Cryptography Reference
In-Depth Information
5.2 Patterns and Compression
Life often reduces to formulas. At least it does on television, where
the solutions appear every 30 or 60 minutes. When you know the
formula, it is very easy to summarize information or compress it.
A network executive reportedly commissioned the television show
“Miami Vice” with a two-word memo to the producer reading, “MTV
Cops.” You can easily specify a episode of “Gilligan's Island” with a
single sentence like, “The one with the cosmonauts.” Anyone who's
seen only one episode of the show will know that some cosmonauts
appear on the island, offer some hope of rescue, but this hope will be
dashed at the end when Gilligan screws things up.
Compressing generic information is also just a matter of finding
the right formula that describes the data. It is often quite easy to find
a good formula that works moderately well, but it can be madden-
ingly difficult to identify a very good formula that compresses the
data very well. Finding a good formula that works well for specific
types of data like text or video is often economically valuable. People
are always looking for good ways to cut their data storage and com-
munications costs.
Compressing data is of great interest to anyone who wants to hide
data for four reasons:
Less data is easier to handle.
This speaks for itself. Basic text can
easily be compressed by 50 to 70%.
Images might be com-
pressed by 90%.
Compressed data is usually whiter.
Compression shouldn't destroy
information in a signal. This means that the information per
bit should increase if the size of the file decreases. More infor-
mation per bit usually appears more random.
Details about
measuring
information are on page
31.
Reversing compression can mimic data.
Compression algorithms
try to find a formula that fits the data and then return the spe-
cific details of the formula as compressed data. If you input
random data into a compression function, it should spit out
data that fits the formula.
Page 183 shows how the
JPEG algorithm can
identify just howmuch
space can be exploited
in an image.
Compression algorithms identify noise
Good compression algorithms
understand how human perception works. Algorithms like
JPEG or MP3 can strip away extra information from a file that a
human doesn't notice. This information can also be exploited
by steganographers to locate places where a significant amount
of noise might be replaced by a hidden message.
