H. G. Dietz
http://aggregate.org/hankd/
Department of Electrical and Computer Engineering
Center for Visualization & Virtual Environments
University of Kentucky, Lexington, KY 40506-0046
Initial release: April 8, 2006
This document should be cited using something like the bibtex entry:
@techreport{lenscap20060408,
author={Henry Gordon Dietz},
title={Accessories For Exposure Calibration},
month={Apr},
year={2006},
institution={University of Kentucky},
howpublished={Aggregate.Org online technical report},
URL={http://aggregate.org/DIT/LENSCAP/}
}
The following incredibily simple observation was presented to the University of Kentucky's intellectual property committee on April 24, 2003. The committee found the idea to be patentable, but felt that the University did not contribute sufficiently to its formulation so as to justify the University having ownership rights to a patent, so all rights were turned over to the inventor -- H. G. Dietz. The idea, as presented then, is now (April 8, 2006) put forth for free use by the general public. The discussion that follows comes primarily from the original disclosure presentation.
The actual color of an object varies with lighting, but human vision compensates for lighting. Unfortunately, camera sensors do not. They simply record a sampling of the intensity of light over a range of wavelengths. The problem of compensating for lighting becomes even more complex because most digital cameras do not really distinguish intensity in each individual wavelength; instead, they use color filter arrays that do not have very sharp wavelength cutoffs and also convolve color sampling with spatial positioning.
Most digital and video cameras try to guess the color (really the intensity of the spectral components) of the ambient light. One approach reads a diffuse sample, usually of light hitting the camera, but there is no guarantee that the light hitting the camera has the same color as the light hitting the scene. Alternatively, cameras use some variant of the classic "gray world" assumption that the average color of a scene would be a neutral gray if the ambient lighting was not biased in favor of specific wavelengths... but the world often is not really gray.
The best solution is to have a manual white balance mode in which the camera can sample a neutral-color reference to determine the lighting -- making the "gray world" assumption correct. The problem is that finding a neutral reference can be surprisingly difficult. White objects are not good because they may corrupt color by saturating one or more color channels on the sensor; black objects give highly ambiguous color estimates. A gray object allows calibrating exposure as well as color. Lots of professional photographers carry 18% reflectance gray cards -- I used to -- but carrying a gray card (or multiple-color calibration card) that is big enough to focus on can be pretty awkward.
The invention is based on a very simple realization: there is no need for the reflective color reference to be big enough to focus on! As long as the blurry image is large enough for color correction sample(s) to be taken from a portion of the target region that is not in the blur radius of a color coming from outside the reference area, blur literally has no impact on the quality of color calibration. Thus, the reflective color reference can be made quite small -- for example, got lenscap?
Modifying virtually any component of the usual camera accessories, from lenscaps to carrying straps, to incorporate small color calibration targets is trivial. For years now, I have used a circle of 18% gray fitted inside my lenscap as an unfocussed calibration target. Using the Canon G1 with the lenscap target shown above, the sampling procedure looks like:
The same technique also works for multi-color calibration targets, although the complexity of a full 24-color gretagmacbeth ColorChecker probably is too high for a small pattern that is used out-of-focus. About 1/4 of a ColorChecker Mini would fit on a typical-size lenscap and, without requiring sharp focus, at least four colors can be placed on a lenscap.
Interestingly, the issue in calibration of exposure and color is not so much careful choice of which color(s) you use as targets as it is a matter of knowing what those colors are supposed to be. The happy result is that you can easily make the small color references using any of a wide range of techniques and materials. For example, I've found that many laser printers can make acceptable approximations to 18% gray cards, and a circle cut from such a printout is easily fixed inside a lenscap. (Actually, there's even a bit of controversy about what percentage reflectance gray one should be using... so feel free to make and use whatever value you like. ;-) Paints, plastics, etc. also can be used to create acceptable gray and/or color patches. Why not just make the lenscap itself out of a neutral gray plastic?
True, it takes a certain level of discipline to always use a color calibration target of any kind and perhaps only profesionals will do that, but that's not really the issue. Most cameras now do pretty well automatically determining the white balance in most lighting situations. This invention isn't a thing you'll use for every shot, but it sure is nice to have it available when the little LCD on the back of your camera displays something nasty!
An August 23, 2007 update: Photographer Stephen Johnson is now selling a 1 inch square plastic gray card suitable as a digital color reference stuck to the outside of a lenscap as the Digital GrayCapTM. The only date on his site is August 21, 2006, which is well after our page was posted and years after our invention, so we doubt his product would prevent anyone else from implementing the ideas made public in our page. He also seems to have missed the concept that one does not need to focus on the target; his site gives only sharp images of his target, although it should work just as well out of focus.
The only thing set in stone is our name.