Pure-tone audiometry is the simplest possible clinical test of the auditory system: present a pure sinusoid at a controlled level and frequency, and ask the patient to indicate when they hear it. The lowest level at which they reliably respond is their threshold at that frequency. Repeating at a small set of standard frequencies — typically 250, 500, 1000, 2000, 4000, and 8000 Hz — produces the audiogram.
The audiogram is to audiology what blood pressure is to cardiology: a single measurement that takes about ten minutes to collect, but whose interpretation organises most subsequent clinical decisions. This lesson develops the test, the chart, and the basic conventions for recording thresholds.
The test
The instrument is the audiometer — typically a small box with two earphones, a bone-conduction oscillator, a tone-generator dial calibrated in steps of 5 dB, a frequency selector, and a patient response button.
The procedure (clinical Hughson–Westlake / “modified Hughson–Westlake”):
Present a tone at a clearly audible level (typically 30 dB above the expected threshold).
After each patient response, decrease the level by 10 dB until they fail to respond.
Then increase by 5 dB until they respond again.
The threshold is the lowest level at which the patient responds to at least 50% of presentations on ascending trials. (Most clinics use 2-out-of-3 or 3-out-of-5 as the operational rule.)
Repeat at each test frequency in both ears.
The whole procedure for an air-conduction audiogram in one ear takes about five minutes. Adding the other ear, plus bone-conduction thresholds (see 2.3), and any required masking, brings a full audiometric evaluation to about 20–30 minutes.
What’s being measured here is behavioural sensitivity — the patient must consciously decide they heard the tone and press a button. This is fast, cheap, and informative, but it requires a cooperative adult who can attend reliably. Children and patients who can’t reliably respond require the objective tests of Ch 6.
The chart
By convention the audiogram is drawn with frequency on a log scale on the horizontal axis (250 Hz to 8 kHz, left to right) and hearing level on a downward-increasing linear scale on the vertical axis (0 dB HL at top, 110 dB HL at bottom). Better hearing thresholds therefore plot higher on the page — a deliberate ergonomic choice that makes more-impaired audiograms drop down visually.
Symbols follow a colour and shape convention nearly universal in clinical audiology:
Ear
Mode
Symbol
Colour
Right
Air conduction, unmasked
O
red
Right
Air conduction, masked
△
red
Right
Bone conduction, unmasked
⟨
red
Right
Bone conduction, masked
[
red
Left
Air conduction, unmasked
×
blue
Left
Air conduction, masked
□
blue
Left
Bone conduction, unmasked
⟩
blue
Left
Bone conduction, masked
]
blue
The “right is red, left is blue” mnemonic is rooted in maritime navigation lights and survives in audiology by historical accident; the symbol choices are codified in the ASHA / American Academy of Audiology standards.
preset:
Drag the threshold symbols up and down to enter pure-tone thresholds. The sidebar auto-categorises the audiogram by degree (using the pure-tone average), shape (by inspecting the slope and any notches), and type (sensorineural / conductive / mixed, determined from the air–bone gap and the bone-conduction levels). Real clinical interpretation also weighs immittance results, OAEs, and patient history — but the audiogram alone often suggests the diagnosis the rest of the workup confirms.
Drag the threshold symbols to enter pure-tone thresholds for each ear at each test frequency. The chart auto-categorises the audiogram by degree (using the pure-tone average), by shape (slope, presence of notches, mid-frequency depression), and by type (sensorineural, conductive, mixed — determined from the air-bone gap, developed in 2.3). Try the presets to see canonical configurations: a flat conductive loss, a high-frequency sloping loss typical of presbycusis, a 4-kHz notch from noise exposure, a mid-frequency “cookie-bite” loss that often points to a genetic etiology.
Degree of hearing loss
The audiogram is the substrate; quantitative descriptors are read off it.
