For example, light from the sun has nearly parallel rays and light from an artificial light source doesn't.
I bet it can be done somehow, but it could be expensive.
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Whitebalance, what is good ?
- Thread starter Eleson
- Start date
I bet it can be done somehow, but it could be expensive.
Petteri Sulonen
Forum Pro
But you're not measuring the light from the sun (or the bulb), unless you're actually shooting the sun (or the bulb). You're measuring the light reflected off an object.
Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
rgmoore
Senior Member
It's worse than that, because WB is ultimately an aesthetic decision. Different photographers taking pictures under the exact same lighting might very well have different ideas about how they want the picture to turn out. One photographer might want to completely correct for any color shift caused by the lighting while another might want to leave some or all of the shift as a way of conveying the mood of the scene to the viewer. A third might want to exaggerate the color to for aesthetic reasons. None of those is a unique right answer.
--
As with all creative work, the craft must be adequate for the demands of expression. I am disturbed when I find craft relegated to inferior consideration; I believe that the euphoric involvement with subject or self is not sufficient to justify the making and display of photographic images. --Ansel Adams
poliscijustin
Veteran Member
It's pretty subjective. You could use WB to completely remove the effect of lighting, but that leaves the world looking kind of flat.
I will say, the big problem with AWB on my D70 is that it is erratic and can give unplesant casts (not as things are independent of light, nor as you remember them). It especially bogs down under mixed lighting. Fortunately, the other WB settings are pretty good.
--
A small but growing collection of my photos can be seen at
http://www.pbase.com/poliscijustin
I will say, the big problem with AWB on my D70 is that it is erratic and can give unplesant casts (not as things are independent of light, nor as you remember them). It especially bogs down under mixed lighting. Fortunately, the other WB settings are pretty good.
--
A small but growing collection of my photos can be seen at
http://www.pbase.com/poliscijustin
Okay, but! I think I've got it now =). You'd need a sensor that simply measures a lot more than visible light. Tungsten doesn't emit UV and there are probably other effects at other wavelengths. For animals that can see UV (4 primary colors), it probably doesn't look the same either.
rgmoore
Senior Member
Yes. It needs to read the photographer's mind so that it knows how he wants the scene rendered ;-)
The underlying problem is that different pictures with the same light need different WB treatment. You need to integrate information about the scene into the WB process. Then you'll be able to tell the difference between a sunset landscape and an indoor portrait and choose an appropriate WB even though the light is the same. This would be something like Nikon's color matrix metering, but a lot more sophisticated. If you were really smart, you'd even be able to recognize when you were photographing a color chart and adjust the white balance to give perfect color accuracy ;-)
--
As with all creative work, the craft must be adequate for the demands of expression. I am disturbed when I find craft relegated to inferior consideration; I believe that the euphoric involvement with subject or self is not sufficient to justify the making and display of photographic images. --Ansel Adams
Petteri Sulonen
Forum Pro
There's not a whole lot of UV in late-afternoon sunlight either. It gets filtered out by the longer trip through the atmosphere.
Point being, if there is a way to tell, it ain't gonna be easy. I have a hunch that AWB works more or less like evaluative metering -- it compares readings off the captured image to a library of readings from other images, tries to find the best match, sees what the colors should be in it, and adjusts WB based on that. IOW, it doesn't even try to look at the light; it looks at the picture instead. You know, "lots of blue on top, lots of green in the middle, a pinkish oval near the center, ah-ha, person against a background of greenery with a blue sky, must be daylight WB."
There's certainly a lot of room for improvement with that approach too, and if done well, it could work very well indeed.
Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
Auto White Balance depends on information that may or may not be in a particular image. The optimum infomration is something bright white, mid-grey and black. The reason is that the camera wants to track these values by bringing up all the color channels to equal values in the 'White'... lower all the color channelsto equal values in the 'black' and match all color channels to an equal value in the 'grey'.
But, often there is none of the purely equal values in the captured image. There might be something close like a light yellow... or an almost black Navy Blue. And, when the camera's firmware tries to 'WHITE BALANCE' these values to white and black, color casting is the result.
The tests in the review, however, should have been straightforward since a grey-scale is normally used to test color correction. This means that if the auto-white balance does NOT produce the desired result with a controlled image, the chances that it will really be off in uncontrolled settings is likely.
But, often there is none of the purely equal values in the captured image. There might be something close like a light yellow... or an almost black Navy Blue. And, when the camera's firmware tries to 'WHITE BALANCE' these values to white and black, color casting is the result.
The tests in the review, however, should have been straightforward since a grey-scale is normally used to test color correction. This means that if the auto-white balance does NOT produce the desired result with a controlled image, the chances that it will really be off in uncontrolled settings is likely.
Phil says that this is good :
http://www.dpreview.com/reviews/hp935/page11.asp
Can't argue with the man ;-)
Jean
http://www.dpreview.com/reviews/hp935/page11.asp
Can't argue with the man ;-)
Jean
Kendall Helmstetter Gelner
Forum Pro
I dont't know that anything can really deal with mixed lighting all that well, it's a really hard problem. It's why I hate to use fill-flash, because to me it almost never looks natural. I think I would rather have a camera with a pop-up 6 foot reflector than a built in flash.
