noncontinuous spectra like fluorescent light. Not quite, of course; in fact some just plumb can't be white balanced (such as certain types of sodium light; not all though).
Eleson said:
So to me there are two issues here:
1. We cannot fully control the color balance with the tools given
to us.
Click to expand...
Actually, we can... well, almost fully, with the second axis that we have.
Eleson said:
2. If we can, we still have no true right answer to what the
correct balance should be in all situations.
Click to expand...
Actually, we do: it's simply a matter of definition. What we don't have is a single answer to which white balance is aesthetically most pleasing for any given situation. Such as, indoor tungsten lighting "should" be balanced out more than warm late-afternoon sunlight, even though objectively speaking both types of light are identical.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
 
Chato said:
The OP is wondering about Mr. Askeys test on a specific camera. I
admit right now that why a sensor cannot be set to take proper WB
is a mystery to me. Many here can provide answers, I cannot.
Click to expand...
It's not that hard a question to answer, actually. There are a number of problems, some fundamentally insoluble.

Suppose you put a separate white balance sensor on the camera. Say, a little blister of milky-white plastic, with a sensor behind it that measures the color of the light hitting it, and corrects white balance accordingly. Simple enough? Sure -- but the trouble is that it measures incident light on the camera which is not necessarily the same light that's on the subject. For example, suppose you're photographing a sunlit subject with your back to the sun. The light on the subject will be sunlight, around 5200 K. However, your camera is in the shade, which means that the light on it will have more blue (from the blue sky) in it, giving it a much higher color temperature -- say, 6500 K or even 8000 K. If the camera uses its incident sensor to white balance the picture, it'll use the wrong WB and will turn out orange.

OK, so forget that and measure the white balance off the picture instead. Now, the problem is... what exactly is white? Suppose you're taking a picture of a piece of orange fabric by daylight, or a piece of white fabric by (orange) tungsten light. Both will look the same color to the sensor -- orange. How does it know whether the color comes from the light or the subject? The answer is that it can't: AWB is a very complicated guessing game, I would imagine even more complicated than evaluative metering, and there's absolutely no conceivable way that the camera would get it right by itself every time.
Chato said:
What I can say is that all of my Olympus P&S cameras, and I've had
four of them, all had better auto WB than my two expensive Nikon
DSLR's. Since I shoot Raw, this is not as serious a problem as it
could be - Nontheless, why a $300 camera had better auto WB that
one that costs $5000 has always been another mystery... :)
Click to expand...
My guess is that dSLR's err on the side of caution and PnS's err on the side of aggressiveness. "If unsure, leave alone" says the dSLR, and "If unsure, give it your best shot" says the PnS.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
 
Petteri Sulonen said:
Chato said:
The OP is wondering about Mr. Askeys test on a specific camera. I
admit right now that why a sensor cannot be set to take proper WB
is a mystery to me. Many here can provide answers, I cannot.
Click to expand...
It's not that hard a question to answer, actually. There are a
number of problems, some fundamentally insoluble.

Suppose you put a separate white balance sensor on the camera. Say,
a little blister of milky-white plastic, with a sensor behind it
that measures the color of the light hitting it, and corrects white
balance accordingly. Simple enough? Sure -- but the trouble is that
it measures incident light on the camera which is not necessarily
the same light that's on the subject. For example, suppose you're
photographing a sunlit subject with your back to the sun. The light
on the subject will be sunlight, around 5200 K. However, your
camera is in the shade, which means that the light on it will have
more blue (from the blue sky) in it, giving it a much higher color
temperature -- say, 6500 K or even 8000 K. If the camera uses its
incident sensor to white balance the picture, it'll use the wrong
WB and will turn out orange.

OK, so forget that and measure the white balance off the picture
instead. Now, the problem is... what exactly is white? Suppose
you're taking a picture of a piece of orange fabric by daylight, or
a piece of white fabric by (orange) tungsten light. Both will look
the same color to the sensor -- orange. How does it know whether
the color comes from the light or the subject? The answer is that
it can't: AWB is a very complicated guessing game, I would imagine
even more complicated than evaluative metering, and there's
absolutely no conceivable way that the camera would get it right by
itself every time.
Petteri Sulonen said:
What I can say is that all of my Olympus P&S cameras, and I've had
four of them, all had better auto WB than my two expensive Nikon
DSLR's. Since I shoot Raw, this is not as serious a problem as it
could be - Nontheless, why a $300 camera had better auto WB that
one that costs $5000 has always been another mystery... :)
Click to expand...
My guess is that dSLR's err on the side of caution and PnS's err on
the side of aggressiveness. "If unsure, leave alone" says the dSLR,
and "If unsure, give it your best shot" says the PnS.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
Click to expand...
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top