Digital video surveillance would not be practical without digital compression. Digitizing video with high resolution produces a massive amount of fast data that is not only difficult to transmit over long distances but also requires massive storage capability. Several video compression techniques greatly ease the data rate and storage problems.
The three most common compression methods used in video surveillance are Motion JPEG (MJPEG), MPEG-4, and H.264. Most digital systems started out with MJPEG, which compresses the video frame by frame. It requires the least amount of CPU processing power, but it also requires the most storage capability and transmission speed. MPEG-4 is more efficient but requires a continuous amount of frame processing. Lots of CPU horsepower is needed, but that’s relatively easy to come by today.
While H.264 is an ITU standard, it’s also Part 10 of the MPEG-4 standard. It provides the best tradeoff between quality and bandwidth. H.264 provides about twice the compression of the basic MPEG-4 for the same video quality, meaning it provides the highest-quality video with the lowest bit rate. Most codecs today are either MPEG-4 or H.264, the percentage being roughly equal with a significant trend toward H.264. Some systems support both MJPEG and H.264.
The actual bandwidth needed to transmit compressed video depends on all the characteristics discussed earlier, including video format, frame rate, and compression type (see the table). The digitized and compressed video is then packetized in TCP/IP. The procedure is usually encapsulated in Real-time Transport Protocol (RTP), which is encapsulated in User Datagram Protocol (UDP). UDP then is encapsulated in IP, usually IPv4, but more recently in IPv6. The resulting packets then typically will get packaged into standard Ethernet frames for their ride over CAT5e/6 or wireless.
In the past, most video was stored on videotape. Most VCRs used a T-120 VHS cassette with a two-hour limit. The slower frame rate greatly increased the amount of video. Many systems had multiple VCRs. In addition, each system had different policies about how much video to archive and for how long. No new systems use VCRs thanks to the introduction of the DVR.
DVRs, which store data on a large hard drive, are common today, especially in cable and IP television (IPTV) systems. Thanks to video compression, huge quantities of video can be easily stored. Special DVR systems for surveillance are common. These include analog camera inputs that are digitized by analog-to-digital converters (ADCs) and compressed before storage. IP cameras provide other inputs.
The typical DVR appliance is a multi-channel, time-lapse video recording device and Web server. A common box comes with four, eight, 16, or 32 input channels. You can change the frame rate, record, or playback function or access it remotely. Most come with at least a 500-Gbyte hard drive, but options of 750 Gbytes and 1 Tbyte are common. H.264 compression is the most widely used. Some units come with a DVD burner.
A key factor to consider is the maximum frame rate of a DVR, which is typically 120 fps. This rate has to be divided up among the cameras. With four cameras, each can record at 30 fps. But if more cameras are used, the frame rate of each will have to decrease. The total number of cameras times the frame rate cannot exceed the 120-fps maximum. As it turns out, most cameras today only record at a 3- to 5-fps rate anyway.
The newer network video recording (NVR) systems use special servers to store the compressed video on a large hard drive or multiple drives in a redundant array of disks (RAID) configuration. Hybrid DVR/NVR systems are also an option for systems that are being expanded from an older installation.