Customer Reviews

The Data Compression Book

5 star:		(2)
4 star:		(2)
3 star:		(2)
2 star:		(1)
1 star:		(2)

 

 

PROs: 1. It is one of very few books on data compression available on the market. 2. Description of the IDEAS of compression techniques is very well written. 3. The books comes with the C code for most algorithms. 4. Fairly wide scope of data compression techniques is presented.

CONs: 1. Possibly for copyright reasons, the formats of commonly used file formats are not disclosed; the enclosed propgrams are generic compression algorithms, which do not create (or open) actual .ZIP, .ARC, or .JPG files, which can be opened by commercial programs. Therefore, this book will not help you to open standard compressed files from your home-made programs. 2. There is a missing link between the well described ideas (general principles) of the compression techniques, and their actual algorithms presented as C programs - namely, the algorithms are not described verbally. You have to analyze typically 6-page-long programs to understand how the actual encoding is done. 3. Although there is a section on sound compression, the MP3 standard is not explained. The same applies to MPEG.

SUMMARY: Good to get a general idea how the data compression is performed. Helpful if you want to develop your own compressed data format. Of very limited help if you want to work with standard compressed files in your own program. Requires knowledge of C and some time to study the enclosed code.

This book's target audience is the novice C programmer who needs to implement data compression of some kind. The authors go to great pains to explain exactly how the code works, but they don't do as good a job on the algorithms themselves. If you are a competent C programmer and/or have any formal training in algorithms, this is probably not the book for you, though it may be a good jumping-off point if it's the only book you can get your hands on.

If you want to implement a compression scheme fast easily, or to have general knowledge about compression algorithms this is a very good book. If you really want to deeply understand compression algorithms, to have some kind of insight about them, and to know about the best algorithms available today - this book is has less benefit (understatement).

This book did a better job than any other of teaching me sliding window dictionary encoding and adaptive huffman encoding. It got right to the point in an easy to understand fashon. There was no thick cloud of theory masking the information I needed. Read this book before you read others on the subject of data compression.

If your primary interest is in writing compression algorithms, this is the book you want. It does a fine job of explaining several different techniques, and shows you how to implement them.

It's light on the theory side, so if you're more interesting in the "why" than the "what", you may want to choose a different book. One exception to this is the explanation of arithmetic coding and statistical methods, which can actually be read and understood in one pass.

This books is for C programmers, interested in understanding Data compression. Both Lossless and Lossy is covered. Easy to read, and each compression is implemented in C. You will not find Mathematical details in this book.

The data compression topics are covered in the most light-handed way possible. The promos for the book advertise that the JPEG compression standard is "discussed", and that source code is provided ... but the source code provided is NOT a JPEG encoder/decoder, but rather an abstract Huffman Table/Quantization program that is ONE HUNDRED FRICKIN' PERCENT USELESS.

From zip archives to jpg images, compressed data are ubiquitous on
hard drives and on the internet. Consequently, a book offering
the intermediate-level programmer a hands-on introduction to data
compression is valuable. Even better is a book that does so using
C language source code compiling to ready-to-use executables.

That better book is The Data Compression Book (2nd edition).
Published in 1996 and now out of print (but available through
used book vendors), this book takes the reader from Huffman
coding to the Ziv - Lempel LZ77 and LZ78 dictionary compression
algorithms that are the heart of much lossless data compression.
Lossy compression of audio and image data are also addressed.
Most valuable to me is the patient yet thorough coverage of the
discrete cosine transform and the entropy (i.e., "zigzag")
encoding central to jpg images. The second edition ends with a
new chapter on fractal image compression, a fascinating topic
sadly encumbered within a minefield of patents. Yet, as those
patents expire, fractal methods introduced here may see a
renaissance.

Anyone integrating the standard compression libraries (e.g.,
zlib, libjpeg, and libpng) into their software will find this
book useful. And if you anticipate modifying those libraries
(perhaps to introduce data hiding (steganography)), this book
will be even more useful through the technical background it
provides.

The book does show its age in places. (1) Source code originally
came on a floppy disk. The used copy I bought lacked it. Emailing
the first author (markn at ieee dot org) brought a prompt reply
with an attached zip file of all source files. (2) Not covered,
of course, are compression methods developed since 1996. These
include the LZMA methods used by 7-zip and Rar, mp3 audio
compression, and the Burrows-Wheeler algorithm used by bzip2. (3)
A graphics display utility program that is part of the source
writes directly to video memory, a classic DOS-era hack. That's
not apt to work anymore! A simple workaround is to convert the
graphics test files supplied with the source to portable graymap
files. Do that by prepending the following three lines of text

P5
320 200
255

to the test image file and by changing the extension of the
resulting file to .pgm. Now most image display programs (e.g.,
Gimp, Irfanview, Xnview) will display the image. (For how this
works, read the Wikipedia article on the Netpbm graphics format.)
Using mingw gcc 3.4.5 on Windows XP, I found compiling and
linking the source code to be problem-free. The authors tried
hard to keep their C code as portable as possible.

Finally, to those who might snub this book for the attention it
gives to first-generation compression methods such as Huffman
coding, I offer this anecdote. Adaptive Huffman, I found, often
beats other more powerful methods, including 7-zip and rar, when
the data set to compress is tiny (i.e., about the size of a
"tweet", 140 bytes or less). Adaptive Huffman wins because,
unlike the more powerful methods, it includes no header or other
overhead in the compressed file. Older can sometimes be better.

searshing for books is very amazing in Amazone.com but i wish if we (customers )have more information about authers where do they teach; thier email andsome things like that so we could make our desecion