The CCD/CMOS sensor battle
There’s been a tremendous explosion of digital technology in photography. Higher resolutions, more sophisticated metering and exposure capabilities and the introduction of digital SLR cameras are prompting professional photographers to make the move towards digital.
Last year, more than 100 million image sensors of all types were sold and the demand for imaging sensors is continuing to climb. Sizes of sensors vary greatly. Some pro medium and large format sensors go as high as 20 megapixels (MP). Kodak has built a line of 16MP sensors that are included in a number of leading medium format digital backs.
The highest resolution digital SLR, the Kodak 14n, has a 14MP sensor, with the Canon 1Ds being close behind at 11MP. The next level for professional SLRs is around 6MP, with a number of models from different manufacturers in that category.
All sensors initially capture their images in a continuous analog signal, through anywhere from hundreds of thousands to millions of picture element positions. Technically they’re not quite pixels at this point, since pixels are digital, but they’re still frequently referred to as such. The values of the individual picture elements of those sensors are then converted to an equivalent digital pixel value.
The CCD
The most common type of sensor is the CCD (charge-coupled device). With a CCD, light is captured with individual photo-diode sensors. The photons that strike the sensor are converted to an equal number of electrons stored at individual sensor positions. Those electrons are then read electronically and stepped off of the charge transfer register. Once off of the CCD array, they are converted to their relative digital value.
CCDs require a specialized chip construction process. Rather than having all the electronics on one chip, a separate chip set is required to handle support functions. There has been some progress made in integrating other electronics functions into the CCD, but for the most part CCD digital cameras require a considerable amount of supporting electronics. Depending upon the camera design, sets of anywhere from three to eight chips are incorporated in the camera’s image capture and conversion process.
Sensor Tricks
With most CCDs, each individual sensor position provides one pixel of digital data. But with some specialized CCDs, such as Fuji’s Super FinePix CCD, additional pixels are added to the image through the electronic conversion process. The end result is that there are more pixels in the final image than there are sensors on the CCD. Fuji has developed very sophisticated electronics to increase the image resolution beyond what the CCD can capture.
Nikon is also going with a type of interpolation to increase the effective resolution of its D1X digital SLRs. Trying to compete with the 14MP Kodak 14n, which is equipped with a Nikon lens mount, and Canon, which has the 11MP EOS 1Ds, Nikon has developed enhanced in-camera firmware and related software to intelligently interpolate the 6MP images that the D1X captures up to 10MP. With both Fuji and Nikon, the resulting image quality is very good, but they’re still forms of interpolation.
All too often, the capture resolution and the effective image resolution are used interchangeably. In most cases, the effective resolution is less than the actual CCD dimensions. That’s because the camera lens doesn’t quite cover the entire sensor, so not all pixels on the CCD are sampled. With interpolation schemes, however, the effective image resolution is higher, sometimes considerably higher, than the capture resolution.
CMOS Arrives
There’s another type of sensor besides CCD that’s becoming popular in digital cameras, and that’s the CMOS (complementary metal oxide semiconductor) sensor. Over the last few years, CMOS sensors have become increasingly common. They are being used in medium and large format digital backs, in professional digital SLRs, as well as some consumer cameras.
Both CMOS and CCD sensors are constructed from silicon. They have similar light sensitivity over the visible and near-IR spectrum. At the most basic level, both convert incident light into electronic charge by the same photo-conversion process.
However, CMOS sensors can be made of the same silicon material as other computer chips. That means all the electronics can be incorporated onto one chip, reducing production costs, space requirements and power usage. With CMOS, it’s possible to produce entire digital cameras on a single chip.
CMOS sensors also have individual picture elements, but, unlike a CCD, the conversion of the electronic signal to a digital value is completed within the individual photo sensor. That makes it possible to read-out the values of the individual sensors in a single step, rather than having to step the electronic signal off of the register, as is the case with CCDs.
CMOS sensors have been around as long as CCDs have. A type of CMOS sensor, called NMOS (n-channel metal oxide semiconductor) was used in the early 1970’s in video cameras. They worked, but image quality was marginal, at best. An unacceptably low signal-to-noise ratio has always been one of the problems with CMOS.
