The traditional disadvantage of CMOS imagers has been that their additional complexity often reduces the percentage of each pixel actually used to collect light (the fill rate). This, in turn, reduces the inherent sensitivity of those photosites when compared to simpler implementations. Fortunately, the ability to incorporate noise reduction technology on the sensor and the use of microlenses have allowed modern designs to largely overcome this problem.
Additionally, as active pixel sensors incorporate an amplifier in each photosite on the sensor they can be sensitive to another form of noise. Small production variances can cause the gain of each amplifiers to be slightly different, which can cause visible noise patturns in the resulting image. Luckilly, this patturn is relatively predictable so many modern cameras incorporate electronics that can automatically remove this fixed patturn noise durring the readout phase. This process also provides the benefit of removing dark current noise, making sensors with this technology well suited to long exposures.
When compared to CCD-based imagers, the additional complexity in the readout mechanism used in CMOS sensors often mean that it is more difficult to maintain high frame rates. As such, many high-end CMOS imagers have taken to using a large number of parallel readout channels to compensate for this disadvantage. For instance, the Canon EOS 1D Mk.II uses a total of eight different channels to achieve its 8.5fps frame rate. While this does restore performance, it can add significant expense to the overall camera design.
The traditional disadvantage of CMOS imagers has been that their additional complexity often reduces the percentage of each pixel actually used to collect light (the fill rate). This, in turn, reduces the inherent sensitivity of those photosites when compared to simpler implementations. Fortunately, the ability to incorporate noise reduction technology on the sensor and the use of microlenses have allowed modern designs to largely overcome this problem.
Additionally, as active pixel sensors incorporate an amplifier in each photosite on the sensor they can be sensitive to another form of noise. Small production variances can cause the gain of each amplifiers to be slightly different, which can cause visible noise patturns in the resulting image. Luckilly, this patturn is relatively predictable so many modern cameras incorporate electronics that can automatically remove this fixed patturn noise durring the readout phase. This process also provides the benefit of removing dark current noise, making sensors with this technology well suited to long exposures.
When compared to CCD-based imagers, the additional complexity in the readout mechanism used in CMOS sensors often mean that it is more difficult to maintain high frame rates. As such, many high-end CMOS imagers have taken to using a large number of parallel readout channels to compensate for this disadvantage. For instance, the Canon EOS 1D Mk.II uses a total of eight different channels to achieve its 8.5fps frame rate. While this does restore performance, it can add significant expense to the overall camera design.