Data compression

Overview

The term "data compression" is used to describe a method by which information in a file is manipulated to produce a smaller file. There are two types of data compression- lossy and lossless.

Basics

Due to real-world limitations on data storage space and bandwidth limitations in transmission, it is often desirable to reduce the size of data files. In some instances; any change to the data is not acceptable and in other cases a close approximation of the information is considered acceptable.

For data that must be preserved intact, lossless data compression can be use to reduce the size of a file with some limitations. The total possible reduction is limited by a number of factors, including the amount of "redundant" information in the file and how quickly the file must be processed. For this reason, files containing text can be reduced significantly while files such as images or audio are less likely to gain much reduction in size using lossless compression.

In the quest for higher levels of data compression for audio applications in telecommunications; research in the perception of audio resulted in a form of data compression employing "perceptual coding." By applying signal processing based on how the human mind perceives audio; varying amounts of the original audio information can be discarded while retaining the over-all character of the original signal. This type of data compression is referred to as "lossy" because some of the original information is "lost" in the process. The quality of the results vary with the level of compression, the complexity of the audio, and the discrimination of the listener. For example; what might be considered "acceptable" for voice communication over a telephone might not be considered acceptable for music applications.