Ever wondered why white noise is white? It's not the only colour of noise you can get. There are a handful of others too -- which are used in a variety of different situations, including acoustics testing, electrical engineering and hard physics.
They're separated from each other by their spectral density -- the way the power that the signal contains is distributed over different frequencies. Different spreads are given different colour names, including white, pink, brown, blue, violet and grey. Here's how they differ from each other, and what they tend to be used for.
White noise
You might think of white noise as the sound of waves on a beach, or talk radio playing quietly -- something that overrides background noise. Sound technicians define it more tightly -- white noise has equal energy per cycle, meaning that its frequency spectrum is completely flat. The band of signal between 20 and 40Hz contains the same power at the band between 2000 and 2020Hz. It's referred to as white because it's analagous to white light -- where every band of the spectrum shines equally brightly, combining to produce white.
White noise's ability to cut through background noise means that it's easy to locate the source, and as a result it's sometimes used as part of a siren on an emergency vehicle. It's also used in audio synthesis to create percussive instruments like cymbals, which produce a lot of noise.
As it contains an equal amount of high frequencies to low frequencies, white noise is also found in sound-masking tools, such as those produced for tinnitus sufferers, and to aid a good night's sleep. Finally, white noise is used as a basis for some random number generators. Random.org uses atmospheric antennae to generate random digit patterns using white noise.
Pink noise
While white noise has equal power in frequency bands of equal width, pink noise's spectrum falls off logarithmically over time, with equal power in bands that are proportionally wide. As a result, the band of signal between 20 and 40Hz has the same power in it as the band between 2000 and 4000Hz.
That's useful because it's the same as humans hear -- each octave in the western musical scale contains the same amount of energy as the octave above and below it, despite encompassing a wider frequency range. Pink noise replicates that phenomenon, making it very useful as a signal to test amplifiers and loudspeakers. Pink noise is also seen in meteorological data series, the radiation output of some astronomical bodies and in the statistics of DNA sequences.
Brown(ian) noise / Red noise
Red and brown noise are the same thing, and are more accurately known as brownian noise. It's nothing to do with the fabled brown note, but instead is the type of signal noise produced by Brownian motion. As it's heaviy weighted towards the lower end of the spectrum, it sounds like a low roar. It's used in climatology to describe climate regime shifts, but there's a certain amount of disagreement (PDF) within the scientific community as to its value for such a purpose.
Blue and Violet noise
On the other end of the spectrum to red and brown noise lie blue and violet noise. Blue noise is a little like pink noise in that it changes proportionally with increasing frequency, but instead of a decrease you see an increase. Retinal cells are arranged in a blue-noise pattern, which a number of studies have shown yields good visual resolution.
Violet noise is, similarly, like an inverted version of brownian noise. Its power density increases per octave with increasing frequency over a finite frequency range. It can also be used in the treatment of tinnitus, as its high frequencies tend to mask the high frequencies heard by many tinnitus sufferers.
Grey noise
So we've seen noises weighted to higher and lower frequencies.
What about both at the same time? That's grey noise -- which has lots of power at the top and bottom end of the frequency spectrum but very little centred around the range of normal human hearing. More specifically, it's designed such that given a regular human ear experiences every frequency equally loud. It's white noise, but tuned specifically to the psychoacoustics of the average human ear.
As a result, it's useful in studies of hearing difficulties, allowing researchers to assess how a particular person's hearing differs from the average.
Others
They're all the colours of noise that have a widely accepted meaning, but there are a selection of others that are less tightly-defined. Orange noise is a spectrum that has a few bands eliminated entirely -- those centred around frequencies of exact musical notes, meaning that it plays everything that isn't perfectly in tune, and nothing that is. These "sour" notes give the noise its name. It can be used to check tuning.
Then there's black noise, which has a number of different explanations. Some define it as pure silence, others describe it as the output of an active noise control system that cancels out existing noise -- essentially negative noise. One maker of an ultrasonic vermin repeller calls its output -- which is actually white noise that's silent below 20,000Hz -- black noise.
Finally, green noise is described as either the mid-point of the white noise spectrum, or the background noise of the world -- as a power spectrum that's averaged over a very long time in several outdoor sites around the planet. It looks rather like pink noise, but with a hump added around 500Hz.
Heard of any other noise colours, or different definitions of the above? Post them in the comments below.
This article was originally published by WIRED UK
