Unidirectional Sound Signage for Speech Frequency Range Using Multiple-Loudspeaker Reproduction System

Human safety is the most important issue in disaster management. Speech is a sound signal containing information that is easily and quickly understood by humans. Using speech as sound signage in emergency systems can effectively increase human safety in low or poor visibility conditions such as in smoke-filled situations. However, reflections of sound through walls, floor surfaces, and ceilings will affect clarity of speech. Unfortunately, because of the characteristics of sound reproduction systems, a single loudspeaker propagates sound waves omni-directionally at low frequencies. This paper proposes a simple multiple-loudspeaker system for reproducing sound with uni-directional characteristics. The proposed system consists mainly of a primary loudspeaker for introducing sound in the desired beam, a secondary loudspeaker for reducing gain in the undesired direction, and digital filters. An adaptive finite-impulse-response (FIR) filter is used to produce the controlling sound by implementing a filtered-x least-mean-square algorithm, and a delay filter for adjusting the time alignment of sound propagation between primary and secondary sources at the control point. Several operational conditions for illustrating real situations and reflections were considered in an anechoic chamber. Experimental results show the directivity patterns of the proposed multiple-loudspeaker system for the required conditions. In a low frequency range, the system is able to control unidirectional propagation; there is a sound beam in the desired direction and, conversely, reduction of gain in the undesired direction around the control point.