Hearing Frequency Test, Explained

After reading this you will know how a high-frequency hearing test brackets your personal ceiling, how an audiogram-style threshold sweep reads across the octave bands, and how to tell a real hearing limit from a limit imposed by your speakers.

What the test measures and one quick example

Sound is pressure that oscillates. The rate of oscillation is the frequency, measured in hertz (Hz) or thousands of hertz (kHz). A young, healthy ear responds to roughly 20 Hz at the low end up to about 20000 Hz at the top. The top of that range is what drifts down with age.

Here is the hook. Play a tone at 15 kHz. Most people under 30 hear a thin, high whistle. Many people over 50 hear nothing at all, even at the same volume. Nothing changed about the tone. What changed is the listener. This test finds the exact frequency where that transition happens for you, then compares it to typical values for your age.

The tool has two modes. The first climbs upward from 8 kHz to find your ceiling. The second holds a fixed set of frequencies (the octave bands from 125 Hz to 8 kHz) and asks how quiet each one can get before you lose it. Together they sketch a rough hearing profile.

When to use it, and when not to

Use this for curiosity, for a classroom demonstration of presbycusis, or to compare devices and codecs. It is a good, honest illustration of how the highest frequencies fade first. It also makes a memorable group activity: play a rising tone to a mixed-age room and watch who stops reacting.

Do not use it as a diagnosis. A clinical audiogram runs in a sound-treated booth with calibrated transducers, and the technician knows the exact sound pressure level at your eardrum. This test knows none of that. Your browser controls a volume number, not a decibel level, and your hardware colors everything.

If you notice ringing, muffled hearing, or a sudden drop in one ear, stop testing and see an audiologist. Sudden one-sided hearing loss is treated as urgent. A browser game is not the place to work that out.

The math of pitch and the octave spacing

Pitch is logarithmic, not linear. Doubling the frequency raises the pitch by one octave, and your ear treats each octave as a roughly equal step. That is why the audiogram bands double each time: 125, 250, 500, 1000, 2000, 4000, 8000 Hz.

f_n = f_0 \cdot 2^{n}

Here f_0 is a starting frequency, n is the number of octaves above it, and f_n is the result. Start at f_0 = 125 Hz. Two octaves up is 125 \cdot 2^{2} = 500 Hz. Six octaves up is 125 \cdot 2^{6} = 8000 Hz. The seven test bands are exactly the integers n = 0 through n = 6.

The ceiling search uses a finer step. It multiplies by a fixed ratio each time rather than doubling. If each step raises the frequency by 3 percent, then

f_{k} = 8000 \cdot 1.03^{k}

where k counts steps above 8 kHz. Ten steps reach 8000 \cdot 1.03^{10} \approx 10750 Hz. Forty steps reach about 8000 \cdot 1.03^{40} \approx 26100 Hz, past any human ceiling. Small ratio steps give fine resolution near your limit.

How the ceiling search brackets your limit

The search does not test every frequency. It uses bracketing: keep a frequency you can hear and a frequency you cannot, then narrow the gap between them. This is the same idea as a binary search, and it converges fast.

Suppose you hear 14 kHz but not 18 kHz. Your true ceiling sits somewhere in that 4 kHz window. Test the midpoint, 16 kHz. If you hear it, the window shrinks to 16 to 18 kHz. If not, it shrinks to 14 to 16 kHz. Each answer halves the remaining range.

w_m = w_0 \cdot \left( \frac{1}{2} \right)^{m}

Here w_0 is the starting window width, m is the number of answers you have given, and w_m is the width that remains. Starting from a 4 kHz window, after m = 4 answers the window is 4000 \cdot (1/2)^{4} = 250 Hz wide. Six answers bring it under 63 Hz. You bracket a sharp ceiling in well under ten clicks.

Reproducing the demo run

Load the tool with its default settings and follow the ceiling search as a 35-year-old with typical hearing might answer it.

Start at 8 kHz. Clearly audible, so answer "I hear it". The lower bound becomes 8 kHz.
Jump up to 16 kHz. Still audible, thin but present. Lower bound moves to 16 kHz.
Jump to 18 kHz. Nothing. Upper bound becomes 18 kHz. Window is now 16 to 18 kHz, width 2 kHz.
Midpoint 17 kHz. Faintly audible. Lower bound moves to 17 kHz. Width 1 kHz.
Midpoint 17.5 kHz. Nothing. Upper bound 17.5 kHz. Width 500 Hz.
Midpoint 17.25 kHz. Barely there. Report the ceiling as about 17.3 kHz, plus or minus 0.25 kHz.

Six answers. The reported ceiling of 17.3 kHz sits right in the expected band for a listener in their mid-thirties. Compare that to the reference table below.

Typical ceilings by age

Presbycusis is gradual and starts at the top of the range. The numbers below are rough averages for people without noise damage. Individuals vary widely, so treat these as a place to stand, not a rule.

Approximate highest audible frequency by age group
Age group	Typical ceiling (kHz)	Notes
Under 20	18 to 20	Full range still intact for most
20 to 29	17 to 18	Slight top-end loss begins
30 to 39	16 to 17	Demo run lands here
40 to 49	14 to 16	Noticeable drop above 15 kHz
50 to 59	12 to 14	Many lose 15 kHz entirely
60 and up	8 to 12	High-frequency loss widens

The curve slopes down steadily. The vertical marker at age 35 lines up with the demo ceiling of about 17 kHz.

Reading the audiogram-style profile

The second mode does something different. Instead of finding the top frequency, it holds each octave band and lowers the level until the tone is barely audible. The quietest level you can still detect is your threshold for that band. Plotting thresholds across the seven bands gives a profile shaped like a clinical audiogram, though without calibrated decibels.

Because the tool cannot know true sound pressure, read the profile as relative, not absolute. A flat line across bands means every frequency reaches you about equally. A line that dips at the right end (the high frequencies need more level to be heard) is the classic early sign of age-related or noise-related loss.

A typical mild high-frequency pattern: the 8 kHz bar sits far above the rest, meaning that band needs much more level before it becomes audible.

Clinical audiograms plot loss downward, so a lower point means worse hearing. This tool plots the level you needed, so a higher bar means worse hearing. The shape carries the meaning either way: watch which bands stand apart from the rest.

Watch the effect of a hardware roll-off

The single most instructive thing to explore is why an "I can't hear it" can be your gear. Speakers and codecs cut high frequencies at some cutoff, and above that cutoff no tone reaches you no matter how good your ears are. Move the cutoff and watch your measured ceiling get capped.

Common mistakes that ruin the result

The biggest error is starting loud. A blasting tone is unpleasant and it temporarily raises your thresholds, so the next few readings come out worse than your true hearing. Start quiet and creep up.

The second error is trusting laptop speakers. Small drivers often roll off above 15 to 17 kHz, and Bluetooth codecs can cut the top end to save bandwidth. Wired headphones over a direct connection give the most honest ceiling. The widget above shows exactly how a low cutoff caps your number.

The third error is expectation bias. If you know a tone is playing, you may click "I hear it" for a tone you cannot actually hear. Look away from the frequency readout while you decide, and if a band feels ambiguous, mark it as not heard. A conservative answer keeps the profile honest.

Related tests on this site

This test sits with the other perception and reflex tools. For a hearing-plus-reflex combination, try the Reaction Time Test, which times how fast you respond to a change. For vision instead of hearing, the Contrast Sensitivity Test uses the same barely-audible logic with faint patterns, and the Color Discrimination Test shrinks a color difference until you miss it. If you enjoyed the timing side of tones, the Tempo & Rhythm Tapping Test measures how steadily you can tap a beat.

Frequently asked questions

Why can't I hear 18 kHz anymore?

The highest frequencies fade first with age, a process called presbycusis. A ceiling near 16 to 17 kHz in your thirties is typical and not a cause for concern. Noise exposure speeds the same drift.

Does the volume slider mean decibels?

No. It sets a digital level your device turns into sound at an unknown physical loudness. That is why the audiogram mode reports a relative profile and not calibrated dB values.

Why did my two ears score differently?

Real hearing often differs slightly between ears, and headphone channels can differ too. A large gap between ears is worth mentioning to an audiologist, but small differences are normal.

Can this test damage my hearing?

Not at sensible volumes. High-frequency tones can be piercing, so the tool warns you first and asks you to start quiet. Keep the level comfortable and limit long full-level exposure.

Why is my ceiling exactly the same on two different devices?

If both devices report the same round number like 15 kHz, suspect a shared hardware or codec cutoff rather than a true ear limit. Test with wired headphones to check.

Try it yourself in the tool