Pure Tone Average Calculator
Pure Tone Average Calculator
Five published pure tone averages from one audiogram — the classic three-frequency average, WHO’s four-frequency average, the AAO-HNS set, the Fletcher index and the occupational 2–4 kHz average — with the spread between them, because they are not interchangeable.
Pure tone average
Five published definitions500 Hz 15, 1000 Hz 15, 2000 Hz 25, 3000 Hz 45, 4000 Hz 65 dB HL — a sloping high-frequency loss — with the WHO four-frequency average selected
The five definitions
- three-frequency average
- the mean of 500, 1000 and 2000 Hz. The oldest definition and still the one most often meant by a bare “PTA”. Fletcher proposed averaging 500 to 2000 Hz in 1929, and the 1959 AAOO hearing impairment calculation was built on it
- WHO four-frequency average
- the mean of 500, 1000, 2000 and 4000 Hz, taken in the better hearing ear. This is the average WHO’s grades of hearing loss are applied to, and the one the Global Burden of Disease estimates use
- AAO-HNS four-frequency average
- the mean of 500, 1000, 2000 and 3000 Hz. In 1979 the impairment calculation was changed to add 3000 Hz rather than 4000 Hz, to estimate speech understanding in noise as well as in quiet. It is the set written into US state compensation rules — Minnesota Rules 5223.0340 subpart 4 names “the four frequencies of 500, 1,000, 2,000 and 3,000 Hertz”
- Fletcher index
- the mean of the BETTER TWO of 500, 1000 and 2000 Hz. Fletcher’s 1950 simplified rule is to “take the average of the two smallest values of loss”, which he reported as almost as reliable as his full formula for predicting hearing loss for speech. It is the only definition here that discards a measured threshold, and on a sloping loss it returns the lowest number of the five
- occupational average
- the mean of 2000, 3000 and 4000 Hz. 29 CFR 1910.95(g)(10)(i) defines a standard threshold shift as “an average of 10 dB or more at 2000, 3000, and 4000 Hz”, and 29 CFR 1904.10 uses the same average for its 25 dB element. It carries no 500 or 1000 Hz at all
- why they diverge
- a definition’s answer depends entirely on whether it admits a frequency the loss has reached. A noise or ototoxic notch at 4000 Hz is invisible to the three-frequency average and to the Fletcher index, visible at a quarter weight to WHO’s average, and carries a third of the occupational average on its own
- units
- dB HL throughout, and dB HL is a scale defined against a reference equivalent threshold level for a named transducer, not an absolute sound pressure. There is no conversion to dB SPL that does not need the ISO 389 table for your own earphone
Worked example
500 Hz 15, 1000 Hz 15, 2000 Hz 25, 3000 Hz 45, 4000 Hz 65 dB HL — a sloping high-frequency loss — with the WHO four-frequency average selected
WHO four-frequency: (15 + 15 + 25 + 65) ÷ 4 = 120 ÷ 4 = 30.0 dB HL
AAO-HNS four-frequency: (15 + 15 + 25 + 45) ÷ 4 = 100 ÷ 4 = 25.0 dB HL — 5 dB lower, and the difference is exactly a quarter of the 20 dB step between 3000 and 4000 Hz
Three-frequency: (15 + 15 + 25) ÷ 3 = 18.3 dB HL
Fletcher index: the better two of 15, 15 and 25 are 15 and 15, so (15 + 15) ÷ 2 = 15.0 dB HL
Occupational: (25 + 45 + 65) ÷ 3 = 45.0 dB HL
Spread 30.0 dB. The same ear is 15.0 dB on one published definition and 45.0 dB on another — a threefold difference on one audiogram, with no measurement error anywhere in it
The consequence is not academic. Under WHO's 2021 grades 30.0 dB is a mild hearing loss; 15.0 dB is normal hearing; 45.0 dB is a moderate hearing loss. Three grades, one ear
Now flatten the audiogram. Set all five thresholds to 40 dB HL and every definition returns 40.0 with a spread of 0.0. That is why a calculator checked only on a flat audiogram cannot tell a correct frequency set from a wrong one
Steepen it instead: 10, 10, 30, 60, 85 gives 16.7, 33.8, 27.5, 10.0 and 58.3, a spread of 48.3 dB. The divergence tracks the slope
Who specifies which average
| Definition | Frequencies averaged | Specified by | Applied to |
|---|---|---|---|
| Three-frequency average | 500, 1000, 2000 Hz | Fletcher (1929); the 1959 AAOO impairment calculation | The ear being described, and still the default meaning of a bare PTA |
| Fletcher index | better two of 500, 1000, 2000 Hz | Fletcher, J Acoust Soc Am 1950 | Predicting hearing loss for speech from an audiogram |
| WHO four-frequency average | 500, 1000, 2000, 4000 Hz | WHO, World report on hearing (2021), and the Global Burden of Disease estimates | The better hearing ear, for the WHO grades and for disabling hearing loss |
| AAO-HNS four-frequency average | 500, 1000, 2000, 3000 Hz | The 1979 revision of the AMA/AAO impairment method; Minnesota Rules 5223.0340 subp. 4; Washington L&I form F252-007-000 | Each ear separately, as the input to the monaural impairment percentage |
| Occupational average | 2000, 3000, 4000 Hz | 29 CFR 1910.95(g)(10)(i) and 29 CFR 1904.10 | Either ear, as the standard threshold shift and the 25 dB element |
The same five definitions on three audiograms
| Audiogram, 500 / 1k / 2k / 3k / 4k | 3-freq | WHO 4-freq | AAO-HNS 4-freq | Fletcher | 2–4 kHz | Spread |
|---|---|---|---|---|---|---|
| 40 / 40 / 40 / 40 / 40 (flat) | 40.0 | 40.0 | 40.0 | 40.0 | 40.0 | 0.0 |
| 15 / 15 / 25 / 45 / 65 (sloping) | 18.3 | 30.0 | 25.0 | 15.0 | 45.0 | 30.0 |
| 10 / 10 / 30 / 60 / 85 (steep) | 16.7 | 33.8 | 27.5 | 10.0 | 58.3 | 48.3 |
| 60 / 40 / 20 / 15 / 10 (rising) | 40.0 | 32.5 | 33.8 | 30.0 | 15.0 | 25.0 |
There is no single pure tone average, and that is the problem
A pure tone average is not one quantity. It is a family of quantities that share a name, differ in which frequencies they admit, and are specified separately by different bodies for different purposes. The classic average takes 500, 1000 and 2000 Hz, the frequencies that carry most of the information in quiet speech. WHO’s grades of hearing loss are applied to the average of 500, 1000, 2000 and 4000 Hz in the better ear. The impairment percentage used in US compensation takes 500, 1000, 2000 and 3000 Hz in each ear separately — the 1979 revision added 3000 Hz, not 4000 Hz. Occupational threshold shift under 29 CFR 1910.95 is defined on the average of 2000, 3000 and 4000 Hz, with no low frequency in it at all. And Fletcher’s own 1950 simplification keeps only the better two of three.
