Home › Speech
Acoustic measures and norms
Acoustic analysis turns a recording of speech or voice into numbers. Those numbers are only as useful as the recording, the software and the reference data behind them. This guide lists the commonly used measures with approximate values.
Read this first. Norms vary with the software, microphone, task (sustained vowel or connected speech), age, sex and language or dialect. The values below are rounded, commonly cited figures for orientation only. For clinical decisions use the reference data supplied with your own system and service.
Summary table
| Measure | What it describes | Approximate typical value |
|---|---|---|
| Fundamental frequency (F0), speaking | Average voice pitch | Adult males about 100–130 Hz; adult females about 190–220 Hz; young children about 250–300 Hz |
| Maximum phonation time (MPT) | How long /a/ can be sustained on one breath | Healthy adults often 15–25 s or more; children shorter |
| S/Z ratio | Compares how long /s/ and /z/ can be sustained | About 1.0; 1.4 or more may suggest laryngeal pathology (Eckel & Boone, 1981) |
| Jitter (local) | Cycle-to-cycle variation in pitch | Below about 1% in Praat-style analysis (software dependent) |
| Shimmer (local) | Cycle-to-cycle variation in loudness | Below about 3.8% in Praat-style analysis (software dependent) |
| Harmonics-to-noise ratio (HNR) | Balance of periodic sound and noise | Healthy adult sustained vowels commonly around 20 dB or higher |
Fundamental frequency (F0)
F0 is the rate of vocal fold vibration and is heard as pitch. It differs between people and changes with age, hormones, emotion and speaking style. Report both the mean and the range (for example in semitones), because a person can have a typical average with an unusually narrow range.
Maximum phonation time and S/Z ratio
MPT is measured by asking the person to take a deep breath and sustain /a/ as long as possible at a comfortable pitch and loudness. It is usually the best of several trials. A short time can reflect reduced breath support, poor vocal fold closure, or simply low effort, so it is a screening indicator and not a diagnosis.
The S/Z ratio divides the longest /s/ by the longest /z/. In people with healthy larynges the two are similar, giving a ratio near 1.0. Because /z/ needs the vocal folds to vibrate, a clearly shorter /z/ can point to a problem at the larynx (Eckel & Boone, 1981). A ratio below 1.0 is common in children and is not usually a concern.
Jitter, shimmer and HNR
These “perturbation” measures describe tiny irregularities in a sustained vowel. They are sensitive to recording quality, background noise and the algorithm used, so values from different programs are not interchangeable. Many clinicians treat them as supporting information and rely more on auditory-perceptual assessment, patient report and examination of the larynx.
Formants
Formants are resonances of the vocal tract. The first two (F1 and F2) mainly distinguish vowels: F1 relates to how open the jaw is, and F2 to how far forward the tongue is. The table gives rounded averages for three American English vowels from Peterson and Barney (1952).
| Vowel | Men F1 | Men F2 | Women F1 | Women F2 |
|---|---|---|---|---|
| /i/ as in “heed” | 270 Hz | 2290 Hz | 310 Hz | 2790 Hz |
| /ɑ/ as in “hod” | 730 Hz | 1090 Hz | 850 Hz | 1220 Hz |
| /u/ as in “who’d” | 300 Hz | 870 Hz | 370 Hz | 950 Hz |
Voice onset time (VOT)
VOT is the time between the release of a stop consonant and the start of voicing. In English word-initial stops, voiced consonants have short VOTs and voiceless consonants have long ones, roughly: /b/ 1 ms, /d/ 5 ms, /g/ 21 ms versus /p/ 58 ms, /t/ 70 ms, /k/ 80 ms (Lisker & Abramson, 1964). VOT is useful in studying speech sound development, accent and some speech disorders.
Recording tips
- Use a quiet room and a decent microphone about 15–20 cm from the mouth, slightly off-axis to reduce breath noise.
- Avoid clipping: peaks should stay below the maximum level.
- Keep the same set-up, task and instructions each time so results can be compared.
- For perturbation measures, analyse a steady middle section of a sustained vowel, not the start or end.
- Free software such as Praat can measure all the values above.
Sources and further reading
- Peterson, G. E., & Barney, H. L. (1952). Control methods used in a study of the vowels. Journal of the Acoustical Society of America, 24(2), 175–184.
- Eckel, F. C., & Boone, D. R. (1981). The S/Z ratio as an indicator of laryngeal pathology. Journal of Speech and Hearing Disorders, 46(2), 147–149.
- Lisker, L., & Abramson, A. S. (1964). A cross-language study of voicing in initial stops: Acoustical measurements. Word, 20(3), 384–422.
- Boersma, P., & Weenink, D. Praat: doing phonetics by computer [software]. www.praat.org