SQUEEZE PLAY
With video compression, CCTV systems with digital video recorders store more data and become network-friendly

By Mark McCall

THERE LITERALLY are hundreds of digital video recorders on the market, each offering a host of features and advantages over traditional VCRs.

Digitized video makes search and retrieval fast and easy. It is possible to go through an entire day of digital video in a fraction of the time it would take with a VCR.

Motion detection, sophisticated editing features and alarms that alert users automatically by pager or by sending live video to remote monitoring stations, are features that make DVRs useful management tools for any size business. The result is an array of possibilities when installing a digital recording unit.

However, the most important advantage of these recorders remains compression, which enables DVRs to store massive amounts of data — eliminating tapes and tape management — and allows for data to be networked more effectively.

Quite simply, the transmission and storage of video would not be possible without a high degree of compression. Similar to other computer applications, video is simply a series of bits. These bits are the computerized “pixels” that make up images. Bit rate, the amount of data it takes to create a complete image, is a key measurement because it determines how much data a computer can store and how quickly a complete image can move from one location to another. Uncompressed video has a very large bit rate. For example, a standard CCTV requires at least 20 mbps per second. At this rate, a CD-ROM could hold just three seconds of uncompressed video, while a standard 30GB computer hard drive holds approximately 25 minutes. Obviously, a smaller bit rate is better.

Transmitting the images through a LAN, WAN or the Internet creates similar difficulties and requires sophisticated compression engines. Uncompressed video would be akin to pushing a zucchini through a straw. It can be done, but it is not pretty. Moreover, the compression has to be suited to packet switching (e.g., TCP/IP) a standard protocol of most networks.

One method to solve the transmission and storage conundrum is by using open compression formats, such as JPEG, MPEG or Wavelet. However, any compression solution comes at a price: image quality.

Compression algorithms all work by eliminating redundant data, or bits. JPEG was originally designed for still pictures and is not very effective. Motion JPEG attempts to provide motion compression, but still compresses at 16:1. MPEG represented a giant leap forward by eliminating much of the redundancy between frames, and achieved compression ratios of at least 100:1. More advanced MPEG algorithms are on the market today (e.g., H.261 and H.263), which have improved the way images are scanned and processed — but still, the files remain large.

As for Wavelet technology, it also has been around for some time, and has made significant advancements in the ability to compress data. Basic Wavelet algorithms seek to remove redundancy by recording only changes in pixelization and eliminating video frequencies that are imperceptible to the human eye. They do so at compression ratios as high as 250:1. This still requires incredibly large hard drive space for storage, and most have to be backed up to DAT tapes.

One solution is to capture fewer lines per inch and at slower frame rates. Unfortunately, the basic quality needs of the security industry make this an unacceptable option. The image must capture sufficient detail; however, every increase in screen size, resolution or frame rates only adds to the bit rate problem. Hence, the rather poor image quality present in many of today’s DVRs.

Due to its complexity and size, even compressed digital video is massive. In fact, few people really comprehend how difficult it is to move digital video through a network. People are more accustomed to sending an e-mail with, at most, a spreadsheet attachment. A standard e-mail is just two to 10 kilobytes, whereas a compressed video file can be 4MB to 5MB for just a one-minute file.

Companies that provide greater bandwidth want people to believe that the simple solution is to build a bigger pipe. In fact, fiber has become synonymous with the promise of faster connectivity — a massive pipe that lies outside a doorstep and delivers data with speed and efficiency. But before it gets to the doorstep, data has to squeeze through a host of components that bring the stream to a trickle. Add to this more applications and users, all running through the same pipe, and it is no wonder IT managers cringe at the thought of streaming video.

The latest generation of video compression seems to offer the best solution for the security industry. These compression algorithms provide the optimal tradeoff between image quality and compression, and are affected by motion within the video.

Using intelligent motion detection, only pixels that change are transmitted and stored, so the bit rate increases proportionally with the motion. Done properly, these systems achieve the industry’s lowest bit transfer rates (typically 5K to 10K per camera), the industry’s highest compression ratio (2400:1 ratio) and, most importantly, maintain high video quality.

Mark McCall is vice president of operations at Ionit Technologies Inc., an Illinois-based company that simplifies security and video management systems by digitally recording images directly onto a computer’s hard drive. For more information, visit www.ionittech.com.