GENIOVOX Project: Computational generation of expressive voice

Oriol Guasch, Francesc Alías, Marc Arnela, Joan Claudi Socoró, Marc Freixes, Arnau Pont

The GENIOVOX project: “Computational synthesis of expressive voice”, with ref. TEC2016-81107-P and funded by the Ministerio de Economía, Industria y Competitividad (Plan Nacional de I+D Excelencia) was carried out in the period 2016-2019. Its two main objectives were the following ones. On the one hand, diphthongs and hiatuses were simulated in three-dimensional (3D) geometries using the finite element method (FEM), based on the resolution of the underlying wave equations. Likewise, techniques were developed to simulate syllables with fricative consonants that did not require the use of high-performance computing. The trick was to approximate the interdental flow acoustic source terms using quadrupole, dipole and monopole distributions instead of getting them from a computational fluid dynamics simulation. In addition to generating diphthongs and syllables with fricatives, the project proposed a first attempt to incorporate some expressive effects through modifications of the vocal tract geometry and the glottal source model. Vowel sounds were computationally generated by convoluting the impulse response of 3D FEM vocal tracts with glottal pulses that incorporated tense, neutral and lax phonations from expressive speech corpora.

doi: 10.21437/IberSPEECH.2021-33

Guasch, O, Alías, F, Arnela, M, Socoró, J.C, Freixes, M, Pont, A (2021) GENIOVOX Project: Computational generation of expressive voice. Proc. IberSPEECH 2021, 151-154, doi: 10.21437/IberSPEECH.2021-33.