2. Why codec

Bandwidth is a limit of the network transmission throughput. It is the resource of the networks. The data rate through the network is bounded above by this bandwidth. If there are many users, the bound may be very weak, because many users will be sharing the network throughput.

For the real-time communication over a network, the size of the image should be big enough to be seen by the communicators and the frame rate should be fast enough to covey the impression of a live conversation for both images and sounds [12].

Because human eyes perceive 25 frames per second as a continuous motion, the normal full motion videos are 25 fps (frames per second) or more. If an image size is 640 * 480 pixels with full resolution, the frame rate is 25 fps, and we use 24 bit colour depth, then the throughput to transmit such a video stream, without compression, would be 184,320,000 bps (bits per second)! If bi-directional transmission is used, the throughput should be doubled, e.g. 368,640,000 bps. It is impossible to realise this communication rate today. That's why image compression is used during networked multimedia transmission. Before being transferred over a network, the video and audio streams are first encoded into compressed packages.

At the receiving end, those packages are decoded to regenerate the original video streams. For example, the motion JPEG compression ratio is somewhere between 2:1 to 100:1, depending on the different quality required. Typically JPEG achieves 10:1 to 20:1 compression without visible loss and 30:1 to 50:1 compression is possible with small to moderate defects. This of course depends on the image composition and high contrast artifacts can cause visible loss even at low compression rates. For very-low-quality purposes such as previews or archive indexes, 100:1 compression is quite feasible[13]. The compression ratio achieved varies with different types of compression. Even with the same compressing method, the ratio varies for the contents of the images.

This data compression technique is used to enable the data of audio or/and video streams to be transferred smoothly through network at a rate that can be tolerated. Networks can only transfer at a limited data rate. Of course, the side effects of the compression and decompression are the time delay experienced in the execution of the algorithm, and signal loss of the communication. In the real-time application, the codec could be accomplished by hardware which is relatively faster with higher quality transferring data but needs extra investment or by software which is relatively slower with lower quality transference data but cheap.