What You Can No Longer Hear

What this page cannot show

Both studies included participants with normal otoscopy and tympanometry, thresholds below 25 dB across speech frequencies and no occupational noise exposure. The curves describe a group selected for normal hearing. Long exposure to industrial noise, concerts or loudspeakers makes the comparison less applicable.

In a clinic, each ear is measured separately with calibrated headphones in a booth. Here you listen with both ears, so the ear with better hearing responds, and the page cannot measure background noise in your room. The tone always falls from loud to quiet; this procedure may estimate a slightly lower threshold than one that alternates levels. Reaction time adds about one and a half decibels at every frequency and has little effect on the shape of the curve.

This is an educational instrument, not a diagnosis or personalised advice. The evidence was checked through 30 July 2026 and has not undergone clinical review; the reference studies are from Czechia and Thailand, with global figures from the World Health Organization. If hearing drops suddenly in one or both ears, seek medical help immediately: it is an emergency. For ringing that will not stop or if speech has become harder to understand in noise, visit an ENT clinic, where hearing is measured with calibrated equipment.

How the figures were computed

Both reference tables were transcribed cell by cell from the published articles, quartiles included, and sit in the archive next to the page. Of the 150 cells transcribed, exactly one carries a median outside its own quartiles: for men aged 25–34 at 9 kHz, the table gives a median of 10 with quartiles of 0 and 5. The rule applied throughout is that a median is pulled inside its own quartiles, so 5 is used there. It is the only intervention on the data.

The curves are interpolated linearly between the published age-band midpoints. The estimate stays between ages 18 and 64, the range covered by the reference. The algorithm finds the age at which the published curve has the smallest squared difference from the shape of your responses after removing the best vertical offset. The two papers label the fourth extended frequency as 11.2 and 11.25 kHz; it is the same audiometric frequency, and the page uses 11.2 kHz throughout.

The coefficient is α in α×(age−18)², fitted separately at each frequency across the medians of the five groups. The lower value at 16 kHz must be read alongside the response rate: for the older groups, the median includes only ears that responded at this frequency.

Cross-check: the Czech curve, read at the band midpoints of the Thai study, differs by 4.2 dB on average across 24 paired cells, with a worst case of 14.1 dB. The oldest Thai band falls outside the Czech range and was left out, because there the comparison would be measuring the difference in age rather than in method.

A synthetic reader whose thresholds match the published curve exactly at the age in the first column, measured through headphones that play the top end quieter or louder. The columns are the assumed tilt at 16 kHz relative to 9 kHz.