A three-CCD camera is a camera whose imaging system uses three separate charge-coupled devices (CCDs), each one taking a separate measurement of the primary colors, red, green, or blue light. Light coming into the lens is split by a trichroic prism assembly, which directs the appropriate wavelength ranges of light to their respective CCDs. The system is employed by still cameras, telecine systems, professional video cameras and some prosumer video cameras.
Compared to cameras with only one CCD, three-CCD cameras generally provide superior image quality through enhanced resolution and lower noise. By taking separate readings of red, green, and blue values for each pixel, three-CCD cameras achieve much better precision than single-CCD cameras. By contrast, almost all single-CCD cameras use a Bayer filter, which allows them to detect only one-third of the color information for each pixel. The other two-thirds must be interpolated with a demosaicing algorithm to 'fill in the gaps', resulting in a much lower effective resolution.
The combination of the three sensors can be done in the following ways:
- Composite sampling, where the three sensors are perfectly aligned to avoid any color artifact when recombining the information from the three color planes
- Pixel shifting, where the three sensors are shifted by a fraction of a pixel. After recombining the information from the three sensors, higher spatial resolution can be achieved. Pixel shifting can be horizontal only to provide higher horizontal resolution in standard resolution camera, or horizontal and vertical to provide high resolution image using standard resolution imager for example. The alignment of the three sensors can be achieved by micro mechanical movements of the sensors relative to each other.
- Arbitrary alignment, where the random alignment errors due to the optics are comparable to or larger than the pixel size.
Three-CCD cameras are generally more expensive than single-CCD cameras because they require three times as many elements to form the image detector, and because they require a precision color-separation beam-splitter optical assembly.
Some design goals for a prism assembly are:
- That the light path lengths for the three colors are the same (with correction for the different index of refraction of the glass at different colors).
- That the separation works regardless of the polarization of the incoming light; this polarization is quite challenging in practice, and there are various strategies for dealing with the resulting polarization-dependent color separation.
- That the output images are oriented the same way around (in the case of CCD image sensors). In the prism assembly illustrated above, where the red light is the direct path, the blue path is reflected once and the resultant image is not laterally inverted, unlike the red and green. In cameras using vacuum tube video pickup devices this was solved by merely reversing the connections for the line scan coils; with CMOS imagers, the row or column address sequence is simply reversed. But with CCD sensors it is necessary to build a mirror image sensor for one channel. The Philips prism assembly (center right) has all three color channels laterally inverted and can thus use three similar CCDs.
The concept of cameras using three image pickups, one for each primary color, was first developed for color photography on three glass plates in the late nineteenth century, and in the 1960s through 1980s was the dominant method to record color images in television, as other possibilities to record more than one color on the video camera tube were difficult.
Three-CCD cameras are often referred to as "three-chip" cameras; this term is actually more descriptive and inclusive, since it includes cameras that use CMOS active pixel sensors instead of CCDs. Camcorders with three chips were called "3CCD" earlier and some are still called "3MOS" (derived from 3xCMOS, Panasonic) today.
Precise alignment of the three CCDs is problematic, since a truly correct (pixel-matched) alignment would require each CCD to be positioned within an accuracy of a small fraction of the size of a single pixel. Even if such precision could be achieved at the time of manufacture, ambient temperature conditions and normal-use physical stresses would play havoc with pixel-precise alignment. This is an issue not just for translational (X,Y) positioning, but also for angular (image rotation) alignment, and for focus (distance from lens) alignment. Single-CCD cameras avoid all these issues by keeping the RGB sub-pixels physically together on the same CCD. Fortunately, human vision extracts most of its detail (luminosity) information from just one channel (green), which greatly mitigates the negative impact of three-CCD misalignment.
Dielectric mirrors can be produced as low-pass, high-pass, band-pass, or band-stop filters. In the example shown, a red and a blue mirror reflect the respective bands back, somewhat off axis. The angles are kept as small as practical to minimize polarization-dependent color effects. To reduce unwanted reflections, air-glass interfaces are minimized; the image sensors may be attached to the exit faces with an index-matched optical epoxy, sometimes with an intervening color trim filter. The Philips type prism includes an air gap with total internal reflection in one light path, while the other prism shown above does not. A typical Bayer filter single-chip image sensor absorbs at least two-thirds of the visible light with its filters, while in a three-CCD sensor the filters absorb only stray light and invisible light, and possibly a little more for color tuning, so that the three-chip sensor has better low light capabilities.
- Cliff Wootton (2005). A Practical Guide to Video and Audio Compression: From Sprockets and Rasters to Macroblocks. Elsevier. p. 137. ISBN 978-0-240-80630-3.
- "3 CCD with Pixel Shift Technology". GL2 Digital Camcorder.
- Forget 3CCD - Panasonic introduces 3MOS | News | TechRadar
- Diagram of the Philips prism and variations
- Prism-Based Color Separation for Professional Digital Photography