white balance tools

Please describe such simple experiment, because I don’t see how this can be determined experimentally.
Take shots of a colourful scene of sufficient dynamic range, including the the gray card in the scene. Shoot the scene exposing gray card differently, those are your "click grey" white balance setup shots; expose the scene for the scene itself. Apply white balances obtained from sampling grey card (using those wb setup shots) to the shot that is exposed for the scene itself.
I've done nearly the same, but just using the gray card itself.
It is not the same. You need the scene.

--
http://www.libraw.org/
 
I've done nearly the same, but just using the gray card itself.
It is not the same. You need the scene.
This is why I asked for a technical reference. I doubt any "serious" photography textbook is going to explain why white balance corrections that produced the same results on mid gray will produce different results on colors. Personally, I doubt they do.

.
 
Sorry to interject, as I haven't read all the posts, but I have an observation and a question.

First the observation:

Although I perfectly understand the need for a reference baseline to judge and adjust the colour of a shot to accurately present the resulting shot, I have often questioned the absolute adherence in anything but a studio set-up.

For instance, if you are shooting in a red light district and everything is bathed in red lights, although the objects everywhere will have their own 'colour' if illuminated properly, wouldn't the idea to be to show the red light? In such a case, white balance will throw everything massively 'out of kilter'.

I know this is an extreme and hypothetical example, but I do wonder where others draw the line.

I can see the need, though, as colour memory is appalling; amongst the worst aspect of memory. It has been demonstrated that you can show a coloured card to someone, then ask them to turn around and pick the correct one from a choice of a number of similar shades. The result is no better than random chance, so using white balance will give an unchanging reference point, but something of the feel for a moment and the conditions may well be lost. For that reason I use it with caution.

My question relates to auto white balance in different areas.

Why, if you use auto-balance in a camera, do you get a different result from hitting auto in Lightroom and auto colour in Photoshop, for example? What are the decisions they need to make that cause such wide discrepancies between calculating supposedly the same thing?

--
2011 : My new year's resolution -
To be positive, not negative.
To help, not to hinder.
To praise, not to criticise.
 
Interesting. So...what does any of that have to do with white balance? Seems like dcpTool is for correcting a profile problem. What does white balance have to do with profiles?

But more important, if there is some issue with the application of white balance based on profiles, then what would make a near-white reference any more accurate than a gray reference? It's entirely conceivable that the near-white reference might introduce a color cast in darker levels because the luminance levels are so far apart and the errors from the “twisted” profile have mounted. In such a case it would make sense to work from the middle. And once the profile is correct, we're back at the question of the difference in precision between near-white and gray, and if it's even visible.

Well, it doesn't look like this going to get answered...so I think I'll just stick with gray.

.
 
So...what does any of that have to do with white balance?
It is the answer to your "why white balance corrections that produced the same results on mid gray will produce different results on colors". One of the answers, the simplest one.

White balance has everything to do with profiles because profiles are corrections that are correct only for the certain white point while white balance sets that white point.

This also means that the more accurate is white balance the more applicable is the profile and the more accurate are colours. Precision of white balance calculations in the fixed point world depends on the absolute values of the dividend and the divider. The closer is the raw value to maximum the more precise is the value of the white balance.

Visible or not depends on the eyes of your client.

--
http://www.libraw.org/
 
White balance has everything to do with profiles because profiles are corrections that are correct only for the certain white point while white balance sets that white point.
White balance most definitely does not set the white point. White balance transforms to match the white point of the color space. As for the rest of what you said...I can't make head or tails of it other than the big-numbers-are-more-accurate part, which was already discussed.

.
 
White balance most definitely does not set the white point
You are confused. White balance is used to adapt the image to the white point of the profile.
You just said that white balance sets the white point. Now you're saying that white balance adapts to some other white point.

Confusion reigns.

So I guess the bottom line is that you have no proof whatsoever that a near-white reference makes visibly improved corrections than a gray reference.

.
 
White balance most definitely does not set the white point
You are confused. White balance is used to adapt the image to the white point of the profile.
You just said that white balance sets the white point. Now you're saying that white balance adapts to some other white point.
The shot is taken under some light, which has it's white point (Adobe are using CCT and tint approximations to define it). White balance compensates, adopts, or sets (all are used interchangeably) a new white point in the shot that matches the given profile - or the profile is calculated on the fly based on the white balance in the shot to match the white point of the shot.
Confusion reigns.
Only if you do not know the topic.

--
http://www.libraw.org/
 
Did you measure the neutral patches on the ColorChecker chart?
Yes, as I use those to profile my cameras in the field.
And did they measure as being neutral?
Not perfectly neutral.
OK, so my policy of putting a WhiBal card in the same scene as my ColorCheck Passport card is sound. Because the ColorChecker neutral patches aren't suitable for setting white balance after creating the profile. But I realize that, because I do need to precisely set exposure for the ColorChecker card in order to do a profile, that the WhiBal card is indeed grayer (and hence less accurate) than it should be.

Do you know of an alternate source of white WB targets? Now that BabelColor says that "Sales of TARGETS are suspended until further notice. "

Wayne
 
Just don't get a cold, or things will start to look blue.
A white handkerchief is all you need for perfect manual WB setting.
--

 
Do you know of an alternate source of white WB targets?
Yes, http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productid=2702
You can request sample calibration data before deciding.
Mmm, they are a bit more expensive than WiBal cards. I think I can live with the small amount of inaccuracy that I am getting with my light gray WhiBal card. Thanks, though.

BTW, if Graystar is listening, A WhiBal G7 Pocket Card is only $19.95. And they are individually measured. I can see raising your eyebrows at an $80 card, but I think that $19.95 is reasonable for a card that is individually measured.
http://mtapesdesign.com/whibal/

Wayne
 
BTW, if Graystar is listening, A WhiBal G7 Pocket Card is only $19.95. And they are individually measured. I can see raising your eyebrows at an $80 card, but I think that $19.95 is reasonable for a card that is individually measured.
Okay...I had seen the pocket sized kit but missed that they're selling the card alone. So yeah...I'd say 20 bucks is probably well spent. I wish it were a little bigger, though. It's fine for dropping in a scene, but for setting a custom white balance on a Nikon, even my 2.5" wide card is a little tight. 3"x4" would still slip into a pocket easily.

.
 

Keyboard shortcuts

Back
Top