Alliance partner, Redline Communications, recently wrote a fantastic blog post that is worth re-posting: Determine the right amount of bandwidth for wireless video surveillance
We’ve been talking a lot here lately about using wireless for video applications. With the broadband wireless technology available today, it’s possible to put high resolution cameras where you need them: on lamp posts for monitoring traffic, in your parking lot for video surveillance, or all over your city for improved security.
Wireless networks are an important component of video surveillance systems enabling the installation of cameras anywhere they needed without having to run cables or fiber. When designing the wireless network the question of how much bandwidth will be needed is going to come up. It’s not a simple question to answer and more bandwidth always translates into more radios. So getting a solid estimate of bandwidth requirements up front will ensure not only excellent performance but excellent performance within the budget.
The amount of bandwidth used by video surveillance cameras is determined by the number of cameras, image resolution, frame rate, and compression ratio. Clearly the more cameras you have, the more bandwidth you will use on the network. A single frame can contain from 30 Kbytes to over 160 Kbytes for high resolution images. At multiple frames per second, each camera could easily require 10 Mbps excluding any network protocol overhead. And if you have multiple cameras the bandwidth increases linearly. Fortunately there are ways to reduce this.
Resolution
Higher resolution means a greater amount of detail is captured in a video image. A standard definition TV image has a resolution of 352240, while your PC typically has 704480. High resolution cameras have resolutions up to 2592 x 1944. As the resolution goes up, so does the amount of bandwidth required, so it’s best to select a resolution level that meets your needs.
Compression
Video compression is an important tool in helping to ease strain on the network allowing high quality video transmission without hoarding bandwidth. Compression works by comparing the differences between frames meaning bandwidth requirement are directly related to what is being monitored. For example, monitoring a static scene requires less bandwidth as compression rates will be very high. On the other hand, monitoring traffic or public areas require higher network bandwidth.
Video surveillance systems typically use M JPEG, MPEG-4 or H.264 compression technology. Which one you use depends on your application and needs. MPEG-4 provides better compression, but lower resolution. It is usually used when you need to conserve bandwidth and storage. MJPEG compression is usually used by the higher resolution cameras. H.264 is the latest compression technique and offers an excellent trade-off between quality and bandwidth. H.264 provides about twice the compression of MPEG-4 for the same video quality.
Frame Rate
The number of frames per second relates to how smooth the video appears. Standard TV runs at 30 frames per second and uses up quite a lot of bandwidth. Fortunately in most video surveillance applications, there is not much motion. For a person walking through the room, a frame rate of 4 fps will be sufficient. Increasing the frame rate increases bandwidth usage. To minimize bandwidth while retaining high quality images, most video surveillance systems increase the frame rate only when motion is detected.
What’s the bottom line?
Due to the many factors that contribute to the video bit rate, there is no exact method to calculate total bandwidth requirements. Each camera manufacturer and each model of encoder will have different requirements, and every installation will introduce unique variables.
Most wireless video surveillance systems are installed outdoor. Weather conditions, the topography, line-of-sight conditions, and interference spectrum all impact the amount of bandwidth on a given link. Careful planners provide an additional 20% bandwidth to mitigate the impact of these variations. The following charts give a reasonable starting point.
| Typical CIF to 4CIF Bandwidth Requirements | |||
| Resolution | Image Rate | Activity Level | Bit Rate (Mbps) |
| CIF (176144) | 3 | Medium | 160 |
| CIF | 7 | Medium | 185 |
| CIF | 15 | Medium | 200 |
| CIF | 30 | Medium | 500 |
| 2CIF (352288) | 3 | Medium | 320 |
| 2CIF | 7 | Medium | 370 |
| 2CIF | 15 | Medium | 400 |
| 2CIF | 30 | Medium | 1,000 |
| 4CIF (704576) | 3 | Medium | 640 |
| 4CIF | 7 | Medium | 740 |
| 4CIF | 15 | Medium | 800 |
| 4CIF | 30 | Medium | 2,000 |
| Typical Megapixel Camera Bandwidth Requirements | ||||
| Compression | Megapixel | Resolution | Image Rate | Bit Rate (Mbps) |
| H.264 | 1 | 1,280 x 1,024 | 5 | 0.7 |
| H.264 | 2 | 1,600 x 1,200 | 5 | 1.03 |
| H.264 | 3 | 2,048 x 1,536 | 5 | 1.6 |
| MJPEG | 1 | 1,280 x 1,024 | 5 | 4.7 |
| MJPEG | 2 | 1,600 x 1,200 | 5 | 6.2 |
| MJPEG | 3 | 2,048 x 1,536 | 5 | 9.6 |
| H.264 | 2 | 1,600 x 1,200 | 15 | 3.8 |
What about the network equipment?
Most wireless equipment specifies bandwidth under ideal conditions. For example measuring bandwidth using only large packet sizes can inflate the amount of bandwidth. Video compression results in a range of packets sizes from small to large. Check the radio specifications or talk to the manufacturer to ensure bandwidth is consistent under varying packet sizes.
Latency can impact bandwidth as latency is related to the amount of time it takes to access the wireless link. High latency means wasted bandwidth since the wireless link is not being used during this access period. Latency also impacts the quality of video surveillance operation. High latency can result in fuzzy or pixelated images and poor performance of video analytics such as license plate recognition. Always select wireless equipment that has low latency and low latency that does not vary under traffic loading. Look for radios with less than 8 msec.
In video surveillance systems the uplink, from the camera to the control center, carries most of the traffic. Some radios use a fixed uplink and downlink ratio which means downlink bandwidth cannot be assigned to the uplink. Choosing radios that provide dynamic time-division duplex (TDD) allows any uplink:downlink ratio to be assigned.
Conclusion
Video stream quality and network bandwidth are directly correlated. With some up front planning to determine the amount of bandwidth required and an understanding of the key factors that impact bandwidth in outdoor wireless networks, it is straightforward to design and deliver excellent system performance to meet your needs.