☉ Why does the Auto White Balance (AWB) of the most advanced camera shoot the photos with wrong colors from time to time?
Except for a
few of the early advanced SLR cameras equipped with an independent
unshaded external white filter for its AWB, such as Nikon D2X,
etc., all the cameras, including DSLR camera, EVIL camera, digital
cameras, video recorders, smartphones, tablet PC and computer video
cam, use the logic calculus AWB (Auto White Balance) to
deduce the camera white balance value instead of getting the actual
white balance value of the framed scene with a real optical filter
to do this job.
Every camera / mobile phone manufacturer has its own unique formula and weighted white color balance calculation methods under the unopened black case operation. That leads to very different color distortion photo at different scenes. Take all the latest camera models of all brands to aim at the same scene and shoot at the same time, you would notice that the colors of each of these captured photos are quite different from each other while the real life overall color of the scene is one only under that particular lighting. Thus, you can easily see how serious calculation errors may occur in the the AWB (Auto White Balance) systems of each camera.
The basic concept of white balance is that white can be restored to white no matter under what light the sources are. The function of the so-called AWB (Auto White Balance) is to restore all the colors under color cast lighting to the colors under the standard illumination automatically. International Commission on Illumination CIE defined the standard illumination is D65 light source with the color temperature of 6500K. The Color Calibration restoration definition is to restore for true color of the rendering scene. Color is the color temperature of the light itself. While the purpose of camera AWB is to correct all the color cast caused by the light source. This is the main reason many cameras capture the photo images in wrong tone of noon during the evening.
There are several camera AWB (Auto White Balance) simulation algorithms:
Grayscale World Assumption
Grayscale World Theory: according to this theory when there is enough color change with one photo, its R, G, B value of all components tends to a status of balance ( even if the numerical value of the three: R, G, B is equal, in other words should be the type of color: black, white of gray.) As regards the white balance, this theory has a wide range of application. One distinguishing feature is that it can use much more image information to make a judgment but when it is used with an image that is composed by one color (single color image), there are some problems that may appear.
Standard Deviation Gray World Assumption
Standard Deviation Gray World Assumption is an algorithm directed towards improving the above-mentioned assumption. It works on the principle of dividing one image in several parts (blocks), then performs a statistical analysis on each part (block) of the image, looks inside its color richness, more proportions on the weighted color while less proportions on fewer colors to reach a final average summation. Then, the RGB value would be modified according to the relatively accurate summation.
Perfect Reflector Assumption
Perfect Reflector Assumption: The point of the maximum brightness of the image is the white point. It assumes that the biggest Y value is the white color of the YCbCr space and uses it to revise the colors of the entire image. It is characterized by considering only that part of the brightest colors and is completed on the contrary to the above Gray World Assumption. This assumption might achieve better results when dealing with particular monotone color image. However, it may have the color cast or overkill circumstances when faced with the color rich pictures since the brightest point is not necessarily the white color.
All the known or undocumented camera AWB logic calculus would encounter occasional or frequent computing errors to shoot out with color inaccuracy photos due to environmental complexity or its logical calculus imperfection. They are extremely unreliable to shoot for a color correct photo. Of course, every camera manufacturers would claim that they have made significant improvements towards colors and AWB on their latest model of the cameral or video CAM when it was announced. The truth is that this kind of claims have been the same over the last two decades. In spite of improvements through camera from one generation to the next generation, its AWB function is still unreliable up to this day. Until the camera manufacturers adopt our patented technology of “built-in or external white balance filter device” in the next generation of new camera, this color inaccuracy problem on every new camera AWB phenomena would be continued.
Until then, “RAYNAULD™ Directivity Color Calibration Filter” is the best and only solution to shoot for an almost perfect color photo anytime anywhere !