Breaking Down Digital Video Barriers
Let Application Specifics Guide Compression Method
As the previous section indicates, compression scheme is the key to achieving powerful video surveillance systems. Through the years, many such schemes have emerged but security systems focus primarily on MPEG and H.264. These algorithms take a reference image every so often and for subsequent images look only at the changes with respect to the reference image. That estimation is what gives it the ability to reduce image sizes as much as they can.
Frame-based prediction is not always the best answer. At lower frame rates where the amount of change from one image to the next can be large, the bandwidth savings that arise from this type of motion estimation are limited. However, newer and affordable advancements are leading the trend toward the use of TV-quality frame rates where codecs like MPEG-4 and H.264 are necessary and practical.
The compression gains with MPEG have made it a very popular standard in video systems. Note, though, that even within MPEG there are several options. MPEG-2 is an older standard used primarily for DVDs and broadcast video, whereas MPEG-4 provides better quality for a given bit rate. The most recent addition is H.264 (also known as MPEG-4 Part 10 or AVC for advanced video coding), with even better compression.
As a rule of thumb, for a standard 30 FPS analog camera with H.264 compression, the bit rate is 1Mbps. In general, MPEG-4 would require 50-percent more bandwidth, and MPEG-2 would require 50-percent more bandwidth again compared to MPEG-4.
The tradeoff for this quality compression, though, is more computational horsepower to run the compression algorithm, which can then translate into higher cost and a longer latency time. However, as signal processors become more powerful and less expensive the impact of this will be mitigated in the next few years.
A systems integrator or security manager must determine the minimal video quality that is acceptable for a system. This will determine the image resolution and frame rate that will, along with the choice of codecs used, dictate the necessary data rate and amount of computational power and storage needed. Today MPEG-2 is used in legacy systems and is not implemented in new installations. MPEG-4 is the midpoint and often used for low-end systems, while H.264 is the common choice for new systems.
Video Analytics Can Help Ease Network and Storage Burdens
Several factors contribute to this shift from MPEG-4 to H.264. Since MPEG-4 was developed, engineers have created new ways to improve compression, and many of these have been implemented in H.264, which allows the use of multiple reference frames instead of just one and supports variable-sized prediction blocks.
Depending on how the features and options are implemented, it can mean a considerable difference in compression efficiency, but with a tradeoff in computational overhead and latency. Thus, people working with H.264 can select from a handful of pred
efined feature sets known as profiles. Of these, the H.264 high profile provides streaming support for low-latency multiresolution video, a key requirement in video surveillance applications. In addition, there are many other “knobs” to tweak when trading off between quality and bit rate, and they must also be set properly.
Latency, as noted earlier, is also an important issue. Today, 100-millisecond latency between the time an event occurs and time it appears on the screen is generally acceptable. This is best handled as a system level issue. If latency is a crucial concern, the system should feed the raw video immediately to a preview display rather than to a recording/storage system and then to the display. Of course, each compression scheme adds some time between the raw data and the stored image, but system developers who are concerned about latency should highly consider using H.264.
A concern with the transmission of video data is and whether there is enough bandwidth to transmit all the image data in real-time. In a hybrid system, which includes analog cameras and DVRs with built-in hard drives, the DVR must have the necessary bandwidth to handle all the video feeds. The same issue arises with purely digital feeds from IP cameras. The choice of codec and bit rate can also determine whether the network and storage are adequate.
In security applications, it is important not to overload the network, so over-provisioning is the best route. 100Mbps Ethernet is generally very reliable for eight 1Mbps cameras, whereas 10Mbps Ethernet might not be. Of course, the final determination depends on which switches and hubs are being used. Storage has always been a concern in security applications, but rather than send all the data all the time, why not limit it to when only key points of interest occur? A new class of intelligent cameras and DVRs are implementing intelligent video analytics, which is the ability to examine an image for user-programmed events and to detect something of interest such as the presence or absence of people. Common examples of video analytics include the detection of potentially suspicious activity such as loitering in sensitive areas, traffic monitoring, airport security and the detection of objects left behind. With the help of analytics, video can be sent or stored only when something of interest happens, reducing overall network and storage demands.
If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!
Related Content
Security Is Our Business, Too
For professionals who recommend, buy and install all types of electronic security equipment, a free subscription to Commercial Integrator + Security Sales & Integration is like having a consultant on call. You’ll find an ideal balance of technology and business coverage, with installation tips and techniques for products and updates on how to add to your bottom line.
A FREE subscription to the top resource for security and integration industry will prove to be invaluable.
