Spectral Approach to the Modeling of the Singing Voice

TitleSpectral Approach to the Modeling of the Singing Voice
Publication TypeConference Paper
Year of Publication2001
Conference Name111th AES Convention
AuthorsBonada, J., Loscos A., Cano P., & Serra X.
Conference Start Date30/11/2001
Conference LocationNew York
AbstractIn this paper we will present an adaptation of the SMS (Spectral Modeling Synthesis) model for the case of the singing voice. SMS is a synthesis by analysis technique based on the decomposition of the sound into sinusoidal and residual components from which high-level spectral features can be extracted. We will detail how the original SMS model has been expanded due to the requirements of an impersonating applications and a voice synthesizer. The impersonating application can be described as a real-time system for morphing two voices in the context of a karaoke application. The singing synthesis application we have developed generates a performance of an artificial singer out of the musical score and the phonetic transcription of a song. These two applications have been implemented as software to run on the PC platform and can be used to illustrate the results of all the modifications done to the initial SMS spectral model for the singing voice case.
Full Documentfiles/publications/aes2001-bonada.pdf
intranet