Expressive Performance Model for a Singing Voice Synthesizer

TitleExpressive Performance Model for a Singing Voice Synthesizer
Publication TypeMaster Thesis
Year of Publication2005
AuthorsAlonso, M.
preprint/postprint documentfiles/publications/077337-PFC2005-Malonso.pdf
AbstractThis thesis describes the design and implementation of an expressive performance model focused on emotions for a singing voice synthesizer. The resulting model can modify the performance features of a musical fragment in order to simulate different moods or emotions in the performer such as happiness, fear, anger, sadness, etc. The model is based on the rule system developed by Johan Sundberg and Anders Friberg at the KTH university in Stockholm and the results of the research on musical expression and singing voice by Patrik N. Juslin and Petri Laukka from the university of Uppsala, Sweden. The singing voice synthesizer used is Daisy, developed at the Music Technology Group of the University Pompeu Fabra in Barcelona.