In my case, I wanted to reproduce the color appearance of the flower as it appeared at the time of the capture.
Appeared to you as you remember it? In an output referred fashion? You can’t put an accuracy metric* on that. At least one that isn’t subjective. You can say, on a scale of one to ten, ten matching as you recall the scene, this image is X number. But that again is totally subjective. I may have been there with you and disagree. I may only give it a 8. If you want to use accuracy as a useful term, we have to measure the scene color, in which case we are back to scene referred colorimetry. We can then use a non ambiguous, non subjective value to the accuracy.
I can guesstimate the length of the flower in your image below. But that’s a guess, its not accurate. Even if I guess exactly right, until you measure it with something that has a somewhat well defined accurate value (say a good tape measure, using a metric like inches or mm), there is no way either of us can say my guess is accurate or not.
Close only counts in (fill in the blank).
Then you say you want to reproduce the color appearance of the flower as it appeared. Reproduced where and how? On a display? Wide gamut or “sRGB”? Or some output device like a print? We’re going to accurately match the gamut, the contrast ratio? Probably not. Again, your idea of a match is subjective. Unless you can measure these attributes and the reproduction, how can we say its accurate? There’s no measurable metric.
IF I measure my display with a sound instrument and it tells me its 150cd/m2, we can use another reference grade device and measure it and determine the accuracy of the first measurement. Otherwise we can look at make a subjective call. But that’s not accurate.
accuracy
n 1: the quality of nearness to the truth or the true value; "he
was beginning to doubt the accuracy of his compass";
"the lawyer questioned the truth of my account" [syn: truth]
[ant: inaccuracy]
* 2: (mathematics) the number of significant figures given in a
number; "the atomic clock enabled scientists to measure
time with much greater accuracy"
Since you were not there when the capture was made, you have no idea of the appearance of the flower. You could adjust the WB to obtain what you consider most pleasing, but you could not reproduce the actual appearance of the flower. It is not subjective.
No I have no idea. But then how do you define my rendering as inaccurate? By how much? A lot? You can say you adjusted the WB to obtain an accurate rendering but
how do you prove it? I have to take your word for it. And again, since this is an output referred image, for a specific output referred device, are you quite certain that the same WB and other attributes of the image will retain that exact color and tone on other devices? And if you say yes, how do you prove it.
Accuracy is an over used buzz word. If you are going to say your rendering is accurate, not pleasing or what you believe is a visual match, I’m going to ask you to prove that your use of this WB is indeed accurate. Tell me how you’ll prove this?