Digital Signal Processing Reference
In-Depth Information
Illustration 40:
Conserving tendency of a noisy signal
Both Illustrations - the time domain above, the amplitude spectrum below - describe a noisy signal, that is
not pure stochastic noise, which displays a conserving tendency (influenced by the signal).
This is shown by the amplitude spectrum below. A line protruding from the irregular continuous spectrum
at 100 Hz can clearly be seen. The cause can only be a (periodic) sinusoidal signal of 100 Hz hidden in the
noise. It forms the feature which conserves a tendency although it is only vaguely visible in the time
domain. It could be "fished out" of the noise by means of a high-quality bandpass filter.
This is in fact the central theme of "information theory". As it presents itself as a theory
formulated in purely mathematical terms, we shall not deal with it systematically in this
topic. On the other hand,
information
is the core term of information and communications
technology. For this reason important findings of information theory turn up in many
places in this topic.
In this topic the foundations of signal processing are physical (see Chapter 1: "In techno-
logy nothing works that is contrary to the laws of nature"). But physics has still not suffi-
ciently succeeded in integrating the concept of information in the fundamental principles
of nature. There is no "Information Principle" in physics. Here the following applies:
