Brown vs pink vs white noise
The short answer: the three colors are the same randomness wearing different spectra. White keeps equal energy at every frequency (0 dB/octave) and sounds bright; pink halves its energy every octave (−3 dB/octave) and sounds like rain; brown quarters it (−6 dB/octave) and sounds like a deep rumble. Which one is "best" depends on the sounds you are masking and the ears doing the listening — so this page ends with a mixer that plays all three.
Side by side
| White | Pink | Brown | |
|---|---|---|---|
| Spectral slope | 0 dB per octave — flat | −3.01 dB per octave | −6.02 dB per octave |
| Power in each higher octave | The same | Half (×½) | A quarter (×¼) |
| How it is made | Random samples | White noise through Kellet's filter | White noise through a leaky integrator |
| Sounds like | Untuned radio, fan on full, air vent | Steady rain, wind in trees | Distant thunder, heavy rain on a roof, cabin hum |
| People reach for it to mask | Voices, clatter, TV through a wall | Mixed household sounds, light traffic | Traffic rumble, bass, HVAC and aircraft hum |
| Common complaint | "Too hissy to sleep to" | "A bit plain" — which is also its virtue | "Too muffled; voices poke through" |
Why the slope changes everything you hear
Hearing is logarithmic: the step from 100 Hz to 200 Hz feels the same size as the step from 5,000 to 10,000 Hz, even though the second spans forty times the bandwidth. A flat spectrum (white) therefore dumps most of its energy into the octaves you are most sensitive to, and sounds bright. Pink compensates exactly for the widening octaves, so every pitch region gets an equal share — the balanced, rain-like middle. Brown over-compensates: nearly all of its energy is bass, which is why it can feel more like a physical presence than a sound, and why it masks low rumbles that white and pink barely touch.
One thing the colors share on this site: loudness. Every generator is normalised to the same RMS level (0.22), so differences you hear between the players are differences of spectrum, not of volume. Plenty of noise apps quietly make one color louder; we would rather you compared them fairly.
Decide by ear
Start with brown alone. If voices or clatter still poke through, add pink in 10% steps until they soften. If the mix feels thick or sealed, trade some brown for white instead. There is no wrong answer — the right mix is the one you stop noticing.
Press play to compare in place — brown underneath, a layer of pink on top. Take pink to zero for pure brown; take brown down and white up to walk toward the bright end.
No timer set — playback continues until you pause it. With a timer, the sound fades out gradually at the end instead of cutting off.
Questions
- Which color is best overall?
- None, in general. Brown wins for sleep and low rumble, white for masking speech and clatter, pink as the compromise most people can leave on all day. Anyone claiming one color is universally superior is selling something.
- Can I mix colors?
- Yes — mixing is often better than any single color, because real rooms contain both low rumble and high clatter. The mixer above (and on the homepage) exists for exactly that.
- Are there other colors — blue, violet, grey?
- They exist as definitions (blue rises +3 dB/octave, violet +6), but they are specialist tools for audio testing rather than listening comfort, so this site sticks to the three colors people actually sleep and work to.
- Do the spectra here match the diagrams?
- The diagrams are idealised straight lines from the definitions. The real generators were measured offline at 44.1 kHz and land within about a decibel of those lines across the audible octave bands (brown's slope levels off progressively toward the very bottom, as the brown noise page explains).