Fundamental Data Compression

Chapter 8

Prediction and transforms

We have seen several lossless compression algorithms. In practice, these algorithms are often used together with some other methods to achieve better overall compression efficiency. This can be done by lining up several different algorithms and using them one after another. Certain changes can be made on the source distribution before applying the lossless compression algorithms. The source data can also be transformed to another domain to achieve better overall compression performance.

The preparation work before applying certain lossless compression algorithms is called preprocessing. Examples of typical preprocessing operations include sampling and quantisation as well as prediction and transforms, all ...