8th European Conference on Speech Communication and Technology

Geneva, Switzerland
September 1-4, 2003


VISPER II - Enhanced Version of the Educational Software for Speech Processing Courses

Miroslav Holada, Jan Nouza

Technical University of Liberec, Czech Republic

In the paper we describe a new version of the software tool developed for education and experimental works in speech processing domain. Since 1997, when the original VISPER was released, we have added several new modules and options that give a student a deeper look at the basic principles, methods and algorithms used namely in speech recognition. Newly included modules allow for visualization of the Viterbi search algorithm implemented either in sequential or parallel way, they introduce the idea of the beam search with pruning and guide a student towards understanding the principle of word string recognition. The VISPER concept of a single graphic environment with mutually linked modules remains untouched. The VISPER II is compatible with all recent versions of the MS Windows OS and it is freely available.

Full Paper

Bibliographic reference.  Holada, Miroslav / Nouza, Jan (2003): "VISPER II - enhanced version of the educational software for speech processing courses", In EUROSPEECH-2003, 3169-3172.