noisemoon › blog › all the noise colors, explained
All the noise colors, explained
Most people know white, pink, and brown noise. There are at least three more in regular use — blue, violet, and grey — with their own characters and use cases. Here's the complete tour, with what each one sounds like and where it shows up.
The naming convention
Noise colors are named by analogy with light. White light contains all visible wavelengths in equal amplitude; white noise contains all audible frequencies in equal amplitude. From there the analogy stretches: pink and brown noise have more "warmer" (lower-frequency) energy, blue and violet have more "cooler" (higher-frequency) energy. Grey noise is the odd one out — it's perceptually flat rather than mathematically flat.
The mathematical handle is the spectral slope, measured in decibels per octave. White is 0 dB/oct. Each step toward warmer colors loses 3 dB/oct. Each step toward cooler colors gains 3 dB/oct.
Quick reference
| Color | Slope | Character | Sounds like |
|---|---|---|---|
| Violet | +6 dB/oct | Very high-end | Hissing, sharp, almost painful at volume |
| Blue | +3 dB/oct | High-end heavy | Hiss with sparkle |
| White | 0 dB/oct | Flat across spectrum | Steady rush, TV static |
| Pink | −3 dB/oct | Balanced toward bass | Rainfall, distant waterfall |
| Brown (red) | −6 dB/oct | Bass-heavy | Surf, distant rumble |
| Grey | (perceptually flat) | Equal loudness, not equal energy | Pink-ish but harder to describe |
White noise
Flat across the audible spectrum. Equal energy at every frequency from 20 Hz to 20 kHz. Slope: 0 dB/oct. The "default" reference noise. Used in audio engineering for testing and calibration, in masking systems for offices, and in sleep settings (often heavily filtered to be more pleasant).
Sounds like a steady rush, similar to TV static or an open shower. Brighter than most people initially expect; a common reaction to first hearing pure white is "this is harsher than I thought." That's because most of what's labeled "white noise" elsewhere isn't actually white — we measured 19 popular YouTube "white noise" videos and found they averaged a slope of −9.46 dB/oct, much closer to brown.
Pink noise
Slope: −3 dB/oct. Halves in power for each doubling of frequency. The most popular noise color for sleep, especially infant sleep, because it's balanced enough to mask environmental sounds but soft enough not to feel harsh over many hours.
Sounds like steady rainfall or a distant waterfall. The most commonly recommended starting point in tinnitus sound therapy. Used in audio engineering for room calibration, because the equal-energy-per-octave property matches how human ears parse loudness across the spectrum better than white does.
Brown noise (also "red")
Slope: −6 dB/oct. The deepest of the three classics. Most of the energy sits in the low end; very little high-frequency content. Sounds like surf at the shoreline or a low rumble.
Brown took off on TikTok in 2022-23 as the popular focus pick for ADHD listeners. The grounding character of brown — low, steady, almost physical — reads as comforting rather than activating for many people. More on the ADHD case.
Confusingly, brown noise isn't named after the color. It's named after botanist Robert Brown, who described Brownian motion — the random walk that mathematically generates this slope. The "red" alternative name comes from the analogy with red light being lower-frequency than other visible colors.
Blue noise
Slope: +3 dB/oct. Mirror image of pink — energy increases at higher frequencies. Sounds bright and hissy.
Blue noise rarely shows up as a sleep or focus aid — it's the wrong direction for human-comfort use. Where it earns its keep is in audio engineering and dithering. Adding blue noise to digital audio before quantization spreads the quantization error to high frequencies where the ear is less sensitive, producing a perceptually cleaner result. Most professional audio software includes blue noise dithering as an option.
Violet noise (also "purple")
Slope: +6 dB/oct. Mirror image of brown. Even more high-end heavy than blue. Sounds sharp and hissy at any meaningful volume.
Like blue, violet noise has technical applications rather than human-listening applications. It's used in differentiating signal-processing operations and in modeling certain physical systems. Some tinnitus protocols use narrow-band violet for masking very-low-pitched tinnitus, because the high-frequency contrast from the violet helps the auditory system attenuate the perceived tinnitus pitch by contrast. A niche use.
Grey noise
Grey noise is the conceptual outlier. Instead of being defined by spectral slope, it's defined by equal-loudness contours — the noise is shaped so that every frequency is perceived as equally loud by human ears, even though the energy isn't equally distributed.
This is psycho-acoustically motivated: human hearing isn't equally sensitive at all frequencies. The Fletcher-Munson curves and their modern equal-loudness counterparts (ISO 226) describe how loudness perception varies with frequency. Grey noise inverts that response so the listener perceives uniform loudness across the band.
In practice, grey noise sounds vaguely pink-ish but with a slightly different character. Some hearing-protection and tinnitus protocols use grey noise specifically because of its perceptual flatness. It's less common than the other colors and not all noise apps support it.
What about other named noises?
You'll occasionally see "green noise" (centered around 500 Hz, mid-band), "black noise" (silence with sparse transients), or "orange noise" (mid-band, claimed to disrupt musical pitch). These are less standardised than the main color set and you'll find different definitions across sources. They're real names in some contexts but the specs aren't universally agreed.
Which one do you want?
- For sleep: pink or brown. Pink for general use, brown for the deepest masking.
- For focus / ADHD: brown. Some prefer pink or white for noisier environments.
- For tinnitus: pink as a starting point.
- For masking voices: white. The full high-frequency energy cuts through speech.
- For audio engineering: white for testing, pink for room calibration, blue for dithering.
- For everyday use: pretty much never blue or violet.
NoiseMoon's white, pink, and brown presets cover the common cases. The warm-bright slider lets you dial between them, so the sharp boundary between "pink" and "brown" is more like a continuous spectrum than three discrete options. We don't currently ship blue, violet, or grey — they're niche enough that we haven't prioritised them, but they may show up later.
← back to noisemoon.com