Databases Reference
In-Depth Information
Scalar Quantization
9.1 Overview
In this chapter, we begin our study of quantization, one of the simplest and
most general ideas in lossy compression. We will look at scalar quantization
and continue with vector quantization in the next chapter. First, the general
quantization problem is stated, then various solutions are examined, starting
with the simpler solutions, which require the most assumptions, and proceeding
to more complex solutions that require fewer assumptions. We describe uniform quantization
with fixed-length codewords, first assuming a uniform source, then a source with a known
probability density function (
pdf
) that is not necessarily uniform, and finally a source with
unknown or changing statistics. We then look at
pdf
-optimized nonuniform quantization,
followed by companded quantization. Finally, we return to the more general statement of the
quantizer design problem and study entropy-coded quantization.
9.2 Introduction
In many lossy compression applications, we are required to represent each source output using
one of a small number of codewords. The number of possible distinct source output values is
generally much larger than the number of codewords available to represent them. The process
of representing a large—possibly infinite—set of values with a much smaller set is called
quantization
.
