Third International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2003)
The time scale modification (TSM) of speech is concerned with the compressing or expanding of audio signals in the time domain without affecting the signals pitch or naturalness. Conversely, the frequency scale modification (FSM) of speech is concerned with altering the pitch and formants of a signal without changing the signal duration. This paper describes a hardware implemented and optimized TSM/FSM system. Biomedical speech related applications for such a system include accelerated aural reading for the blind and improved speech recognition In a voice controlled robotic system for the disabled, the speech can be effectively slowed down to improve the recognition rate. Other applications of the system include speech synthesis, foreign language learning, audio typing, and voice transformation.
Index Terms. TSM, FSM, VLSI
Full Paper (reprinted with permission from Firenze University Press)
Bibliographic reference. Breen, D. / O'Neill, R. / Smith, T. D. / Schwarzbacher, A. Th. (2003): "VLSI implementation of a TSM/FSM algorithm", In MAVEBA-2003, 205-208.