Hardware Compression System vs. Software Compression System

March 12, 2010

PC-based DVR systems have become the choice for surveillance system integrators and users due to their higher scalability, flexibility and ease of customization. However, choosing a hardware compression system or software compression system can make a big difference.

By Hikvision Digital Technology
PC-based DVR systems appeal to system integrators and users because of their unique flexibility and scalability. By adding more compression cards, the system can support more video channels, and desired functionality can be achieved by adding daughter boards. In addition, the provided SDK allows the DVR system to be customized to meet specific requirements. A PC-based DVR system captures surveillance video in real-time and saves it to your hard drive. To do this, it has to gather the signal, and compress it, and save it to a video file on your drive. Every card captures a certain number of frames per channel (for example, a 16-channel card that captures 100 frames per second in total, captures 6.25 per channel, whereas one that captures 400 fps, can capture 25 per channel). Then, the system must compress it and save it. Either using a software process, or the DVR card itself does it.

It's important to note that a PC-based DVR system can be made using either a software compression card or a hardware compression card. As the name suggests a software compression system (SCS) uses the card to capture the signal, and software to compress it. It's your computer's CPU and RAM that do this work, which comes at a high cost of your system resources.

A hardware compression system (HCS) not only uses the card to capture the signal, but will also compress it. It will have one or more chipsets on the card that are responsible for capturing, compressing and saving the video to the hard drive.

Software Compression System
The biggest problem of software compression DVR system is CPU Loading. This means that your computer's CPU, already busy with other things, is being heavily used by the DVR system. This may not be a problem with DVR systems running at a lower frame rate, but with faster, real-time systems, where the processor has 400 4CIF images to compress every second, the CPU can quickly max out. Why is it a problem? There are several reasons.

First, not all frames are the same size, nor they do not all require the same amount of work to process, which means the workload varies constantly regardless of the compression codec being used. Also, the CPU is busy processing other things like GUI (graphic user interface) changes, playback functions, motion detection and network operation, to name just a few. With the CPU under such heavy use, several large frames in a row can cause serious problems.

With the normal operation of your PC you see this all the time. One process simply waits its turn at the CPU, so you'll see a window take extra long to redraw, or the screen will take longer to refresh. This is not normally a problem. But with DVR operation, such 'traffic jams' are bad news. At 400 frames per second (4CIF resolution), you computer gets overloaded pretty quickly and starts to drop frames. Frame drop means gaps or 'blips' in your recorded video and audio artifacts. You may not notice occasional drops, but if your CPU drops more than a few, you will see it. In severe cases these traffic jams can cause your computer to crash completely.

Because a hardware compression DVR system uses its own chips for compression, the load on the computer's CPU and RAM is a great deal lower; frame drops almost never occur on these systems.

Hardware Compression System
A hardware compression system has one or more chipsets on the card that are responsible for capturing, compressing and saving the video to the hard drive. This means it has good performance coupled with low PC CPU and RAM resource requirements . For this reason a hardware compression DVR system is far more stable than software compression DVR system.

And because hardware compression cards usually use high performance chipsets to compress frames, the compression process will not be affected or interrupted by other programs. It brings more fluid video image to customers.

Meanwhile, with a low CPU occupancy rate, hardware compression system can easily accomplish additional functions by using spare system resources of computer, such as intelligent analysis, license plate recognition and so on. Network connection performance can be better as well, in particular fast transmit speed and multi-channel simultaneous connection.

Hardware compression systems can also easily support up to 64 channels in one PC with real time monitoring and real time recording (every channel is 25fps in PAL or 30fps in NTSC). Most software compression system supports only 16 channels.