Lawrence J (2023)
Publication Type: Thesis
Publication year: 2023
DOI: 10.13140/RG.2.2.34073.19045/1
Analyzing spatial sound fields typically necessitates a large number of microphones placed at
specific spatial points. The usage of moving microphones provides an alternative method for
analysis, potentially reducing the required number of microphones. To investigate the potential
of a single fast moving microphone we previously developed the rotating equatorial microphone
prototype. The circular motion of this microphone introduces distortions into recorded sound
sources which can be used to determine their direction of arrival (DOA).
In this thesis an algorithm is derived which compensates these distortions for arbitrary sound
sources given their DOA. Additionally, a metric is introduced which quantifies the distortions
present in an audio recording. Subsequently, arbitrary sound sources with an unknown DOA are
localized by applying the algorithm for various DOA guesses and determining the DOA with
minimum distortion using the previously defined metric. With this approach we localized four
simultaneous wideband audio signals in 2D space and two wideband audio signals in 3D space
with ±10◦ accuracy in simulations. We then verified the approach in practice for one and two
simultaneous audio sources. We found that accurate localization was only possible in 2D space
for audio sources mainly consisting of constant tones in the mid frequency range.
APA:
Lawrence, J. (2023). Sound Source Localization with the Rotating Equatorial Microphone (REM) (Master thesis).
MLA:
Lawrence, Jeremy. Sound Source Localization with the Rotating Equatorial Microphone (REM). Master thesis, 2023.
BibTeX: Download