The new kids on the block are the so called 4k and 2k cameras from Dalso Arri Kinetta and anyone else who wants to buy a few chips from Rockwell. They use Bayer filters to massage a colour image from a single chip.
So what is a Bayer filter?
A individual pixel or photosite reacts to photrons no matter what colour. It does not know if the photrons come from a red green blue or grey subject. So CCDs and CMOS sensors are monochromatic.
How do we get them to recognize colour? There are two methods in common operation in TV cameras. A light splitting prism splits the incoming image from the lens into three identical images. Red green and blue colour filters are applied to three surfaces of the prism, then the ccd or cmos sensor is fixed behind the filter.
So each ccd still monochromatic, now only reacts to a particular colour.
So how does a single chip sensor work?
Most single chip CCD detectors that have a checker board pattern of different colored pixels. In the long established Bayer pattern 50% of the pixels are green, 25% red, and 25% blue. Averaged together, a Bayer chip misses 67% of the color information in the image. But complex calculations from surrounding pixels are made to interpolate the missing color at each pixel. What is more difficult to construct is the image detail contained in a missing color, which can result in reduced sharpness and color artifacts in the image. Bayer pattern color cameras do not produce equivalent resolution to true RGB pixel cameras.
Sony's Genesis camera uses a striped ccd sensor. The monochromatic sensor is fitted with red green blue filters. Unlike the Bayer filter concept, where there is unequal number of red green and blue filters, the striped filter has equal number of red green and blue. So no complex algorithms are required to make up colours.
Since Bayer was not designed for motion picture production there are questions as to whether motion induced artifacts can be reduced to acceptable levels.
Dalsa uses single chip CCD bayer, the Arri D20 uses single chip CMOS Bayer