Google
¡¡

Home | Product | Solution |Purchase | contract-us

¡¡
¡¡
How to Calculate the Bandwidth Required for Your Network Cameras

HIGH-end Video Server Module board W1860

W1804 Video Server

W1802D both D1 Video Server

W1801D single channel video server D1

W1802C both CIF Video Server

W1801C CIF single channel video server

When cameras are attached to your network they use up some of the available network capacity. Network capacity is also referred to as bandwidth. We are often asked "low" much bandwidth will all the cameras take on my network in other words, if I add IP cameras to my network, will everything slow down? The answer depends not only on the number of cameras you have, but also the specifications of the camera.

The important specifications include: resolution, the compression scheme used, and the frame rate of the cameras. This article shows you how to calculate the bandwidth utilization on your network and suggests how to maximize the efficiency of your camera system.

What is Bandwidth?
Network bandwidth is defined as the total data capacity that can be handled by a network. If too many computers try to talk on the network at the same time, the network will slow down. It's important to know what the maximum capacity of the network is and how each device affects the total network performance. Bandwidth on a network is usually defined in bits/second (bps). A network uses serial data. For example a 10Base network has total capacity of 10 M bits/sec, and a 100 Base network has a total bandwidth capacity of 100M bits/sec (Mbps).

Number of Cameras
Obviously the more cameras you have, the more bandwidth you will use on the network. For example if one camera provides video at a data rate of 0.960M bits/sec then 10 cameras will use up 9.6M bits/sec. This is not too bad if you have a 100 M bit/sec network. But, if you have 100 cameras, the data capacity increases to 96M bits/sec which is a very significant part of the total network bandwidth. Don't panic yet. There is a way to handle this.

Resolution
Resolution determines how much information is in each frame of video. The higher the resolution, the more data is contained in each frame. You should first determine the resolution you will need. For example you may need higher resolution if you want to identify someone's face, but lower resolution if you only need to see that a person walked by. A resolution of 352x240 is about what you see on your TV, while 704x480 is almost what you see on your computer screen (when set at SVGA, which is 800 x 600). There are higher resolution cameras available that would allow you to electronically pan/tilt and zoom around an image. For example the IQ camera is capable of resolution of up to 2048 x 1536. This is high enough to allow a user to digitally pan/tilt/zoom even after the video has been stored.

Compression schemes
There are two types of compression schemes used with IP type cameras MJPEG and MPEG-4. Most of the IP cameras use MJPEG compression because it provides higher resolution and is accepted as legal evidence by the courts. MPEG-4 provides better compression, but lower resolution. It is usually used to view live video at 30 fps.

MJPEG compresses each frame of video. MPEG-4 not only compresses each frame, it also improves the compression by sending only the difference between each frame. Since there are very few unique frames we don't talk about frame sizes, instead we measure the data rate based on the amount of motion or changes between frames. For example if there is a lot of motion, the data rate can be as high as 512K bits/sec for each camera. If there is very little motion, the data rate can be lower than 32K bits/sec.

Frame Size
The size of the frame (using MJPEG compression) is determined by the resolution.

We use the following estimates of frame size based on the resolution.

Resolution

Frame Size in Bytes

352 x 240

10 KB

704 x 480

30 KB

1024 x 768

60 KB

1280 x 1024

80 KB

1600 x 1200

140 KB


Frame rate
The number of frames per second relates to how smooth the video appears. Video viewed at 30 frames per second looks like your TV, but uses up quite a lot of bandwidth and storage (if you decide to store the video). For example, if the camera is set at a resolution of 704 x 480, it will have a frame size of 30 KB, which is equal to 240K bits. If the frame rate is set for 1 frame/sec, the bandwidth or data rate generated is 240K bits/sec. If we double the frame rate to 2 frames/sec, the data rate from the camera is doubled to 640K bits/sec; and at 30 frames/sec the data rate is 7,200K bits/sec

How to do the Calculations
In our calculations we converted Bytes to serial bits on the network. There are 8 bits in one Byte. To calculate the bandwidth that will be taken by a certain number of cameras (using MJPEG compression) use the following formula:

No. of cameras x Frame Size (in Bytes) x 8 (Bits/Byte) x frame rate (frames/sec) = Data rate (b/sec)

For example if we have 10 cameras set at a resolution of 704 x 480, the chart says we have a frame size of 30 Kbytes. If we want to use a frame rate of 4 fps the calculation is:

10 x 30KB/f x 8b/B x 4 f/sec = 9,600 bits/sec. or 9.6Mb/sec

If we connect these 10 cameras to a network that has a total bandwidth of 100 Mb/sec, we are taking up a very small part of the available bandwidth.

Controlling the bandwidth
Based on the factors we discussed, there are trade-offs that will allow you to use IP cameras effectively on your network. For example, you usually don't need a very fast frame rate to see all the video. If you are viewing race cars streaking by your camera at 100 mph, you will certainly require a fast frame rate, but usually we are watching cars in a parking lot or people walking through a room. These slower moving objects can usually be captured using a frame rate as low as 1 or 2 frames/sec.

Resolution can be controlled as well. You can set each camera to the resolution that is required for the view and application. For example if you just need to see that some one was in a room and the room is not too large, you can reduce the resolution to 350 x 240. This reduces the frame size dramatically and consequently reduces the bandwidth required. Remember that all the cameras don't have to be at the same resolution, some can be at 350 x 240 and some can be at 700 x 480 or higher.

Another way to control the bandwidth is to select a camera that allows you to adjust the aspect ratio. You don't have to stay with the CCTV 4:3 aspect ratio.
Large Camera Configurations
Suppose the number of cameras that you want to connect is very large. For example 100 cameras can take up 96Mb/sec which is close to the total capacity available on the network.

Here is how to handle this. Instead of placing the cameras on your main network, use a network switch to isolate the cameras from the rest of the network. Place the PC running the video software (video server) on this same network. The video server would allow you to view a small subset of the cameras at one time and notify you of any alarm conditions from any of the other cameras. You can create your own separate network or use the VideoZerver which has a built-in network switch. It not only isolates the cameras, but also allows you to control, view and store the video.

IP/Network camera server/Wireless Network Camera/Knowledge Articles:

Network camera server:

Wireless network camera:

IP cameras Related Articles: