Noise – Nuisance, Useful or Truly Valuable?

There was the French philosopher who, when he learned what prose was, realized that he had been speaking it all his life.

Noise is quite the opposite; we think we know what it is from childhood, and those of us who study engineering are surprised to find that it is often something else. The dictionary has over a dozen definitions of noise, the one closest to what we expect is ‘ an unwanted signal or disturbance. ' But how unwanted is it really?

Three Categories of Noise

Scientific types should recognize that there are three types of noise; the nuisance type that comes from the common definition, useful noise, and third, surprising and valuable noise. Nuisance noise needs no further introduction — engineers try to suppress it, whether it be sound, electromagnetic or even just informational noise. Useful noise would include such things as white noise generators that some people use to help fall asleep, or noise introduced to confuse an enemy such as radar jamming.

The third and rarest kind of noise is that which is unexpected and leads to the unraveling of technical mysteries. Perhaps the most famous such discovery was that which lead to a Nobel Prize in 1974 for accidentally detecting a faint noise which turned out to be residual radiation from the ‘ Big Bang. ' The experimenters, Penzias and Wilson, spent countless hours trying to get rid of it before they realized what it was and what it meant.

Taming Noise By Knowing Noise

Noise is characterized by its energy distribution. So-called white noise contains a full spectrum of energy, hence a uniform distribution in the frequency domain. Other common types are Poisson (or shot noise), exponential, and gaussian or normal. Knowing the kind of noise helps one to predict the probability of an instance of a given piece of unwanted energy, as well as recognize the probable source of the disturbance, (i.e. heat, … ) Other ways of characterizing a noise environment are the usual statistical parameters (mean, variance, etc.) and signal-to-noise ratio.

In consumer electronics, there have been two notable events in noise suppression. Recording noise reduction techniques were introduced by Dolby laboratory in the 1960s. Dolby noise reduction is a form of dynamic preemphasis employed during recording, plus a form of dynamic deemphasis used during playback, that work in tandem to improve the signal-to-noise ratio. More recently, noise canceling headphones have come to market; these work by actively generating a signal that interferes with the noise signal so that the user ‘ hears ' nothing.

In the area of national defense, entire industries have sprung up for finding a signal (perhaps a faint one) in noise (perhaps a lot of noise). Typically, the problem involves testing the hypothesis of whether an expected signal is there or not. It is a classic statistical problem, and the consequences of making a mistake, either a false alarm or a missed detection, can be dire.

There are also noise problems associated with stealth technology, either concealing a target from radar, or on the other side, overcoming the stealth techniques to make the target apparent.

When stealth technology was first introduced, it became clear that it could be defeated by looking for areas of too much calm — a sure sign that there was a stealth device in the area. This was countered by introducing just enough noise into the realm to fool the detectors into thinking there was the natural amount of background noise and nothing else.

Noise isn ' t always an unwanted disturbance

So we come to realize that these ‘ unwanted disturbances ' are not always such a nuisance after all. As Mark Twain said about a Wagner opera, it ' s not as bad as it sounds. Noise can be tamed, used to one ' s advantage and even serve as a valuable research tool. If anyone tries to tell you otherwise, well, that's just so much noise.