Speech Audiometry

Speech audiometry measures how well a listener can detect and understand speech rather than pure tones. It adds a functional dimension to the audiogram by establishing the level at which speech becomes audible (the speech reception or recognition threshold) and how accurately words are identified at a comfortable level (word recognition score), including increasingly under more realistic listening conditions such as background noise.

Definition

Speech audiometry is the measurement of a listener's ability to detect, repeat, or recognize speech material at controlled presentation levels, yielding threshold measures (such as the speech recognition threshold) and suprathreshold word recognition scores.

Scope

This entry covers the speech detection and recognition thresholds, suprathreshold word recognition scoring, the materials used (spondees and phonetically balanced word lists), the rationale for cross-checking pure-tone thresholds, and the move toward speech-in-noise measures. It is a reference description of the methods, not clinical interpretation guidance.

Core questions

• At what level does speech become just audible or just intelligible?
• How accurately are words recognized at a comfortable listening level?
• Does the speech recognition threshold agree with the pure-tone average as a cross-check?
• How does adding background noise change recognition relative to quiet?

Key concepts

• Speech detection threshold
• Speech recognition (reception) threshold
• Spondee words
• Phonetically balanced word lists
• Word recognition score
• Performance-intensity function
• Speech-in-noise testing
• Cross-check with pure-tone average

Mechanisms

Threshold measures present speech material, classically two-syllable spondee words with equal stress, at decreasing levels to find the point of audibility or 50 percent recognition; suprathreshold measures present phonetically balanced monosyllabic word lists at a comfortable level to score the percentage correctly identified (Hudgins et al. 1947; Carhart 1951). Plotting recognition against presentation level yields a performance-intensity function whose shape can suggest cochlear versus retrocochlear involvement. The speech recognition threshold is expected to agree closely with the pure-tone average of the threshold at speech frequencies, providing an internal cross-check on the behavioural audiogram (ASHA 1988). Because everyday listening occurs amid competing sound, speech-in-noise paradigms present target speech against a masker to estimate the signal-to-noise ratio needed for a criterion level of understanding.

Clinical relevance

Speech audiometry describes the functional consequence of a hearing loss for understanding language and provides a behavioural cross-check on pure-tone thresholds. It is used to document communication ability and to characterize how recognition changes with level and with noise. This entry explains how the measures are obtained; it is not a basis for individual diagnosis or treatment.

Epidemiology

Difficulty understanding speech, particularly in noise, is one of the most common complaints associated with hearing loss, and speech recognition measures are a standard part of audiological evaluation across the lifespan. Speech-in-noise testing has grown in use because everyday difficulty often exceeds what the pure-tone audiogram alone predicts.

History

Recorded speech tests for clinical use were developed in the 1940s at the Harvard Psycho-Acoustic Laboratory (Hudgins et al. 1947), and Carhart consolidated the basic principles of speech audiometry shortly afterward (Carhart 1951). Standardized spondee and phonetically balanced materials and guidelines for speech threshold measurement followed (ASHA 1988), and attention later shifted toward speech-in-noise measures that better reflect real-world listening.

Key figures

• Raymond Carhart
• Hallowell Davis
• S. S. Stevens

Seminal works

• hudgins-1947
• carhart-1951

Frequently asked questions

What is the difference between a speech detection threshold and a speech recognition threshold?

The detection threshold is the level at which speech is just audible as sound, while the recognition (reception) threshold is the level at which the listener can correctly repeat about half of the speech material.

Why test word recognition in addition to thresholds?

Word recognition scoring at a comfortable level measures how clearly speech is understood, not just whether it is audible, and the pattern across levels can help distinguish cochlear from neural problems.