Development of the Cantonese speech intelligibility index

Area of Science:

• Audiology
• Speech Science
• Linguistics

Background:

• Developing a Speech Intelligibility Index (SII) is essential for accurate speech assessment in various languages.
• Existing speech intelligibility tests, like the English Hearing In Noise Test (HINT), may not fully capture the nuances of tonal languages.
• The Cantonese version of the Hearing In Noise Test (CHINT) requires a language-specific SII to reflect its unique acoustic properties.

Purpose of the Study:

• To derive a Speech Intelligibility Index (SII) for the Cantonese Hearing In Noise Test (CHINT).
• To determine optimal signal-to-noise ratios and filtering conditions for CHINT.
• To compare the frequency importance function of CHINT with English speech materials.

Main Methods:

• Conventional procedures were used to derive the SII for CHINT sentences.
• Two studies were conducted to establish optimal signal-to-noise ratios and filtering conditions.
• Normal hearing subjects listened to CHINT sentences in speech-spectrum shaped noise.

Main Results:

• The frequency importance function for CHINT indicates a greater reliance on low-frequency information for speech understanding compared to English HINT.
• The derived SII provides a measure of speech intelligibility specific to the Cantonese language.
• Optimal signal-to-noise ratios and filtering conditions for CHINT were identified.

Conclusions:

• Low-frequency audibility is more critical for Cantonese speech intelligibility than for English.
• The differences in frequency weighting may be due to the redundancy of the Cantonese test material, the tonal nature of the language, or both.
• The developed SII enhances the accuracy of speech intelligibility assessment for Cantonese speakers.