4th International Conference on Spoken Language Processing
Philadelphia, PA, USA
Synthesis systems that convert orthographic text into speech usually make assumptions about the input that are no longer valid when used in combination with a scanner and OCR software.This paper describes our experience of adapting our TTS system for use use in such a reading machine.
As synthesis systems move from the laboratory to applications, some surprises may be in store. This paper is intended to share our experiences with other developers in order to prevent them from repeating our mistakes.
Bibliographic reference. Portele, Thomas / Krämer, Jürgen (1996): "Adapting a TTS system to a reading machine for the blind", In ICSLP-1996, 184-187.