RTP: Audio and Video for the Internet

In this chapter we have discussed various ways in which errors due to packet loss can be corrected. The schemes in use today include various types of forward error correction and channel coding, as well as retransmission of lost packets.

When used correctly, error correction provides a significant benefit to the perceived quality of a media stream, and it can make the difference between a system being usable or not. If used incorrectly, however, it can lead to a worsening of the problems it was intended to solve, and it can cause significant network problems. The issue of congestion control ”adapting the amount of data sent to match the network capacity, as discussed in more detail in Chapter 10, Congestion Control ”forms an essential counterpoint to the use of error correction.

One thing that should be clear from this chapter is that error correction usually works by adding some redundancy to a media stream, which can be used to repair lost data. This mode of operation is somewhat at odds with the goal of media compression, which seeks to remove redundancy from the stream. There is a trade-off to be made between compression and error tolerance: At some stage, extra effort spent compressing a media stream is counterproductive, and it is better to use the inherent redundancy for error resilience. Of course, the point at which that line is passed depends on the network, the codec, and the application.