CCD image sensors contain hundreds of thousands (or millions, in the
case of megapixel cameras) of individual picture elements, called
pixels. Each pixel contains a light sensitive element and one
capacitor. The capacitor stores a charge that is proportional to the
amount of light incident on the pixel’s surface, which is then
transferred
to a circuit that converts the charge to a voltage and digitizes it. A
CMOS sensor is constructed from a similar array of pixels, but does not
use a capacitor to store the charge for each pixel. The rows of pixels
are activated sequentially, and the amount of light incident on the
pixel’s surface is converted to a voltage and read directly, at the
time of exposure. Of particular concern for video surveillance is
picture quality across a broad spectrum of
lighting conditions. In this aspect, CMOS has some significant weaknesses compared to CCD, as CMOS
technology tends to have inferior low light ability, does not
compensate well in strong back-lighting conditions and is prone to
excessive noise and shadows in low light conditions. To deal with these
problems, a newer technology, called ‘Wide Dynamic Range’ (WDR), has
evolved and shows great promise. A WDR camera scans the same
frame twice, once by a slow sensor, and then again by a high speed sensor. The two frames are then processed,
pixel by pixel, and result in a single frame output. WDR technology compensates for bright background and
low light conditions, producing clear, low noise images with good contrast. To date, WDR technology has not
been applied to any IP high definition megapixel cameras. Another area
of concern to video surveillance applications are video distortions
known as motion artifacts. Again, CCD performs better than CMOS in high
motion conditions due to the different type of shuttering used. Shuttering refers to the manner in which a video
camera presents light to the sensor. A CCD sensor uses a “global shutter”, which means that the entire sensor is
enabled at once, taking a snapshot of the entire frame. Each pixel’s output is stored in its capacitor and is read
by the circuitry prior to taking a snapshot of the next frame. A CMOS sensor uses a “rolling shutter”.

Due to the lack of charge storage in a CMOS sensor, the information from each pixel is read sequentially in small
strips of pixels, starting from the top and moving down to the bottom
of the array exposing only a portion of the array at a time, hence the
name, “rolling shutter”. Because a CMOS sensor reads different portions
of a frame at
different times, motion within the frame while it is being captured creates such artifacts as skew, wobble, and
in some cases, partial exposure. You may ask, what does all of this
have to do with IP cameras vs. analog cameras? The point is that the
two types of cameras are more similar in image capture technology than
they are different. However, when it comes to the method of video
transmission the differences are significant.
