A Method for Transcribing the Manual Components of Cued Speech

Area of Science:

• Linguistics
• Speech-Language Pathology
• Human-Computer Interaction

Background:

• Cued Speech facilitates visual communication of spoken language.
• Accurate transcription of Cued Speech hand signals is essential for research and application.
• Existing methods may lack systematic approaches for analyzing hand signal components.

Purpose of the Study:

• To describe a novel method for transcribing Cued Speech hand signals.
• To establish procedures for analyzing cue components: steady-state portion, handshape, and placement.
• To evaluate the reliability of the proposed transcription method.

Main Methods:

• Development of specific procedures for identifying key features of Cued Speech hand signals.
• Systematic transcription of Cued Speech materials from multiple individuals.
• Reliability assessment through intra-rater and inter-rater analyses on repeated transcriptions.

Main Results:

• The transcription method demonstrated very good average intra-rater and inter-rater reliability.
• Reliability remained high across different cuers and even with coarticulated gestures.
• The method proved effective for analyzing Cued Speech production systematically.

Conclusions:

• The developed transcription method offers high reliability for analyzing Cued Speech.
• This method can benefit applications requiring quantitative analysis of Cued Speech production.
• Transcription principles may enhance automatic Cued Speech recognition systems.