Last year, more than 100 million image sensors of all types were sold and the demand for imaging sensors is continuing to climb. Sizes of sensors vary greatly. Some pro medium and large format sensors go as high as 20 megapixels (MP). Kodak has built a line of 16MP sensors that are included in a number of leading medium format digital backs.
The highest resolution digital SLR, the Kodak 14n, has a 14MP sensor, with the Canon 1Ds being close behind at 11MP. The next level for professional SLRs is around 6MP, with a number of models from different manufacturers in that category.
All sensors initially capture their images in a continuous analog signal, through anywhere from hundreds of thousands to millions of picture element positions. Technically they’re not quite pixels at this point, since pixels are digital, but they’re still frequently referred to as such. The values of the individual picture elements of those sensors are then converted to an equivalent digital pixel value.
The CCD
The most common type of sensor is the CCD (charge-coupled device). With a CCD, light is captured with individual photo-diode sensors. The photons that strike the sensor are converted to an equal number of electrons stored at individual sensor positions. Those electrons are then read electronically and stepped off of the charge transfer register. Once off of the CCD array, they are converted to their relative digital value.
CCDs require a specialized chip construction process. Rather than having all the electronics on one chip, a separate chip set is required to handle support functions. There has been some progress made in integrating other electronics functions into the CCD, but for the most part CCD digital cameras require a considerable amount of supporting electronics. Depending upon the camera design, sets of anywhere from three to eight chips are incorporated in the camera’s image capture and conversion process.
Sensor Tricks
With most CCDs, each individual sensor position provides one pixel of digital data. But with some specialized CCDs, such as Fuji’s Super FinePix CCD, additional pixels are added to the image through the electronic conversion process. The end result is that there are more pixels in the final image than there are sensors on the CCD. Fuji has developed very sophisticated electronics to increase the image resolution beyond what the CCD can capture.
Nikon is also going with a type of interpolation to increase the effective resolution of its D1X digital SLRs. Trying to compete with the 14MP Kodak 14n, which is equipped with a Nikon lens mount, and Canon, which has the 11MP EOS 1Ds, Nikon has developed enhanced in-camera firmware and related software to intelligently interpolate the 6MP images that the D1X captures up to 10MP. With both Fuji and Nikon, the resulting image quality is very good, but they’re still forms of interpolation.
All too often, the capture resolution and the effective image resolution are used interchangeably. In most cases, the effective resolution is less than the actual CCD dimensions. That’s because the camera lens doesn’t quite cover the entire sensor, so not all pixels on the CCD are sampled. With interpolation schemes, however, the effective image resolution is higher, sometimes considerably higher, than the capture resolution.
CMOS Arrives
There’s another type of sensor besides CCD that’s becoming popular in digital cameras, and that’s the CMOS (complementary metal oxide semiconductor) sensor. Over the last few years, CMOS sensors have become increasingly common. They are being used in medium and large format digital backs, in professional digital SLRs, as well as some consumer cameras.
Both CMOS and CCD sensors are constructed from silicon. They have similar light sensitivity over the visible and near-IR spectrum. At the most basic level, both convert incident light into electronic charge by the same photo-conversion process.
However, CMOS sensors can be made of the same silicon material as other computer chips. That means all the electronics can be incorporated onto one chip, reducing production costs, space requirements and power usage. With CMOS, it’s possible to produce entire digital cameras on a single chip.
CMOS sensors also have individual picture elements, but, unlike a CCD, the conversion of the electronic signal to a digital value is completed within the individual photo sensor. That makes it possible to read-out the values of the individual sensors in a single step, rather than having to step the electronic signal off of the register, as is the case with CCDs.
CMOS sensors have been around as long as CCDs have. A type of CMOS sensor, called NMOS (n-channel metal oxide semiconductor) was used in the early 1970’s in video cameras. They worked, but image quality was marginal, at best. An unacceptably low signal-to-noise ratio has always been one of the problems with CMOS.
posted by eecalvin at
5:19 PM
0 Comments:
Post a Comment
<< Home