On a flat audiogram this does not matter: every definition returns the same number. On a high-frequency loss it matters a great deal, and high-frequency loss is the commonest shape there is — presbyacusis, noise damage and aminoglycoside or platinum ototoxicity all begin at the top of the audiogram. The worked example above is an ordinary sloping loss, and the five definitions return 15.0, 18.3, 25.0, 30.0 and 45.0 dB HL on it. The lowest and the highest differ by a factor of three.
Two practical consequences follow. First, a number quoted without its definition cannot be checked and should not be acted on; “PTA 30 dB” is not a result until the frequency set is named. Second, a definition is chosen by the scheme receiving the number, not by preference. If the question is which WHO grade applies, the WHO four-frequency average in the better ear is the only correct input, and substituting the three-frequency average will understate a high-frequency loss by a grade or more. If the question is a compensation percentage, 3000 Hz belongs in and 4000 Hz does not, however much the 4000 Hz notch looks like the real injury.
An average also throws away the shape, which is often the diagnostic information. Two ears averaging 40 dB — one flat, one normal to 2000 Hz and 80 dB at 4000 Hz — are different ears with different causes and different consequences for speech in noise, and the average cannot tell them apart. Read the audiogram, then compute the average the receiving scheme asks for. Audiometric results are equipment- and calibration-dependent: the dB HL scale is defined against the reference equivalent threshold levels of the ISO 389 series for a particular transducer and coupler, so the same ear can read differently on two audiometers, and the BSA recommended procedure puts the uncertainty of any threshold measurement at no better than plus or minus 5 dB. Your own audiogram and the calibration standard it was recorded against govern, not this arithmetic.
Frequently asked questions
What frequencies are in the pure tone average?
It depends which average. The classic three-frequency average uses 500, 1000 and 2000 Hz. WHO’s grades use 500, 1000, 2000 and 4000 Hz in the better ear. The AAO-HNS and AMA impairment percentage uses 500, 1000, 2000 and 3000 Hz. OSHA’s standard threshold shift uses 2000, 3000 and 4000 Hz. The Fletcher index uses the better two of 500, 1000 and 2000 Hz.
Which pure tone average should I use?
The one the scheme receiving the number specifies, not the one you prefer. A WHO grade is defined on the 500/1000/2000/4000 Hz average in the better ear; a US compensation percentage is defined on the 500/1000/2000/3000 Hz average in each ear. Using the wrong set does not give a slightly different answer, it gives an answer to a different question.
How much do the definitions actually differ?
Not at all on a flat audiogram, and by tens of decibels on a sloping one. On the worked example here — 15, 15, 25, 45, 65 dB HL — the five definitions give 15.0, 18.3, 25.0, 30.0 and 45.0 dB HL, a spread of 30 dB. Steepen the slope to 10, 10, 30, 60, 85 and the spread grows to 48.3 dB.
What is the Fletcher index?
Fletcher’s 1950 simplified rule: average the two smallest losses among 500, 1000 and 2000 Hz and discard the third. He reported it as almost as reliable as his full formula for estimating hearing loss for speech. It is the only definition here that drops a measured threshold, and on any sloping loss it returns the most optimistic of the five numbers.
Why does the audiogram have 3000 Hz on some reports and not others?
3000 Hz and 6000 Hz are intermediate frequencies. The BSA recommended procedure tests them where there is a difference of 20 dB or more between adjacent octave frequencies, and occupational and ototoxicity protocols test them routinely, so a diagnostic audiogram may not carry 3000 Hz while a surveillance one does. Without it the AAO-HNS average cannot be computed at all.
Related calculators
References
- Fletcher H. A method of calculating hearing loss for speech from an audiogram. J Acoust Soc Am. 1950;22(1):1–5.
- American Academy of Audiology. Pure tone average and speech in noise. Reston, VA: AAA.
- Minnesota Rules 5223.0340, Hearing loss, subpart 4. Minnesota Revisor of Statutes.
- Washington State Department of Labor and Industries. Hearing loss worksheet, form F252-007-000. Olympia, WA: L&I.
- Occupational Safety and Health Administration. 29 CFR 1910.95 — Occupational noise exposure. Washington, DC: US Department of Labor.
Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/