Kodak, a company built on a foundation of innovations for photography and film, has struggled somewhat as the world’s transitioned toward a more digital environment. Some product lines have thrived while others have been restructured. The company has made large bets on technology, particularly efforts to push into ink-jet printing and now, a possible new breakthrough for digital photography.
Late last week Kodak unveiled a technology it claims will improve the images of digital camera’s by a factor of 2x to 4x without requiring an increase in the size of the camera’s image sensor. If the technology is as promised, is cost effective to produce, and becomes widely adopted, Kodak could be in a position to generate substantial licensing revenue. The breakthrough could, even potentially, be big enough to redirect the course of the entire digital camera marketplace.
To understand the technology at stake requires a basic understanding of how digital cameras work. So in basic terms: digital camera’s use a grid like array of sensors to simulate film and capture an image. The grid (which is a semi-conductor technology) is composed of thousands of small light sensitive sensors (called Pixels). Each Pixel recognizes a dot of light (like grain in film) as bright or dark. The computer-brain of the camera then converts that information into its memory. Repeating the pattern of dark and light dots of lights like a mosaic recreates the picture. The more pixels on the grid, the more information the camera can record; and the more information stored, the better the picture. (The same is true in film, larger format negatives captures more detailed pictures by storing more information in the form of more data about the pattern of light).
Where film and digital diverge is the point of detail each pixel can record. In a digital sensor every one of the pixels recorded is monochromatic and the technology allows them to recognize only a single color of light – red, green or blue. To compensate, digital cameras lay out the different color-sensing sensors in a standardized grid pattern and then tell the computer to estimate what may have been left out and fill in the gaps to recreate a seamless picture.
As a crude example, imagine drawing a picture in color. Then take a black pen and make a thousand dots on the drawing you just made. Each black dot represents information that the sensor didn’t’ record because the pixel underlying it wasn’t sensitive to the color (wavelength of light)you originally used. Film would have recorded all of the information (e.g. all of your picture). A digital sensor could only record part of it. A computer algorithm is responsible for guessing what was left out and connecting the dots appropriately (a process called interpolation).
What Kodak unveiled Thursday was a new method of laying out a sensor grid. It includes the usual red, green and blue specific sensors but in addition it also includes panchromatic, or "clear" pixel sensors that are sensitive to all wavelengths (colors) of visible light. The process, which relies heavily on new algorithms as well, means a camera will record more actual information. (In the example of making a thousand dots on a drawing, you might now only make 600 dots). It’s a breakthrough which will dramatically change the quality/quantity of data recorded on any camera using the technology. A camera using this technology that has a sensor capable of recording 5 Mega-pixels of light information might be able to record the same amount of information as a camera that requires a 10 megapixel sensor today without the same innovation.
The benefit of the improvement is substantial. Pictures will be possible in lower light situations. Blurry pictures will be less likely. More data being captured in a smaller area (smaller array of sensors) means smaller camera’s can yield higher quality images then presently possible. Further, smaller sensor arrays will require less power to run meaning less battery capacity is needed to power the image-sensing hardware of the camera. These combined factors could result in dramatically higher quality resolution for cameras in smaller-footprint portable devices, like cell phones, whose function limits their size (and the ability of engineers to incorporate larger sensor arrays).
Thanks to Kodak’s innovations, smaller batteries and smaller sensor areas can still yield high resolution – both freeing product designers to further miniaturize and to innovate with features. Kodak’s invention could also mean higher quality professional equipment won’t have to keep growing in size to increase in resolution. (For camera’s to increase in resolution either the size of the individual sensors have to decrease, or the size of the grid/array of sensors has to increase).
Kodak has filed at least 15 patents on the project. It’s worth noting, however, that the Kodak development, while deservingly heralded as a breakthrough, is not the first effort to make a digital sensor receptive to all colors at the same time, or capture more data.
Another Silicon Valley company called Foveon developed and brought to market a competing technology a few years ago which has not been widely adopted (arguably do to cost). In Foveon’s case, they recognized that light penetrates silicon (the material in a camera’s sensor) at different depths depending on the color of the light (due to different colors of light having different wavelengths) . To capture images Foveon built a camera sensor by layering three image sensors on top of each other in the space of a single pixel.
If a beam of light hits a Foveon X3 sensor, the deepest layer will be triggered by only one color (the other two colors not penetrating the top two layers of silicon). The middle layer, in contrast, might be hit with two colors of light and so on. Substracting from what level does, or does not, get triggered, it’s easy for the computer to extract which color should be recorded.
Using this technology, Foveon sensors provides true color data recognition at each pixel. Unlike Kodak’s approach which expands and improves on the more common interpolation approach to creating a picture, Foveon’s 3 layer semi-conductor image sensor, and the signal processing necessary to interpret the data it records, is far more complex method. With triple the number of sensors, it is also much more expensive to build. Additionally, to my knowledge, the Foveon sensor technology only works on CCD based image sensors (CCD and CMOS are the two types technology used for camera image sensors). The Kodak technology will work on either type of image processing/sensing technology, expanding its reach.
Foveon’s technology hasn’t been widely adopted by camera manufacturers. Kodak’s breakthrough might be. Kodak has the money, means and industry relationships to better bring to market their innovation. More importantly, their breakthrough isn’t likely to add substantial cost to the camera. Further, if it is correct that Kodak’s technology applies to both forms of image sensing technologies, and Foveon’s does not, Kodak stands to be in a position to serve the needs of both high end, and low end, manufacturers.
The first application of the new Kodak technology is expected around first or second quarter 2008. That means, by summer or fall 2008, it could start to find its way into the consumer marketplace.
That’s still a long ways off, but mixed with efforts in ink-jet printing, it could be just the turn-around Kodak’s been needing.