Silas Antonisen
Academic Researcher & PhD Student
Universidad de Granada
About Me
I am a PhD student and academic researcher at the University of Granada. My research focuses on music information retrieval (MIR) and computational linguistics, with a particular emphasis on automatic lyrics transcription, lyrics translation, and singing voice synthesis. My goal is to develop a system enabling singers to perform in languages they do not speak through a process I term singing-voice-to-singing-voice translation (SV2SVT).
Interests
- Music Information Retrieval
- Computational Linguistics
- Singing Voice Synthesis
- Transcription and Translation
- Signal Processing
Education
-
PhD in Telecommunications Engineering
Universidad de Granada (Current)
-
MSc in Vision, Graphics and Interactive Systems
Aalborg University (2021-2024)
-
BSc in Robotics
Aalborg University (2018-2021)
Publications
PolySinger: Singing-Voice to Singing-Voice Translation from English to Japanese
Silas Antonisen and Iván López-Espejo
25th International Society for Music Information Retrieval (ISMIR 2024) Conference
November 10-14, 2024, San Francisco, CA, USA
The speech domain prevails in the spotlight for several natural language processing (NLP) tasks while the singing domain remains less explored. The culmination of NLP is the speech-to-speech translation (S2ST) task, referring to translation and synthesis of human speech. A disparity between S2ST and the possible adaptation to the singing domain, which we describe as singing-voice to singing-voice translation (SV2SVT), is becoming prominent as the former is progressing ever faster, while the latter is at a standstill. Singing-voice synthesis systems are overcoming the barrier of multi-lingual synthesis, despite limited attention has been paid to multi-lingual songwriting and song translation. This paper endeavors to determine what is required for successful SV2SVT and proposes PolySinger (Polyglot Singer): the first system for SV2SVT, performing lyrics translation from English to Japanese. A cascaded approach is proposed to establish a framework with a high degree of control which can potentially diminish the disparity between SV2SVT and S2ST. The performance of PolySinger is evaluated by a mean opinion score test with native Japanese speakers. Results and in-depth discussions with test subjects suggest a solid foundation for SV2SVT, but several shortcomings must be overcome, which are discussed for the future of SV2SVT.