Digital Video: How It Works

Anything digital in the CCTV market is popular now, especially digital image recording products. In fact, nearly half the respondents to a 1997 SECURITY Magazine survey indicated that they view digital storage as a strong emerging technology. More than one-fourth of respondents to a 1999 SECURITY Magazine survey said they plan to purchase a digital image storage system within the next 12 months. Why all the attention?

Digital recording offers clear advantages, including better-quality images, quick searches, continuous recording during playback, and the need for less physical storage space for the media.

The buzz at security trade shows recently has been digital video image quality. There are claims of high-resolution- even broadcast-quality images with extremely small storage sizes. Though digital technology is quickly changing the way we gather and process images for security purposes, and though the quality claims sound appealing, how do you separate from fiction?

Where Do Digital Images Come from?

An image starts as a group of light rays coming into a camera. The light rays meet a charge coupled device (CCD) element in the camera - a flat spread of sensors that convert the light into charges and store the information. These sensors are, on average, 640 picture elements wide by 480 elements deep.

The stored charges contain both color and black-and-white data (a composite image). To assemble an image you can see, a digital video decoder must arrange the charge information into primary color channels - red, green, and blue (RGB) - similar to what a television would do.

Once the image is decoded into three channels, it is then converted by an analog-to-digital converter (A/D) in your digital video recorder into eight bit samples per color channel. Three color channels times eight bits per channel equals 24bits. Each set is referred to as a pixel (picture element), and represents one dot painted on a color monitor screen (also called a cathode ray tube or CRT).

In summary, cameras create composite images, and digital recorders decode these images into three color channels, and then digitize the image for storage on a computer hard drive or transmittal through a computer network.

Analog vs. Digital Images

Once the image has made the transformation from analog to digital, many options are available that are not available for an analog image. Analog images only can be displayed on a monitor, recorded to tape, or perhaps printed on a video printer. Digital images may be sent anywhere in the world in the blink of an eye. Via worldwide networks such as e-mail or interoffice Intranets, an image can be sent within minutes of capture.

Digital images can be less expensive to print and easier to enhance, as well. Unlike analog images, digital images can be printed on a low-cost, color ink-jet printer. For easy enhancement, digital images may be imported into a software processing package, such as Adobe PhotoShop.

Resolving the High-Resolution Question

The term high-resolution is often misused and misunderstood in the security industry. High-resolution can refer to spatial resolution (number of pixels per line) or the number of lines per frame. It also may mean the amount of color information in each pixel of the image.

Most digital video equipment will digitize video at either 640x480 pixels or 720x480 pixels. The amount of color information in each pixel will vary, but the systems with the highest resolution will digitize each pixel as 24bit RGB. This retains the full amount of color information possible in digital format.

The digital video decoder that is used typically determines digital image sizes. There are many types, but the most popular formats convert images into 720x480 resolution. This size is best because it contains a larger amount of data per image. An image with a 720x480 resolution is greater than a 640x480 system.

The total image size is determined, then, by multiplying the spatial resolution with color resolution. For example, a 720x480 image with 24-bit resolution (which is 3 bytes) in 1,036,800 bits, or more than 1MB - a fairly large image size not suitable for digital surveillance. That's why compression is so important.

Getting Compressed without Getting Depressed.

In a typical digital video recorder, images are converted to digital, and then they are compressed. Compression processes the data in a image to squeeze the image into a smaller space, for example, in your digital memory or through a phone line.

In today's digital recorder market, there are three types of compression: JPEG, Wavelet, and H.263.

JPEG is the oldest and most established scheme that currently exists. In JPEG compression, the digital image is separated into 8x8 blocks of pixels. Each lock is then assigned a number and coded. The software examines the blocks and decides which blocks are redundant and therefore not essential to creating the image. The program transmits the blocks that are essential, which is a reduced number based on the level of compression requited by the system settings.

Wavelet video compression, rather than operating on 8 x 8 pieces of the image, operates on the entire image. The transformation uses a series of filters that determines the content of every pixel in the image. Each filter outputs a set of coefficients, which represents the result of that filter. Because Wavelet technology works on the entire image, there is no mosaic effect when the image is viewed. Both Wavelet and JPEG produce "lossy" images, meaning they have a loss of image information. However, with Wavelet technology, lossy effects are not apparent until one reaches very high compression rates/ratios. Wavelet technology also provides very efficient motion detection. Because the system can compare the tiniest piece of visual information to the same pixels in the previous image, it can determine a change very accurately.

H.263 is a standard form of compression that is commonly used in video-teleconferencing applications. This technology is very similar to JPEG, except that it only transmits the pixels in each image that have changed from the last image, rather than full images. Because two consecutive images from a camera are typically the same, the H.263 standard capitalizes on this, and therefore employs a frame-differencing technique, which sends only the differences from one frame to the next.