Generating virtual microphone signals in noisy environments

Abstract

Spatial sound acquisition methods typically capture the sound scene with reference to the position of the recording device. Using a recently proposed virtual microphone (VM) technique, the position and characteristics of the recording device (such as the directivity response and orientation) can be modified. This technique relies on synthesizing a VM signal at an arbitrary position, which sounds perceptually similar to the signal that would be recorded with a physical microphone placed at the same location. In this paper, we present a method to generate a VM signal in the presence of noise. Noise reduction is accomplished using a parametric multichannel Wiener filter, where a trade-off parameter is applied in order to achieve a constant residual noise level in the generated VM signal, irrespective of the VM position. The simulated experiments show the applicability of the method for signal extraction in the presence of additive noise.