In our daily life, sound is all around us, moving freely in any direction. For example, when listening to a concert in a concert hall, you can hear the sound of instruments from different directions and feel the voice change as the vocalist moves on the stage.
This presentation from School of Engineering PhD student Huanyu Zuo will look at the recreation of such a lifelike audio experiences. This is a challenge for a spatial sound field reproduction system, one that is vital to many commercial applications, such as home entertainment systems (e.g., 5.1/7.1 systems), modern cinemas (e.g., Dolby Atmos and DTS:X), and VR/AR.
This hybrid event will be held in-person on campus and remotely via Zoom.
In this thesis, we focus on this challenge and find ways to create an immersive sound over a spatial region so that the listener inside the region can experience a realistic but virtual replication of the original sound.
A theory of spatial acoustic vectors is first developed, where the spatial distributions of particle velocity and sound intensity are derived from sound pressure. To extract the desired sound pressure from a mixed sound field environment, a 3D sound field separation technique is also formulated. Based on this theory, a series of reproduction techniques are proposed to improve the perceptual performance.
The outcomes resulting from this theory are:
- derivation of a particle velocity assisted 3D sound field reproduction technique which allows for non-uniform loudspeaker geometry with a limited number of loudspeakers
- design of particle velocity based mixed-source sound field translation technique for binaural reproduction that can provide sound field translation with good perceptual immersion over a large space,
- derivation of a intensity matching technique that can reproduce the desired sound field in a spherical region by controlling the sound intensity on the surface of the region, and
- two intensity based multizone sound field reproduction algorithms that can reproduce the desired sound field over multiple spatial zones.
Finally, the above techniques are evaluated by comparing to the conventional approaches through numerical simulations and real-world experiments.
Huanyu Zuo is a PhD student with ‘Audio & Acoustics Signal Processing Group’, School of Engineering, at The Australian National University. He received a Bachelor and Masters in electrical engineering from the Beijing Institute of Technology, in 2014 and 2017 respectively. He had an audio research internship at HUAWEI in 2020. His research interests include spatial sound recording and reproduction, 3D sound field separation, and psychoacoustics.