What hushnap’s Sounds Look Like: Frequency Charts for All 15
We measured every sound in the player and drew its frequency chart, so you can see why brown noise feels deep, why shushing sounds bright, and how they all differ.
By hushnap.com · Measured
How to read these charts
Each chart shows how loud a sound is at each pitch. Low, rumbly frequencies are on the left, high hissy ones on the right. The line is drawn relative to the sound’s own loudest point (0 dB at the top), so a line that drops toward the bottom means that range is much quieter in that sound. A flat line means every pitch is about equally strong.
These are not loudness readings for your room. They show the shape of the signal hushnap sends to your speaker, not how loud it is. Your speaker and room change what reaches the crib, so to check the actual level, follow our guide to measuring with your phone.
White, pink and brown noise, side by side
The three “colors” of noise differ only in how quickly they fade from low to high pitch. In the textbook definitions, white noise stays flat, pink noise drops by about 3 dB for every octave you go up, and brown noise drops by about 6 dB per octave. Our guide to the noise colors explains what they are and what sleep research has (and hasn’t) found about them.
Here is how hushnap’s versions measure, fitted between 100 Hz and 4 kHz:
| Sound | Textbook slope | Measured slope |
|---|---|---|
| White noise | 0 dB per octave | 0.1 dB per octave |
| Pink noise | about −3 dB per octave | -2.9 dB per octave |
| Brown noise | about −6 dB per octave | -5.2 dB per octave |
White and pink land almost exactly on target. Brown comes out a little shallower than −6 because of the gentle pull back toward zero we build into it. That keeps the random walk from wandering off, but it also levels the curve off below about 150 Hz. Above a few hundred hertz it falls the full 6 dB per octave, so it still sounds and measures like brown noise. All three dip at the far right because we deliberately soften the highest frequencies so none of them sound harsh.
Weather & nature
Rain
A steady bed of pink noise with the highs rolled off above 3.8 kHz, plus about ten tiny droplets a second. Each droplet is a burst of noise lasting 8 to 28 milliseconds, filtered to a random pitch between 1.8 and 5.3 kHz, fired at uneven intervals so it never ticks like a clock.
Strongest at the deepest end of the range · center of energy 1.0 kHz · energy below 250 Hz 42.7%, 250 Hz to 2 kHz 37.5%, above 2 kHz 19.8%
Waves
Brown noise whose brightness sweeps slowly up and down (the filter moves between about 200 Hz and 1.2 kHz on an 11 second cycle) while the volume swells on a 14 second cycle. A faint layer of hiss above 2.5 kHz rises and falls on its own 20 second cycle, like foam on the crest.
Strongest at the deepest end of the range · center of energy 478 Hz · energy below 250 Hz 70%, 250 Hz to 2 kHz 27.4%, above 2 kHz 2.6%
Wind
White noise with most of the sound above roughly 500 to 700 Hz filtered away. The filter’s cutoff, its sharpness and the volume each drift on their own slow cycles (about 9, 6 and 8 seconds), which is what makes it gust.
Strongest around 250 Hz · center of energy 433 Hz · energy below 250 Hz 34.1%, 250 Hz to 2 kHz 65.4%, above 2 kHz 0.5%
Storm
A deep rumble of brown noise below about 380 Hz that swells every 20 seconds, a quiet layer of rain-like hiss underneath, and sharp crackles of high-pitched noise lasting 0.4 to 0.9 seconds every 1 to 3 seconds.
Strongest at the deepest end of the range · center of energy 528 Hz · energy below 250 Hz 70.5%, 250 Hz to 2 kHz 24.4%, above 2 kHz 5.1%
Noise colors
White noise
Random samples, each one independent of the last, with the very top end gently rolled off above 9 kHz so it isn’t harsh.
Strongest around 6.3 kHz · center of energy 5.9 kHz · energy below 250 Hz 1.8%, 250 Hz to 2 kHz 13.5%, above 2 kHz 84.8%
Pink noise
White noise passed through a standard set of filters (Paul Kellet’s pink-noise approximation) so that each octave carries about the same energy, then rolled off above 9 kHz.
Strongest at the deepest end of the range · center of energy 2.0 kHz · energy below 250 Hz 38.3%, 250 Hz to 2 kHz 31.5%, above 2 kHz 30.2%
Brown noise
A random walk: each sample nudges the last one up or down a little, with a gentle pull back toward zero so it never drifts away. Rolled off above 8 kHz.
Strongest at the deepest end of the range · center of energy 481 Hz · energy below 250 Hz 62.7%, 250 Hz to 2 kHz 32.3%, above 2 kHz 5%
Household
Fan
White noise filtered down to below about 1.1 kHz for the rush of air, plus a quiet 110 Hz motor tone that wobbles slightly in pitch.
Strongest around 111 Hz · center of energy 855 Hz · energy below 250 Hz 14.3%, 250 Hz to 2 kHz 82%, above 2 kHz 3.7%
Hairdryer
White noise squeezed into a band around 2.8 kHz, with the center of the band wobbling a little about once a second, plus a faint 180 Hz motor tone.