--
---> Kendall
http://InsideAperture.com
http://www.pbase.com/kgelner
http://www.pbase.com/sigmasd9/user_home
It can't be exclusively scene based, otherwise this would not work :
http://www.dpreview.com/reviews/hp935/page11.asp
Same scene, different light, different AWB needed.
My guess is that the algo also look at the histograms for R, G and B and try to align the "range". I remember when HP launched the R707, they had a whitepaper on their auto-WB algo...
Jean
Petteri Sulonen
Forum Pro
It can't be exclusively based on balancing the RGB channels either, for the same reason. Same scene, so "auto-levels" should yield the same end-result.
It's probably a combination of both.
Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
It's probably a combination of both.
Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]
GiorgioPM
Leading Member
A good EVF would be the cheapest solution to this problem.
--
GiorgioPM
--
GiorgioPM
GiorgioPM
Leading Member
And this is the reason I shut a whole roll of dia to wonderful golden trees lighted in the sunset against a incredible blue-violet mountain background and then when dia where developed, the trees were not golden and the background an awful dirty brown-gray (ouch @@@): optical effect.
The film is objective, our vision subjective (W/B).
W/B is one of the most attractive feature of digital photography (if you have an EVF ..., hope soon EVILs).
Sorry I Disagree
--
GiorgioPM
GiorgioPM
Leading Member
Completely agreed.
Happy to find others waiting for EVILs.
(I'm waiting for an EVIL, to change camera)
--
GiorgioPM
GiorgioPM
Leading Member
People were used to travel riding horses for thousand of years: why invent cars?
--
GiorgioPM
William R Stirling
Leading Member
I don't know about you folks, but I had a plethora of color-correction filters in my bag back in the film days. God I love digital! ;-)
By the way, seems like there might be some techinical issues with AWB and what happens to blue channel noise when you try to balance incandescent light. Mfgs might be trying to control added noise by limiting the amount of gain applied to the blue channel. Just a thot.
Personally I never use AWB ... I leave that decision to post processing. There I can even use selective white balance if needed. Example: Image has large area of green foliage which creates green cast on subject. If I 'color balance' the entire image I lose the vibrant greens. If I just color correct the subject I can 'have my cake and eat it too'. ;-)
Just like my film days, I still shoot RAW and correct in the 'darkroom'. ;-)
Wayne Larmon wrote:
[snip]
cheers!
Rick Stirling
By the way, seems like there might be some techinical issues with AWB and what happens to blue channel noise when you try to balance incandescent light. Mfgs might be trying to control added noise by limiting the amount of gain applied to the blue channel. Just a thot.
Personally I never use AWB ... I leave that decision to post processing. There I can even use selective white balance if needed. Example: Image has large area of green foliage which creates green cast on subject. If I 'color balance' the entire image I lose the vibrant greens. If I just color correct the subject I can 'have my cake and eat it too'. ;-)
Just like my film days, I still shoot RAW and correct in the 'darkroom'. ;-)
Wayne Larmon wrote:
[snip]
[snip]
--
cheers!
Rick Stirling
The proof is in the results.
It's not how the cameras do white balance you should be paying attention to. Just compare the sample photos of many cameras and pick what you like. Of course, it's subjective from one person to another...to a point.
In reality, it's not how the buyers (most of them, at least) compare one model to another, they compare specs and the features over resulting images.
The makers make cameras the consumers buy, so, they often overlook the factors that buyers overlook.
No doubt, this post will be greeted by a lot of denials and attacks. I'm prepared to duck
See this thread.
http://forums.dpreview.com/forums/read.asp?forum=1000&message=20159771
I believe it's not only about the white balance but color spacing as well.
=====================
resellerratings.com
B&H------------1996(reviews)----9.43(6 month rating)
Adorama------276------------------5.26
Smilephoto---19--------------------2.63-------------(out of business)
http://sitekreator.com/allgoo19
It's not how the cameras do white balance you should be paying attention to. Just compare the sample photos of many cameras and pick what you like. Of course, it's subjective from one person to another...to a point.
In reality, it's not how the buyers (most of them, at least) compare one model to another, they compare specs and the features over resulting images.
The makers make cameras the consumers buy, so, they often overlook the factors that buyers overlook.
No doubt, this post will be greeted by a lot of denials and attacks. I'm prepared to duck
See this thread.
http://forums.dpreview.com/forums/read.asp?forum=1000&message=20159771
I believe it's not only about the white balance but color spacing as well.
=====================
--
resellerratings.com
B&H------------1996(reviews)----9.43(6 month rating)
Adorama------276------------------5.26
Smilephoto---19--------------------2.63-------------(out of business)
http://sitekreator.com/allgoo19
Eleson
Senior Member
Yes I can, - But I could be wrong
For me, if everyone of those test shots came out identical,
I'm not sure I would like the pictures that cameras produces.
I'm still confused, but on a higher level ...
For once, a thread I've started set of in exactly the direction I hoped for, with lots of insight and opinion.
--
Cheers
Erland
Wayne Larmon
Forum Pro
I believe that the technical problem is that there is only one color temperature where all three channels of the sensor are at maximum (around daylight?) The farther away that you are from that temperature, the more that one (or two) channels need to be attenuated, to correct color temperature. When they are attenuated enough, you get noise.
The fix, believe it of not, is to use color correction filters in front of your lens, to correct the color temperature so that it is closer to the sensor's native color temperature.
The more things change, the more they remain the same.
Wayne
Keyboard shortcuts
- f
- Forum