EXPERIMENTAL INVESTIGATION OF ACOUSTIC CHARACTERISTICS OF 3D HUMAN VOCAL TRACT MODEL WITH NASAL CAVITIES V. Radolf * , J. Horáček * , J. Košina * , T. Vampola ** Abstract: The following experiments were carried out to be later used in the verification of a complex mathematical model of human voice production. Acoustic resonance characteristics of a 3D human vocal tract model without and with nasal and paranasal cavities were measured in two different ways: The excitation was realized by (1) self-oscillating vocal folds replica and (2) by sine-tone sweeps from an earphone placed instead of the vocal folds. The resulting resonance and antiresonance frequencies were found to be comparable for both excitation signals. Keywords: formant frequencies, modelling of phonation, biomechanics of voice 1. Introduction Human voice is generated by self-oscillating vocal folds excited by air flowing from the lungs. The vocal folds vibration modulates the stream of air producing the primary sound of voice. This sound signal propagates inside the supraglottal cavities (i.e., in the vocal tract) from the vocal folds to the lips and the nostrils which modify its quality. The acoustic resonances of the vocal tract create so-called formants, which occur as peaks in the envelope of the voice spectrum. The formants define vowels and the voice timbre. The final sound quality of human voice is thus given both by characteristics of the vocal fold vibration and by vocal tract properties (Sundberg, 1987). In the present paper effects of nasal cavities together with all paranasal cavities of the vocal tract are studied. The nasality or so-called velopharyngeal insufficiency is modeled by interconnecting acoustic cavities of the nasal tract with the vocal tract model at the velum (soft palate). The objective of the experimental modelling is to offer data that can be used for verification of a complex mathematical model of phonation. 2. Methods A three-dimensional (3D) model of the vocal tract (VT) for the vowel [a:] was created from the Computer Tomography (CT) measurement of a female subject during phonation, see Vampola et al. (2015). The complex 3D volume model of acoustic nasal cavities was developed from a detailed CT investigation of the head of another subject of the same gender, similar age and size. The physical model was made from the volume model by 3D printing, see Fig. 1. Two kinds of excitation of acoustic cavities were realized. First by self-oscillating vocal folds replica and second by an earphone placed instead of the vocal folds. The acoustic pressure inside the mouth was measured with a B&K 4138 miniature microphone (range 6.5 Hz - 140 kHz) 3 mm from the lips and a special B&K 4182 microphone probe (range 1 Hz - 20 kHz) measured the sound 3 mm from the nostril * Ing. Vojtěch Radolf, PhD., Ing. Jaromír Horáček, DrSc., Ing. Jan Košina: Institute of Thermomechanics, The Czech Academy of Sciences, Dolejškova 1402/5; 182 00, Prague; CZ, radolf@it.cas.cz ** prof. Dr. Ing. Tomáš Vampola: Faculty of Mechanical Engineering, Czech Technical University in Prague; Technická 4; 166 07, Prague; CZ 705 24 th International Conference ENGINEERING MECHANICS 2018 Svratka, Czech Republic, May 14 –17, 2018 Paper #63, pp. 705–708, doi: 10.21495/91-8-705