White balance comparisons

[ellis_vener_photography]

I'm working up a review for "Professional Photogrpaher" Magazine of white

balance reference targets.

 

I have no idea how photo.net will crunch this one but I thought you might find

it interesting in terms of thinking about whithe difference betwee naccurate

white balance and pleasing color.

 

I'm going to assume that by this time you are intelligent to be workign on an

accurately calibrated and profiled display. (I use the X-Rite Eye One Pro

photospectrometer and i1 Match software for that).<div></div>

[ellis_vener_photography]

if you down load it to your desktop and open it in Photoshop, the R/G/B values in the middle of the gray section of the Custom card should read 161/161/161 or within a point or two of that.

 

Anything other than that and either photo.net or your browser or both are doing something to the file.

[Tim_Lookingbill]

Thanks for taking the time to post here on such an important subject, Ellis.

 

And as you've said in another thread on the same subject, the custom version gives a good

starting point when applying color enhancements.

 

From the looks of the custom one it appears to be a more overall balanced image

compared to the others probably making it easier applying extra saturation without

inducing weird hue shifts and uneven hot spots. I get this on my RAW shots and I've been

wanting to try WhiBAL for some time.

 

They have a video tutorial showing how spectrally neutral their white target is using a

spectrophotometer. I never new how off just plain old bright white white paper could be

until I saw that tutorial.

[david jeffery]

FWIW My RGB was 162/160/161

 

Thanks Ellis!

[john_bright]

Ellis. My monitor is calibrated by eye to suit myself and printer and I get a reading of 161/159/160 (Custom Card)in the middle. If I move the eyedropper around fractionally, I can get 161/161/161 and various other readings all very close. What does this mean?

[hakon_soreide]

It's so easy to adjust white balance, and I don't see the need to get accurate colours, so I just move the LR sliders until it looks good. A photo is never reality.

[david_hill3]

Thanks, Ellis; this looks very useful.

 

I got 161/161/161 straight away for the gray, and 207/207/207 in the white area, but in the black area there was always some difference between R and B, eg 7/6/4 or 8/7/5. Does this mean my (hardware) calibration is OK for mid-tones and highlights but a little off in the shadows?

[ellis_vener_photography]

Hakon,

 

It is more accurate to say that a photograph is its own reality. also your needs for your photos are everyone elses.

[ellis_vener_photography]

Meant to say:

 

Hakon:

 

Your needs for your pics are different as are everyone elses needs for their photographs -- clearly one size does not fit all, but it's good to always good practice to have a baseline reference to refer back to.

 

David:

 

That is a very small difference: too small to worry about it.

[mendel_leisk]

Applying Filter|Blur|Gaussian Blur with 5 radius will largely eliminate your sample variations. Having done that I get:

 

142/157/178 166/159/171

 

162/159/166 161/161/161

 

180/162/160 186/162/150

[Tim_Lookingbill]

Ellis,

 

Since WhiBaL is used as a base reference, do you know if someone has used it to come up

with and post on the web Lab or RGB number formulas to accurately depict the look of

other color temps that aren't R=G=B neutral looking like sunsets, tungsten, and any other

non6500K color temps depicted accurately in natural scenes?

[ellis_vener_photography]

Tim, wouldn't all of that vary on the atmospheric conditions at sunset?

[Tim_Lookingbill]

Yes, but because of the spectrally neutral properties of the WhiBaL target I think you'ld

probably get a close enough base starting point as well.

 

Probably eyeballing would be better. I find I do quite a bit of RGB sampling of images

found on the web rendered with pleasing nonR=G=B color temps and thought maybe

someone nailed down numbers using the WB target that are close enough.

[Carl Stone]

Ellis, that spread shows different results than what I get from Capture NX. In NX, using flash

WB always gives me the warmest colors, while in your spread, the warmest colors are from

using daylight WB. Comments?

[Carl Stone]

From the shot using custom WB, I get 162-162-162 on my iMac. It's calibrated with the

cheapo Spyder 2 Express.

[ellis_vener_photography]

Carl,

 

Possibly these things are in play in comparing my results to yours

 

1.) different lighting conditions between your tests and mine.

 

2.) Differences in raw converters

 

3.) Differences in programming of presets between Adobe and Nikon.

 

4.) differences in reference targets

As I said I am working up an article , it may be a multi parter. where these issues are addressed; different raw converters on the same files, differences in reference targets, "accuracy" vs. "pleasing rendition. and a couple of other issues. I'll try not to make it geeky and jargon laden.

[ellis_vener_photography]

Carl you posted agai nwhile I was typing: not fair!

 

But to clear up some

confusion: the numbers Photoshop tells you to have nothing to do with how your profile interprets them . The goal of display calibration and profiling is to, within the limits of a.) your display and video system's limits, b.) the limitations of the colorimeter or photospectrometer and c.) the programming of the profiling software being used , make sure that what you are seeing on the display is a faithful rendition of that data.

[Carl Stone]

It can't be too detailed for me, as I need all the help I can get! :o) I'm looking forward to the

article, Ellis.

 

Additionally, I still haven't bought a decent archiving program, and I'm leaning toward

Lightroom. It just seems to me that Lightroom to PS would be a smoother workflow than

Aperture to PS or NX. My choice is less troubling if you consider that Aperture won't work on

my old 1.6 GHz iMac. The only issue that I see is having to learn how to use Lightroom/CS 3,

to get the results that I get from Picture Project/NX.

[Carl Stone]

"make sure that what you are seeing on the display is a faithful rendition of that data."

 

Well, my print colors match my monitor using the profile from my local Costco's printer. I

can't justify the cost a good printer because I just don't print enough to warrant it.

[roger_smith4]

"Yes, but because of the spectrally neutral properties of the WhiBaL target I think you'ld probably get a close enough base starting point as well."

 

I think you're getting at a starting point created by using a WhiBal under *daylight* and keeping that same setting when shooting a sunset (as it should be predictably warm relative to daylight and you don't want it to be neutral).

 

I generally shoot with white balance set to daylight for in-camera jpegs unless I'm under mixed or tungsten lighting.