White noise is a signal made of uncorrelated samples, such as the numbers produced by a random generator. When such randomness occurs, the signal will contain all frequencies in equal proportion and its spectrum will turn flat.
Most white noise generators use uniformly distributed random numbers because they are easy to generate. Some more expensive generators rely on the Gaussian distribution, as it represents a better approximation of many real-world random processes. Both generators will sound the same though, and will exhibit the same flat spectrum. They will only differ by the distribution of their sample levels.
Listen to our two examples: both noises play at same loudness; however the Gaussian version peaks at a higher level (0 dBFS v/s -6 dBFS). This can be explained as follows: compared to the uniform distribution whose sample amplitudes are equally distributed between a minimum and a maximum value, the Gaussian distribution produces a higher density of low level samples. To keep the loudness constant, Gaussian noise must then produce higher peak amplitudes. In other words, high level samples are less frequent in Gaussian noise than uniform noise, but much higher in amplitude.
White noise has been named by analogy to light, which turns white when all frequencies are summed up into a single beam. As light changes its color when altering its frequency distribution, noise can be "colored" too, by shaping its frequency content. The best known colors are pink and brown.
There are many different kinds of waves. Sound relates to pressure waves, only audible to us between 20 Hz (bass) and 20 thousand Hz (treble). Light relates to electromagnetic waves, only visible to us between 430 trillion Hz (red) and 750 trillion Hz (violet). There is no relationship between sound and light waves: noise colors are just a handy analogy, nothing more.
White noise has equal power in equal bandwidths. For example, the 10 Hz bandwidth between 20 Hz and 30 Hz contains the same amount of sound power as the 10 Hz bandwidth between 10,000 Hz and 10,010 Hz.
For the human auditory system, white noise sounds much brighter than what one would expect from a "flat" spectrum. This is because human hearing senses frequencies on a logarithmic scale (the octaves) rather than a linear scale.
On the logarithmic scale, white noise packs more energy in the higher octaves, hence its bright sound. By flattening its spectrum logarithmically, white noise will turn pink.
In audio applications, white noise is used as a reference tone to check frequency responses: play back white noise through your system and check its output with a linear spectrum (FFT) analyzer. The response should keep flat when averaged over time.
If your spectrum analyzer does not operate on a constant bandwidth, but constant percentage bandwidth - such as a 1/3 octave spectrum analyzer - use pink noise instead.
White noise can be used to measure the adverse effects of room modes as well, although a low frequency sine sweep will be better for such a purpose.
In healthcare applications, white noise is used to mask tinnitus, a buzzing, ringing, or whistling in your ear, occurring without any stimulus.
If you are interested...
Our white noise sample file has been generated using wavTones' professional grade White Noise Generator.
Is AudioCheck free? Not for me. Your support keeps this site running. Any donation will be rewarded with • uncompressed .wav files downloads for every test • increased durations and sample rates up to 192 kHz in the Tone Gen section • a suggestion box on every page
|← EU residents click here|
If you already are a patron, please log in.