SteadyState Vowel Duration

Vowel duration is one of the most frequently examined temporal properties in experimental phonetics and psycholinguistics. While all vowels change in length during articulation, researchers often focus on the steadystate portion the segment of the vowel that is produced with relatively stable articulatory configuration and acoustic quality. Understanding steadystate duration helps us answer questions about languagespecific timing patterns, prosodic effects, speech perception, and the interaction of phonetics with phonology.

What Is SteadyState?

In a vowel, three temporal phases can be distinguished:

• Onset (or attack): the transition from the preceding consonant into the vowel.
• Steadystate: the interval when the articulators hold a relatively constant shape, producing a stable formant structure.
• Offset (or release): the transition from the vowel into the following consonant.

The steadystate is typically defined as the part of the vowel that lies between the end of the onset transition and the beginning of the offset transition. In acoustic terms, it is the region where formant frequencies (F1, F2, F3) show minimal slope.

Why Focus on SteadyState Duration?

• Crosslinguistic comparison: Many languages use vowel length contrastively (e.g., Japanese, Finnish). Measuring only the steadystate portion reduces influence from neighboring consonants, yielding cleaner contrasts.
• Prosodic marking: Stress, tone, or boundary tones often lengthen the steadystate part of a vowel while leaving the transitions relatively unchanged.
• Speech perception: Listeners rely heavily on steadystate cues (formant trajectories, spectral shape) to identify vowel quality; duration of this region influences perceived length.
• Articulatory modeling: The steadystate corresponds to a stable vocal tract configuration, making it a natural target for biomechanical simulations.

Measuring SteadyState Duration

There is no single universally accepted algorithm, but the most common approaches share two steps: (1) locate the vowel boundaries and (2) isolate the steadystate interior.

1. Detecting Vowel Boundaries

Boundary detection usually relies on intensity, spectral, or formant criteria. A simple method:

1. Compute a shorttime energy contour.
2. Mark the first point where energy exceeds a threshold (onset).
3. Mark the last point where energy stays above the threshold (offset).

2. Identifying the SteadyState Region

Two widely used techniques are described below.

LinearRegression Windowing

Slide a window (e.g., 20ms) across the vowel and compute the slope of F1 and F2 within each window. The steadystate begins at the first window where both slopes fall below a preset limit (e.g., |slope|<5Hz/ms) and ends at the last such window.

PercentileBased Trimming

Determine the total vowel duration (t_total). Trim a fixed percentage (often 10%15%) from the beginning and end; the remaining middle portion is taken as the steadystate. This method is quick but assumes that the trimmed portions are largely transitional.

Practical Example (Praat script fragment)

Factors Influencing SteadyState Duration

Several linguistic and extralinguistic variables modify how long the steadystate portion lasts.

Phonological Length Contrast

In languages with phonemic vowel length, the steadystate is the principal locus of the contrast. Short vowels may have steadystate durations of ~5080ms, while long vowels often exceed 150ms. The difference is usually proportional across speakers.

Stress and Prominence

Stressed syllables tend to have longer steadystates than unstressed ones, even when vowel quality is identical. The effect can be approximated by a multiplicative factor of 1.21.5.

Speech Rate

When speakers accelerate, both the overall vowel duration and the steadystate shrink, but the proportion of steadystate relative to the whole vowel may stay roughly constant ( 7080%).

Phonetic Context

Adjacent consonants affect transition length. A highfriction consonant such as /s/ can produce a longer onset, leaving a shorter steadystate, while a lowfriction stop like /b/ yields a brief transition and a relatively longer steadystate.

Dialectal and Individual Variation

Even within the same language, dialects differ in how much timing is allocated to the steadystate versus the transitions. Individual speakers also show systematic differences related to vocal tract size and habitual speaking style.

Applications

• Speech technology: Automatic vowellength classification in ASR systems often uses steadystate duration as a primary feature.
• L2 acquisition research: Learners of a language with vowellength contrasts frequently mistime the steadystate, leading to perceptual errors.
• Clinical phonetics: Disorders such as dysarthria may compress the steadystate, providing a diagnostic cue.
• Historical linguistics: Lengthening or shortening of vowel steadystates can trace sound change processes (e.g., compensatory lengthening).

Further Reading

• Crystal, D. (2008). Dictionary of Linguistics and Phonetics. 2nd ed. WileyBlackwell.
• Keating, P. (1989). Prosodic Domains and Lexical Stress. In *The Structure of Phonology* (pp. 299322). Blackwell.
• Ladefoged, P. & Johnson, K. (2014). Acoustic Phonetics. 3rd ed. Wiley.
• Scott, S. (2008). Vowel Duration and Perception. *Journal of Phonetics*, 36(4), 675689.
• Yuan, J., & Liberman, M. (2008). Acoustic and Perceptual Evidence for a Vowel Length Contrast in Mandarin. *Phonetica*, 65(2), 7387.

By isolating and analysing the steadystate portion of a vowel, researchers obtain a clearer picture of temporal patterns that are fundamental to both the phonetic signal and its linguistic interpretation.