Improved voice acquisition using acoustic metamaterials

Voice acquisition and recognition are important, because sound is the most convenient and efficient media to realize seamless and hands-free human-machine interaction. Many applications, such as smart phones, humanoid robots, smart home, autonomous vehicles, and Internet of Things, demand for high-quality voice acquisition, especially in a noisy environment. In this report, we show that the evanescent portion of human voice can be amplified by setting an acoustic metamaterial in front of a microphone; thus, the signal to noise ratio can be enhanced remarkably. Such an effect is demonstrated through a double negative acoustic metamaterial slab with simultaneous negative density and bulk modulus induced by coexisting and coupled Helmholtz and Fabry-Perot resonances. This work provides a passive and physical means, rather than the traditional active and algorithmic method, to improve the quality of voice sound acquisition.