The Mathtalk Program

Key Words

Algebra, blind people, synthetic speech, prosody, browsing, glancing, algebra earcons.

Mathtalk and Maths

Mathtalk was developed by Robert Stevens at the University of York. Many of the ideas developed within this project have been taken on and expanded within the Maths project (Mathematical Access for TecHnology and Science for visually disabled people). This page describes Mathtalk, but most of the material is equally relevant to Maths. (The sound recordings in the examples on these pages have been recorded to high and low qualities. The high quality ones sound better - but will take longer to arrive. You can choose which quality you want to hear by selecting either 'High' or 'Low'.)

Introduction

Algebra notation is a vital component of many scientific, technical and mathematical disciplines. The inability to use standard algebra notation efficiently and effectively can be a disadvantage in education and subsequent employment. Algerba notation relies heavily on the use of pencil and paper, often using complex structures that lack redundancy. The loss of any one symbol or grouping boundary within an expression can radically alter its meaning. Not being able to use pencil and paper, or braille can mean mathematics becomes inaccessible to many blind people. The Mathtalk program speaks standard algebra notation using a speech synthesiser. It can be used to give blind people the opportunity to read algebra actively, rather than listen passively. To do this the design of the Mathtalk program concentrates on the issues of external memory and control of information flow. Paper holds information external to the reader's own memory, thus relieving him or her of the burden of holding all that information. The visual system is able to shift attention rapidly and focus upon any part of the structure with accuracy.

Using Prosody to Enhance the Display

The prosodic component of speech has been used to improve the utility of synthetic speech as an external memory. Prosody (the rhythm, pitch, amplitude and tempo of speech) can be used to indicate the structure of an utterance. It also aids retention of an expression's content and reduces the mental workload associated with the listening task. The Mathtalk program adds prosody to spoken algebra to make it easier to listen to.

Controlling the Information Flow

The second element of the Mathtalk program is the control of the information flow. This is achieved using browsing functions and an associated command language. A structure based browsing through a language designed to give the speed and accuracy necessary for active reading.

Glancing at an Algebra Expression

The last component of the Mathtalk program was an audio glance. This could be used to give a high-level view of the structure of an expression, without any of its detail. Such a glance could be used to help a reader plan use of browsing functions and anticipate elements of the expression to come. The audio glance is based on earcons (the audio equivalent of icons) and were called algebra earcons. Each structural component of an expression is associated with a musical timbre and the structure of the expression, together with the prosodic form gave the algebra earcon.

The Maths Project

The Mathtalk program only allows reading. The Maths project aims to produce a mathematics workstation upon which someone can read, write and manipulate mathematics using a multi-media interface containing speech, braille and enhanced visual display.

Maths Project publications.

Return to Robert Stevens' home page.

Valid HTML 4.01!