Audio Rendering of Mathematical Formulae Using MathML and AudioMath

Technical, scientific or even simple documents presented online that involve mathematical expressions, issue a big problem with accessibility regarding visually impaired persons. One possible solution is to parse and interpret the mathematical contents and convert them into an audio format. This is a complex and multidisciplinary problem, which AudioMath [1], a work in progress at LPF-ESI [2], studies and aims to contribute to solve. AudioMath provides conversions from mathematical expressions in W3C’s MathML [3][4] format into plain text, something that a capable text-to-speech (TTS) engine would “understand” and read-out. This paper reviews the state of the art in publishing mathematical documents in the Internet, and introduces the AudioMath prototype and its working mechanisms, as an accessibility tool for use together with a TTS engine or to use as a standalone application, with great advantages for visual impaired persons. The main scientific and technical challenges, specifically at the levels of interpretation of the content of the MathML coded expressions and conversion into spoken form, are considered and the current status of the work is described.