Strongest around 2.8 kHz · center of energy 4.5 kHz · energy below 250 Hz 0.4%, 250 Hz to 2 kHz 11.9%, above 2 kHz 87.7%
Vacuum
Brown noise squeezed into a band around 260 Hz, a buzzy 95 Hz motor tone, and a thin whine of hiss above 3.5 kHz.
Strongest around 99 Hz · center of energy 1.5 kHz · energy below 250 Hz 49.5%, 250 Hz to 2 kHz 39.1%, above 2 kHz 11.4%
Car ride
Deep brown noise below about 240 Hz for road rumble, a low 70 Hz engine tone, and a gentle bounce in volume a little over twice a second.
Strongest around 70 Hz · center of energy 129 Hz · energy below 250 Hz 91%, 250 Hz to 2 kHz 9%, above 2 kHz 0%
Womb & soothing
Heartbeat
Two soft, quickly fading low thumps per beat, a “lub” at 55 Hz and a softer “dub” at 50 Hz, synthesized directly as tones. At the player’s setting the pattern repeats about 69 times a minute.
Strongest around 56 Hz · center of energy 55 Hz · energy below 250 Hz 100%, 250 Hz to 2 kHz 0%, above 2 kHz 0%
Womb
Very deep brown noise below about 300 Hz with a volume pulse about 69 times a minute. It is a simple impression of a muffled, rhythmic whoosh, not a recording from inside a womb.
Strongest at the deepest end of the range · center of energy 152 Hz · energy below 250 Hz 82.5%, 250 Hz to 2 kHz 17.5%, above 2 kHz 0%
Shush
White noise squeezed into a band around 2.1 kHz, with its volume shaped by two slow, unevenly timed cycles so it rises and falls like breath.
Strongest around 2.2 kHz · center of energy 3.9 kHz · energy below 250 Hz 0.1%, 250 Hz to 2 kHz 24.6%, above 2 kHz 75.3%
Hum
Two pure tones, 196 Hz (the note G) with a slight vocal-style vibrato and a quieter 294 Hz (D) above it, breathing in and out about every 4.5 seconds.
Strongest around 198 Hz · center of energy 213 Hz · energy below 250 Hz 82.9%, 250 Hz to 2 kHz 17.1%, above 2 kHz 0%
All 15 at a glance
| Sound | Center of energy | Below 250 Hz | Above 2 kHz |
|---|---|---|---|
| Heartbeat Deep | 55 Hz | 100% | 0% |
| Car ride Deep | 129 Hz | 91% | 0% |
| Womb Deep | 152 Hz | 82.5% | 0% |
| Hum Deep | 213 Hz | 82.9% | 0% |
| Wind Mid-range | 433 Hz | 34.1% | 0.5% |
| Waves Deep | 478 Hz | 70% | 2.6% |
| Brown noise Deep | 481 Hz | 62.7% | 5% |
| Storm Deep | 528 Hz | 70.5% | 5.1% |
| Fan Mid-range | 855 Hz | 14.3% | 3.7% |
| Rain Broad | 1.0 kHz | 42.7% | 19.8% |
| Vacuum Broad | 1.5 kHz | 49.5% | 11.4% |
| Pink noise Broad | 2.0 kHz | 38.3% | 30.2% |
| Shush Bright, hissy | 3.9 kHz | 0.1% | 75.3% |
| Hairdryer Bright, hissy | 4.5 kHz | 0.4% | 87.7% |
| White noise Bright, hissy | 5.9 kHz | 1.8% | 84.8% |
“Center of energy” (the spectral centroid) is the average pitch of a sound, weighted by how much energy sits at each pitch. It’s a simple way to rank sounds from deepest to brightest.
Does a deeper or brighter sound help a baby sleep better?
These charts describe the sounds. They don’t show which one works best for sleep, and no research we know of settles that for babies. Most of the studies on noise colors were done in adults, and we found no sleep studies on brown noise at all. Our guide to the best sounds for newborn sleep covers what has actually been tested. In practice, the right sound is the one that settles your baby at a low, safe volume.
You can try any of these in the player and keep the volume low. The loudness guide explains why.
How we measured
- Each sound was generated by the same code the player uses, running in the Web Audio engine of a headless Chromium browser at 48 kHz.
- We captured 40 seconds of each sound after a short warm-up, taking an FFT of 16,384 samples about every 100 ms, and averaged the power across all frames. Sounds that change over time, like waves and storm, are shown as their average.
- Levels are grouped into 1/6-octave bands from 20 Hz to 20 kHz and shown relative to each sound’s loudest band. The chart is clipped at −60 dB.
- The random number generator was seeded, so re-running the measurement gives nearly the same result (within about 0.1 dB) unless the sounds themselves change.
- This is the digital signal before your speaker. Speakers, and small ones in particular, don’t reproduce every frequency equally, and rooms add their own echoes, so what you hear will differ from the chart.
The measurement script is scripts/measure-spectra.mjs in our code, and the full data is available as a JSON file.
Using these charts and data
You’re welcome to use the charts and the data, in a blog post, an article, a class or your own project. They’re shared under the Creative Commons Attribution 4.0 license (CC BY 4.0). That means you can copy, adapt and share them, including commercially, as long as you credit hushnap.com and link to this page.
A credit line like this is perfect: Source: hushnap.com sound spectra, https://hushnap.com/sounds. If you use them somewhere interesting, we’d love to hear about it.