Vocal training with a semi-occluded vocal tract has been used for a long period of time. One way to obtain a semi-occlusion is by phonating through a straw or tube while keeping the free end of the tube in air or submerged into water.

This thesis aims at investigating pressure and flow characteristics of three types of semi-occluded vocal tract exercises: (i) straw/tube phonation with the free end of the tube in air, (ii) tube phonation with the free end in water and (iii) flow ball devices, which are narrow tubes combined with a basket containing a styrofoam ball that lifts off when air is blown into the tube.

Studies I, III and IV investigated pressure characteristics of these devices as functions of flow. Data were collected with a flow driven vocal tract simulator with an outlet for straw/tube/flow ball connection. Studies II and V investigated changes in oral pressure, vertical laryngeal position and fundamental frequency for vocally healthy volunteers during resonance tube phonation in water at two submersion depths.

The results of study I showed that a change in tube diameter affects the pressureflow relationship more than a relative change in tube length for tube dimensions commonly used in voice exercises. When the tubes were submerged into water, the flow could not start until the pressure provided by the water depth was overcome, but as flow increased the pressure-flow relationship was similar to that of the tube in air, but with an upward shift in back pressure related to the pressure provided by the water depth. This was also confirmed in study IV. The oscillating part of the back pressure was analysed in study IV, showing that the amplitude of the oscillations increased with increasing water depth up to 3 cm depth. The amplitude of the pressure oscillations were similar at 3-7 cm water depths.

Results from study IV showed that increasing the flow through a tube submerged in water affected the bubble characteristics. At low flows, the bubbles were emitted one-by-one in a periodic manner, at medium flows the bubbles were emitted in pairs of two and at high flows the bubble formation were chaotic and no clear pattern in the bubble characteristics could be identified. Bubble characteristics differed slightly between tubes with 8 and 9 mm diameter.

The results of study III showed that the pressure-flow relationship for two flow ball devices were similar to those of straws. A flow of 0.2 L/s was required for ball lift off for both devices. Back pressure thresholds for ball lift off were 5 cmH2O for one of the devices, and 20 cmH2O for the other one.

The results of study II and V showed that resonance tube phonation in water provides an average oral pressure slightly above the equivalent water pressure. The vertical laryngeal position, as measured by dual channel electroglottography, dropped for most participants during tube phonation in water, and a fundamental frequency modulation appeared during bubbling for the participants in study V.

The results of this thesis presents new information about the physical properties involved in straw phonation with the free end in air and tube phonation with the free end in air and submerged in water. They further suggest that vocal training with phonation through tubes submerged in water can affect the phonatory system in different ways.