The most common single-number summary is the pure-tone average (PTA) — the arithmetic mean of thresholds at 500, 1000, and 2000 Hz, which are the frequencies most heavily weighted for speech. The PTA is used to assign a degree of hearing loss:
PTA (dB HL)
Degree
≤ 25
normal
26–40
mild
41–55
moderate
56–70
moderately severe
71–90
severe
≥ 91
profound
These cutoffs are clinical conventions, not biological discontinuities. A patient with a PTA of 27 is in the “mild” category but functions essentially normally for one-on-one conversation; a patient with a PTA of 39 is at the upper end of “mild” but already misses considerable speech information in noise. The categories are useful for record-keeping, candidacy criteria (e.g., cochlear-implant rules used to be tied to specific PTA thresholds), and insurance coding. The actual functional impact of a given audiogram depends on the configuration, the patient’s communication demands, the listening environment, and a host of cognitive and social factors that the audiogram alone does not predict.
Configuration
The shape of the audiogram also matters. Standard descriptors:
Flat — thresholds within ~20 dB across the test frequencies.
Sloping — thresholds worsen as frequency increases. The most common configuration in adult-onset hearing loss; characteristic of presbycusis and noise-induced damage.
Rising — thresholds worsen as frequency decreases. Less common; sometimes seen in Ménière’s disease (early stages) or in certain conductive losses.
Notched — thresholds are dramatically worse at one frequency than at adjacent ones. The classic “noise notch” at 4 kHz is the audiometric signature of noise-induced hearing loss (the cochlea is most vulnerable to acoustic damage at ~4 kHz, partly an anatomical resonance, partly a consequence of the ear canal’s amplification near 2–3 kHz exposing 4 kHz to peak intensities).
Cookie-bite (or mid-frequency) — thresholds depressed in the mid-frequency range while low and high frequencies are spared. Often a marker of genetic / hereditary etiology.
The configuration is the second axis along which the audiogram is interpreted, after degree. Together they suggest what the underlying disease process probably is, and what tests should follow.
History
⏳The history— Carhart, ASHA, and the standards behind dB HL
Pure-tone audiometry as we now practise it descends from the Western Electric 1A audiometer (1922), the first electronic instrument capable of generating calibrated tones at controlled levels. Through the 1930s, audiometers were calibrated in dB SPL — the physical acoustic pressure — but this made audiograms hard to interpret: a 30 dB SPL threshold at 250 Hz means something completely different from a 30 dB SPL threshold at 4 kHz, because human hearing sensitivity is wildly frequency-dependent.
Raymond Carhart, working at Northwestern after WWII, championed the use of hearing level (HL) — a frequency-dependent reference that subtracts the normal-hearing threshold at each frequency, so that “0 dB HL” means “at threshold for an average normal-hearing young adult at this frequency.” This makes the audiogram horizontal for a normal-hearing patient and renders the shape of any hearing loss directly visible.
The conversion table — the Reference Equivalent Threshold Sound Pressure Level (RETSPL) — was standardised first by ASA (the American Standards Association, 1951) and revised by ASA/ANSI (American National Standards Institute, 1969), then internationally by ISO 389 (1964 onwards). The current US standard is ANSI S3.6-2018, with frequency-specific RETSPL values for several earphone types (TDH-39 / TDH-49 supra-aural, insert phones, circumaural). Different earphones have different RETSPLs; the calibration converts ear-canal pressure to dB HL via the earphone’s known acoustic response.
The 5-dB step size on the audiometer dial is also Carhart’s: he advocated for it as the resolution that balances test-retest reliability (about ±5 dB even for well-trained patients) against test time. Audiometers with 1- or 2-dB steps exist but are mostly used for research and threshold-microstructure studies, not routine clinical work.
What’s next
The next lesson, 2.2 — Decibels: SPL, HL, SL, unpacks the y-axis of the audiogram. dB HL is one of four decibel scales an audiologist encounters; understanding how they relate is essential to reading the audiometer’s readout correctly and to communicating with other professionals who use different scales.