Contribution
Time-Varying Voice Directivity in Real-Time 6-DoF Room Acoustic Simulation
* Presenting author
Abstract:
Source directivity is an integral part of room acoustic simulation, affecting not only the intensity and frequency balance of the direct sound depending on the source orientation, but also room acoustic parameters such as the direct-to-reverberant energy ratio. While the directivity of a loudspeaker usually does not change over time, the directivity of other sound sources, such as the human voice, does. Whether these time-dependent directivity changes need to be fully considered for a plausible binaural rendering of human speech remains an open research question, and to investigate this, rendering solutions capable of reproducing speech with time-varying (dynamic) directivity are required. This work introduces a new feature for liveRAZR, a real-time six-degrees-of-freedom (6-DoF) room acoustic simulator, that allows the rendering of any sound source with time-varying directivity using audio inputs in Ambisonics format. The study further presents the signal processing to transform speech recordings into Ambisonics signals usable in liveRAZR, including the synthesis of Ambisonics speech signals with time-varying directivity using phoneme-specific directivity data sets in SOFA format and single-channel speech recordings. A final listening experiment compares time-varying and stationary voice directivity.