At a high level, image sensors can be grouped into two general camps - charge coupled device (CCD) sensors and sensors that use complimentary metal oxide semiconductor (CMOS) technology.  Up until now, almost all consumer digital cameras have used CCD sensors. But recently, Kodak announced the EASYSHARE C513 Camera - the first EASYSHARE camera to use a CMOS image sensor (the KODAK KAC-05011, to be precise).  So we thought it might be interesting to give the lowdown on this change and the current state of CCD and CMOS technologies by posing a few questions.

What's the major difference between CCD and CMOS sensor technologies - are they really that different from each other?

CCD technology was developed specifically for imaging - converting light into electrons - so in general it provides very high image quality.  However, while CCD sensors excel in this area, that's all they can do - they need additional support chips to convert that charge into a digital signal, provide the proper pulse sequences to clock out information from pixels in the sensor array, etc.

CMOS image sensors, on the other hand, share the same base technology that is used for general electronic components, like computer processors or memory chips.  That means that a single CMOS image sensor can not only convert light into electrons, but also handle some of these other support functions all on a single piece of silicon.  Because a CMOS sensor can be "smarter" than a CCD sensor, the use of CMOS image sensors can result in a simpler camera design (because fewer parts are required).  CMOS designs also tend to run at lower voltages, which helps to reduce power consumption (leading to longer battery life).

The problem, of course, is that the image quality available from CMOS image sensors historically has not been as good as what you can get from CCDs.  So you needed to trade off these two points - image quality vs. integration and power.  That's why you've tended to see CMOS used in applications like camera phones (where small size and battery life are key) and CCD for other consumer and professional cameras (where image quality is still paramount).

Is that why Kodak hasn't used a CMOS sensor in an EASYSHARE camera before?

Exactly - the image quality available from CMOS just hasn't been good enough until now.  Customers have an expectation of the performance and image quality that needs to be available from a camera with the Kodak name on it.  This is the first time we've been able to meet those quality expectations with using a CMOS image sensor.

I thought CMOS was already being used in cameras - not only lower end products like camera phones, but in professional cameras as well.

It is - but there are some important differences.  First, the CMOS sensors used in professional cameras use pixels that are about 10x bigger than those used in consumer cameras.  All things being equal, big pixels give better image quality than small pixels. So some potential CMOS image quality issues are mitigated by using these big pixels.  Second, those CMOS sensors are generally made using specialized processes that don't allow essential support functions such as analog to digital conversion to be included on the sensor, so additional parts are required in the camera to provide these functions. And third, professional cameras tend to include significant image processing horsepower and memory - more than what is available in a consumer digital camera - that is then used to manage image quality. These three things combine to provide very high image quality, but at a price.  Kodak is aiming down a different path - providing small pixel CMOS image sensors that include essential support functions, with image quality that does not require high-end image processing.

So has CMOS finally caught up to CCD?  Is the image quality available from both technologies now basically the same?

I wouldn't say they're the same, but CMOS is certainly closing the gap quickly.  There's a lot of work that has been going on the past few years to improve the quality of CMOS pixels. And while that work certainly isn't over, it's now starting to bear fruit in consumer applications outside of camera phones, areas where image quality requirements can be even higher.

What's special about Kodak CMOS that allows it to be used for this camera?

Kodak has developed a number of different technologies that improve the image quality available from CMOS pixels.  For example, Kodak developed the use of low dark current pinned photodiodes and a four-transistor pixel architecture, which provide lower noise readout methods. 

Kodak has also developed shared pixel designs, where the transistors used to read out a single pixel are actually shared across multiple pixels.  So instead of 16 transistors used in 4 pixels (4 pixels x 4 transistors per pixel), we only need to use 7, or the equivalent of 1.75 transistors per pixel.  That gives us the same high image quality available from the 4-transistor design, but frees up a lot more space in the pixel to collect light - something that is really important as you keep making the pixel smaller.

In addition to these fundamental points, we continue to make improvements to our designs and have rolled our most recent technology into this new sensor.  When taken all together, they add up to a CMOS sensor that provides the image quality needed for this new camera.

"Most recent technology" - does that mean that the sensor in this camera uses one of the new high sensitivity patterns that Kodak announced last month? 

No, this sensor uses a standard Bayer pattern.  As we announced last month, we don't expect to provide samples of sensors using those new patterns until the first quarter of 2008.

Comments