Photosites and pixels

Each sensor has arrays of hundreds of thousands of "photosites" that detect light. More photosites generally translate to more pixels of resolution. Because each pixel can contain several photosites (e.g., to detect red/green/blue wavelengths), and there is different technology for sensors and photosites, the number of photosites does not necessarily correspond to the number of pixels.

Nevertheless, camera manufacturers usually specify the resolution of the cameras based on the number of pixels that the sensor generates (usually in a native "raw" format). Most DSLRs can store images in different resolutions. For example, Nikon DSLRs enable users to specify small and large JPEGs. When saving native raw files, however, most DSLRs save the full resolution emitted by the sensor. An exception is if the camera is a "full-frame" DSLR that has a "cropped frame" mode. For example, the Nikon D3 has a native FX mode and a DX mode to emulate older Nikon DSLRs. (We'll discuss full frame and cropped frame next time.)

There are numerous advantages to having more pixels. Generally, the more pixels available, the larger a printed image can be without degradation in quality (e.g,. sharpness). This is only really a problem if you are planning to print larger than poster size. Very few prosumers will do that, but it could be nice to know that you can do that.

There are also drawbacks to having too many pixels. All things being equal, with higher pixel density (i.e., pixels within some defined area) generally comes more noise. The files stored by the camera are much larger. When postprocessing files, larger files take much longer to postprocess.

While pixel count should be a consideration when selecting your first or next DSLR, it need not be your primary consideration. Most modern DSLRs have more than 10 or more megapixels (10 million pixels), which is plenty for printing a poster.

Links:

1) Wikipedia article on Foveon X3 Sensor used in some Sigma digital cameras