In this guide
White noise has equal average power per equal-width frequency band; pink noise has equal average power per octave, with less power per hertz at higher frequencies. Brown noise tilts further toward low frequencies. Playback equipment and level affect what you actually hear.
White noise and pink noise differ in how their power is distributed across frequency. Ideal white noise has equal average power in equal-width frequency bands. Ideal pink noise has equal average power in each octave, so its power per hertz decreases as frequency rises. Brown noise has an even stronger low-frequency emphasis. These are statistical descriptions; a short recording will not produce a perfectly smooth display.
When auditioned at a comfortable level, white noise often sounds brighter and hissier, pink noise less bright, and brown noise deeper or more muffled. The exact result depends on the source, the speaker or headphones, the room and your hearing. Audacity's noise generator reference describes the three choices.
White noise vs pink noise vs brown noise
| Noise | Ideal spectral pattern | Typical character |
|---|---|---|
| White | Equal average power per equal-width band, such as each 100 Hz interval. | Relatively bright, broadband hiss. |
| Pink | Equal average power per octave; power per hertz falls as frequency increases. | Less bright than white noise, with more relative low-frequency weight. |
| Brown or Brownian | Power per hertz falls more steeply with frequency than pink noise. | More low-frequency emphasis and a deeper, muffled character. |
These are relative comparisons at comparable listening levels. Different generators, filtering and playback systems can change the result. Brown noise is also called Brownian noise; the name describes a mathematical process, not the color of a source.
Why equal bands and equal octaves differ
An octave doubles in frequency. The band from 100 to 200 Hz is 100 Hz wide; the next octave, 200 to 400 Hz, is 200 Hz wide. With ideal white noise, each equal-width 100 Hz band has the same average power, so the wider second octave contains twice the power of the first. With ideal pink noise, the two octaves have the same average power, so the higher octave has less power per hertz.
This distinction matters when reading an analyzer. A fixed-width frequency display and an octave-band display group energy differently. The sound frequency guide explains hertz and octaves; the spectrogram guide covers a time-varying display.
What are the practical uses?
- Audio and acoustic testing: known noise signals can help excite a range of frequencies for measurement. A meaningful measurement also requires appropriate calibration, microphone, analysis settings and interpretation; simply playing noise does not calibrate speakers.
- Masking: a continuous sound can make some other sounds less noticeable, depending on its level, spectrum and the competing sound. It does not remove the source or guarantee privacy, concentration or sleep.
- Editing and sound design: broadband noise can be shaped into textures or used in an experiment. It differs from recording actual room tone, which captures the particular environment around a dialogue edit.
Neither a noise color nor a digital peak setting tells you the actual acoustic level at your ears. Keep listening comfortable, especially for continuous playback. Do not turn up a signal to compensate for frequencies you cannot hear.
How to compare noise types in Audacity
- Open a new project in Audacity so you do not overwrite material in another recording. Choose Generate → Noise.
- Select White, Pink or Brownian, choose a short duration and a modest amplitude, and generate the signal. Repeat on separate tracks or in separate projects for the other types. Audacity documents the options in its Noise manual.
- Listen briefly at a low, comfortable device volume. Compare one track at a time; equal generator amplitude is not a guarantee of equal perceived loudness or equal average power.
- Export a compatible short WAV if you want to inspect it. The Saturalabs spectrum analyzer displays a local FFT snapshot at a chosen time. Compare several positions, and check the axis and analysis settings before drawing conclusions.
A single FFT snapshot of a random signal fluctuates and is not a definitive measurement of its long-term spectral density. An averaged measurement and the right bandwidth are needed for a meaningful comparison. The analyzer does not generate noise, calibrate speakers or measure sound pressure.
Is white noise the same as background noise?
No. In everyday speech, people may call a fan or hiss “white noise,” but a real recording can have hum, tones, changing levels or a non-flat spectrum. A sample that sounds steady is not proof of an ideal white-noise spectrum. For cleanup, first listen to a section with and without speech; noise reduction can also alter wanted audio. The background-noise guide helps choose an approach.
Common questions
Which noise is best for sleep or focus?
Preferences and situations differ. These color labels describe frequency balance, not a proven outcome for every listener. Compare at a comfortable level and avoid making medical claims based on the label.
Does pink noise contain only low frequencies?
No. Pink noise is broadband: it includes higher frequencies too, but has less power per hertz at higher frequencies than at lower frequencies.
Is a tone the same as noise?
A steady sine tone concentrates energy at one frequency; broadband noise distributes energy across many frequencies. The tone generator creates a sine tone, not white, pink or brown noise.