An acoustic-perceptual method for the quantitative evaluation of hypernasality

Area of Science:

• Computer Science
• Linguistics
• Speech Science

Background:

• Assessing speech quality is crucial for diagnosing and treating speech disorders.
• Traditional methods often rely on subjective perceptual evaluations.
• Objective and quantifiable measures of speech production are needed.

Purpose of the Study:

• To propose a computer implementation for objectively assessing speech quality.
• To evaluate speech from perceptual and communicative viewpoints.
• To provide a numerical specification of phonetic segment adequacy.

Main Methods:

• Utilizing a standardized set of speech samples (vowels and consonants) for phonological evaluation.
• Employing a simulated auditory analyzer to generate spectral representations.
• Implementing a psychoacoustic algorithm to compute perceptual distinctiveness between spectra.
• Comparing analyzed speech items against reference samples from normal speakers or pre-intervention recordings.

Main Results:

• The computer program outputs a numerical score indicating the functional adequacy of speech segments.
• The system allows for quantitative analysis of phonological distinctiveness.
• Objective comparison of patient speech to normative data is enabled.

Conclusions:

• The proposed computer implementation offers an objective method for speech quality assessment.
• This tool can aid in the evaluation of phonological function and communicative adequacy.
• The system provides a valuable adjunct to traditional speech analysis techniques.