Department of
Computational Perception
|
|
Department of
Computational Perception
|
|
The Vowel Worm: Real-time Mapping and Visualisation of Sung Vowels in Music
Description
In search of meaningful new ways to extract and visualise features and qualities of the singing voice, we developed the Vowel Worm. The Vowel Worm projects sung and spoken vowels into a continuous articulatory space. The underlying mathematical models are of low computational cost which makes it possible to predict and display vowel quality in real-time.
To demostrate this capability we provide videos, so that the interested reader might get a better unterstanding of the presented concept.
Reference:
H. Frostel, A. Arzt, and G. Widmer. "The Vowel Worm: Real-time Mapping and Visualisation of Sung Vowels in Music", in Proceedings of the 8th Sound and Music Computing Conference (SMC 2011), pp. 214 - 219, Padova, Italy, July 2011.
>> PDF,
BibTeX
Videos
| Popular | Opera |
| Commodores - Easy | Richard Wagner - Excerpt from Rheingold |
|
The Vowel Worm seems to work quite accurate on popular songs like the one presented here.
|
Operatic singing cases the Worm to move outside the boundaries of the IPA Vowel Chart from time to time. Especially high pitch and vibrato lead to an imprecise prediction. This can be due to the speech-only corpus used for the training of the regression models.
|
| Marvin Gaye - Let's Get It On | Richard Wagner - Excerpt from Rheingold |
| Tori Amos - Me And A Gun | |