A normal hearing test does not rule out hearing loss: standard audiograms stop at 8 kHz, and a 2022 meta-analysis found people with tinnitus and normal audiograms had, on average, poorer hearing from 10 to 18 kHz.
Tinnitus with a normal audiogram is the less common situation: in one Chinese hospital series of 500 people with chronic tinnitus, 96.6% had hearing loss on a standard test. For the minority whose test comes back clean, one candidate explanation is hidden hearing loss, damage to nerve connections that a threshold test cannot show. This page covers a simpler one: ordinary hearing loss at pitches the standard test never plays.
Why the standard test stops at 8 kHz
A standard audiogram tests pitches from 250 Hz to 8,000 Hz, the range how to read your audiogram walks through. Young, healthy ears hear well above that, up to 20 kHz. The band from 9 to 20 kHz is called the extended high frequencies, and routine clinical testing leaves it out; a 2020 review described it as largely ignored by clinicians and researchers alike.
That omission matters because of where those pitches are handled. The highest frequencies are processed at the base of the cochlea, and a 2022 review in The Journal of the Acoustical Society of America describes that region as especially sensitive to ageing, disease, ototoxic drugs and possibly noise. Its authors suggest a loss there may be an early warning of damage.
What testing above 8 kHz finds in people with tinnitus
The best summary is a 2022 meta-analysis in Ear and Hearing. It screened 261 articles and found nine studies comparing adults who had tinnitus and a normal audiogram with control groups. The tinnitus groups had significantly worse hearing at 10, 12.5, 14, 16 and 18 kHz. Heterogeneity was below 50% throughout, meaning the studies broadly agreed, and there was no sign of publication bias. The authors conclude that in most cases tinnitus is associated with some degree of cochlear mechanical dysfunction that conventional audiometry alone may not detect.
The individual studies fill in the picture:
- Young adults. A 2021 case-control study compared 28 people aged 18 to 35 with tinnitus and normal standard audiograms against 34 of the same age without tinnitus. Extended high-frequency loss was more common with tinnitus, and hearing began to fall away at a lower frequency: 10.4 kHz against 12.3 kHz.
- Tinnitus on one side. A 2024 study of 96 people with normal audiograms found that those with one-sided tinnitus had higher thresholds than controls at every frequency tested above 8 kHz, and differences between their two ears at 11.2, 12.5 and 14 kHz. A 2015 study of 75 patients at a Regensburg tinnitus clinic likewise linked the side of the tinnitus to asymmetric high-frequency hearing.
- Several findings at once. A 2022 Korean clinic review of people with tinnitus and normal hearing found extended high-frequency loss in 35.4%. Reduced tolerance of loud sound was more common still, and 75.6% had at least one of the findings the clinic looked for, a list that included psychiatric symptoms.
Not every study agrees, and the finding has limits
One study found nothing in the extended range. A 2022 study in Acta Oto-Laryngologica compared 102 people with tinnitus and normal standard audiograms against age- and sex-matched controls and found no significant difference between 9 and 14 kHz. Instead, thresholds between 2 and 8 kHz were significantly higher in the tinnitus ears, though still inside the normal range, and the authors concluded that extended high-frequency testing appears unnecessary for patients with normal hearing. Read beside the meta-analysis, it still points to differences that a "normal" label can hide, only at a different place on the chart.
It is an association. Every study here compares groups at one point in time. That shows extended high-frequency loss and tinnitus tend to occur together, not that one causes the other; the same noise exposure or ageing could plausibly produce both.
It is common in people with normal audiograms generally. In a 2019 study in PNAS, 74 of 116 mostly young adults with "normal hearing" had some hearing loss above 8 kHz; the study did not report whether any of them had tinnitus. A loss on your chart is consistent with your tinnitus, but it does not prove it is the explanation.
Studies disagree on whether it tracks how bothersome tinnitus is. A 2024 study of 38 patients with tinnitus and normal standard hearing found no relationship between Tinnitus Handicap Inventory scores and extended high-frequency thresholds. The Regensburg study, by contrast, found higher questionnaire scores in patients with abnormal high-frequency results. In the Korean review, psychiatric symptoms were the only finding independently associated with tinnitus distress.
How this relates to hidden hearing loss
The two ideas are often run together, and they are different claims:
- Extended high-frequency loss is ordinary hearing loss, a raised threshold, at pitches the standard test does not play. A suitable audiometer measures it directly.
- Hidden hearing loss, or cochlear synaptopathy, is loss of the connections between the inner ear's hair cells and the auditory nerve. It is thought not to affect thresholds at all, which is why a threshold test cannot show it directly.
They can travel together. In a 2016 Boston study, college students at high risk from noise had normal standard thresholds but worse hearing from 10 to 16 kHz than a low-risk group. They also showed an electrical response pattern consistent with selective nerve loss, did worse at recognising words in noise, and reported heightened reactions to sound.
They can also come apart. A 2026 study in Ear and Hearing looked at adults with normal hearing to 8 kHz and, on average, above it. In them, tinnitus was associated with reduced auditory nerve responses consistent with nerve fibre loss, and with raised gain at the cochlear nucleus, the first brainstem stop for those signals. And a 2026 study of 60 adults with normal audiograms, not selected for tinnitus, found raised extended high-frequency thresholds went with subtle outer hair cell weakening at 1 to 4 kHz rather than with synaptopathy.
