Development of an algorithm for improving quality and information processing capacity of MathSpeak synthetic speech...

Area of Science:

• Assistive Technology
• Speech Synthesis
• Mathematics Education

Background:

• MathSpeak is a system for rendering mathematical expressions into synthetic speech.
• Current MathSpeak systems lack adequate pausing, leading to unclear utterances.
• This ambiguity hinders information processing for users, particularly those with print disabilities.

Purpose of the Study:

• To develop an algorithm for inserting pauses into MathSpeak synthetic speech.
• To improve the clarity and intelligibility of MathSpeak renderings.
• To enhance the accessibility of mathematical content for individuals with print disabilities.

Main Methods:

• An algorithm was developed to insert pauses based on recordings of math teachers.
• Efficacy testing involved college students without disabilities and students with visual impairments.
• Measured parameters included reception accuracy, memory retention, and speech quality rankings.

Main Results:

• The developed algorithm resulted in statistically significant improvements across all measured parameters.
• Reception accuracy and short-term memory retention were enhanced.
• Perceived quality of synthetic speech renderings showed significant improvement.

Conclusions:

• The pause-insertion algorithm significantly enhances MathSpeak synthetic speech quality and information processing.
• This advancement improves the capacity of individuals with print disabilities in STEM fields.
• The findings support greater inclusion and success in academic and career STEM objectives.