Browse Books

Reviewer: Stefan A Robila

Data compression is generally understood as the process of transforming data for the purpose of size reduction. Data compression is currently employed wherever information is handled, be it image (still or video), text, or sound. Understanding the methods used to compress data, and the costs associated with these methods (in terms of size reduction factor and loss level, as well as computational complexity), is an important endeavor for any reader, and an essential one for anybody with a general computing interest. The author’s comprehensive and informative presentation of these topics courageously does away with most of the theoretical topics needed to support the concepts—which is most welcomed. Salomon’s book has reached its fourth edition as a revised and improved version. It is more than double the size of the first printing, which was published only ten years earlier. The new version is identified as a “Springer major reference work,” rendering the text “free from constraints of having to compress the author’s style,” and thus hinting that the text is the most complete reference compared with any previous work. The book’s ambitious goal, to provide an overview of the compression field, stays the same, and is met by adding newly developed algorithms for traditional problems, such as MPEG-4 audio lossless coding (ALS) or differential file compression. New problems and approaches are also introduced, for example hyperspectral data compression. The writing style combines formal discussion of the methods with personal insights and stories. In addition, there are many quotations, which aim to provide a less rigorous view of the field. The text is easy to read, and is free of significant errors (apart from a striking overprinting of a graph over text on the third page). A collection of exercises for each chapter and a solid glossary are included. However, the book seems hardly suitable for textbook use: the topics are too broad for a course to adequately cover, and the problem solutions are included. The book is organized into eight large chapters. Some focus on the approach used for compression (basic techniques, statistical, dictionary, and wavelet), while others are centered on the nature of the data (image, sound, and video)—it is unfortunate that the author did not clearly split the book into two sections: one aimed at general approaches, and the other at techniques for particular data types (including regular text or generic files). The confusion only increases with the last chapter, “Other Methods.” This section describes traditional approaches, which do not fit any of the previous categories, and compression of newly introduced data types (such as hyperspectral data). An important and beneficial aspect of the book is the structure used for introducing most of the topics—this is exemplified in Section 4.8 on JPEG compression. Salomon begins by presenting the topic in general, and doing away with most of the theory. Next, the reader is provided with in-depth coverage in supporting references. Salomon then proceeds to discuss each of the compression and decoding steps in more detail by analyzing the theory behind each of them (for example luminance, discrete cosine transform, and coding). This way, the book is attractive to both the incidental reader (who needs a brief description and no details) and the more thorough one (who may use it for starting a research or teaching project). Overall, it is obvious that this book has benefited from ten years of writing and revisions. As the author acknowledges, it includes any corrections that have become obvious since the previous edition. In addition, at least 17 of the chapter sections (more than 10 percent of the book’s content) were significantly revised or newly introduced, with most of these changes found in the last two chapters (“Video Compression” and “Other Methods”). The book’s size is alarmingly large, but appropriate for a complete reference work; it is rightfully titled. Online Computing Reviews Service