Sometimes neither shows up. A 2017 Manchester study matched young adults with and without tinnitus for hearing all the way up to 14 kHz; tinnitus was still linked to greater lifetime noise exposure, but not to the nerve-response measures synaptopathy should affect.
So a result above 8 kHz is one window into the cochlea, not a full account. The 2022 review makes a similar point about speech: a loss there may be a marker for sub-clinical damage at lower frequencies rather than a cause in its own right. Tinnitus and hearing loss covers a third window, the sounds the outer hair cells themselves produce.
Noise is one common thread. A 2025 meta-analysis in Trends in Hearing reviewed 30 studies covering about 2,500 adults with normal audiograms. Occupational noise was associated with worse thresholds from 9 to 16 kHz; recreational noise showed a less consistent link. The authors found some publication bias at 10 kHz and say longitudinal studies measuring noise dose are needed to establish cause. Age is another thread, covered below. Noise-induced tinnitus covers the exposure side.
What a result above 8 kHz can and cannot change
It does not unlock a different tinnitus treatment. None of the studies above tested a treatment aimed at extended high-frequency loss. The 2014 AAO-HNS tinnitus guideline recommends cognitive behavioural therapy for people with persistent, bothersome tinnitus, and a hearing aid evaluation where hearing loss is documented.
What it can offer:
- A possible explanation. A measured loss above 8 kHz is a concrete, physical finding, consistent with the meta-analysis's reading that tinnitus is usually associated with some cochlear dysfunction.
- A baseline. A 2025 study retested 71 adults aged 19 to 38 with normal audiograms about 24 months apart. Thresholds above 8 kHz shifted faster than standard ones, and a worse starting threshold above 8 kHz was associated with faster decline in the standard range. That makes such a loss a reason to take hearing protection seriously and to keep the result for comparison.
- Monitoring during ototoxic treatment. ASHA's 1994 guidelines on drugs that damage the cochlea call high-frequency audiometry the method of choice for the earliest detection of ototoxic hearing loss, normally testing from 9 to 20 kHz. Ototoxic medications covers which drugs are involved, and ototoxicity monitoring covers the test schedule and grading scales.
Should you ask for it, and can you get it?
It is reasonable to ask. At a US tertiary referral centre, 25 of 40 patients (62.5%) with various hearing-related symptoms and normal standard audiograms had hearing loss at the highest frequencies, and those with the loss were older on average; the abstract does not say how many had tinnitus. The authors suggest extended high-frequency testing may be considered when standard audiometry is normal, while noting that more data are needed before there is an evidence-based clinical algorithm for it. The authors of the 102-person study reached the opposite view. Neither is a guideline.
Availability is the practical obstacle. According to the 2022 review, the test presents no particular difficulty in the clinic; the biggest obstacle is a lack of specialist equipment, and clinical use is limited but increasing. ASHA notes that the frequencies tested vary slightly with the audiometer used. Asking when you book whether a clinic can test above 8 kHz saves a wasted trip.
Know the limits before you read a result:
- Age matters far more above 8 kHz. A 2024 Thai study of 134 adults with normal standard hearing and no tinnitus found median thresholds of 20 dB HL or better at most extended frequencies up to age 40 (30 dB HL at 16 kHz). In those over 40 and up to 70, medians ran from 20 dB HL at 9 kHz to 70 dB HL at 14 and 16 kHz. The international standard for age-related thresholds, ISO 7029, stopped at 8 kHz when a 2014 study set out to extend it, so age-matched reference values above 8 kHz are still being established.
- It measures thresholds, not nerve connections. A normal result above 8 kHz does not exclude the nerve damage described above.
- It cannot say why. A loss there fits with tinnitus, but it is also common in people without it, as the tinnitus-free Thai volunteers over 40 show.
When a normal result is not the end of it
Some features call for more assessment whatever the audiogram shows. The 2014 AAO-HNS guideline asks for a prompt, comprehensive hearing assessment when tinnitus is in one ear, has lasted six months or more, or comes with hearing difficulties. Its recommendation against scans applies specifically to tinnitus that is not one-sided, not pulsing, and comes without neurological signs or hearing that differs between the ears, so any of those features puts someone outside it. A sudden drop in hearing is different again: the US National Institute on Deafness and Other Communication Disorders says to treat it as a medical emergency and see a doctor immediately. When to seek care sorts these by urgency, and pulsatile tinnitus, tinnitus in one ear and sudden hearing loss cover them in turn.
If your main difficulty is following speech in noisy places, that is a separate question. The 2019 PNAS study found extended high-frequency loss correlated with self-reported difficulty hearing in noise. The 2022 review notes that link may not be causal; if it is, amplifying these pitches might help, provided the technical difficulties can be overcome. The digits-in-noise test probes that listening task directly.
If the conversation stopped at "your hearing is normal", what the assessment involves and being told nothing is wrong cover how to take it further. Tinnitus with a normal hearing test is real, and it has a research literature of its own.