Phase 2 - optical readout of novel directional MEMS microphone technology

新兴的语音应用，如自动语音识别，正在推动对定向麦克风的需求，这种麦克风可以从特定方向拾取声音，同时不受其他来源的环境噪声的影响。当前的定向声音捕获解决方案基本上依赖于三角测量：** 全向麦克风的空间阵列通过测量多个阵列 ** 元件之间的时间延迟来实现定向灵敏度。这些阵列很大，而且只在声波 ** 波长（低频>10米）小于阵列尺寸 ** 的频率下定向。Soundskrit正在开发一种新型的微型定向片上麦克风的基础上，从基础声学研究出现的最新进展。他们使用非常柔软的纳米纤维，这样它们就可以随着声压波的强度和最重要的方向移动。然而，** 这些纳米纤维的极端柔软性使得测量它们的位移（即，录音）极具挑战性。极软的纤维使其与传统的电位移 ** 测量传感器不兼容，如电应变仪（这会给纤维增加不可接受的重量）或 ** 电容读数器（这会产生比 ** 声波施加的力大得多的寄生静电力）。** Soundskrit正在研究两种不同的方法来解决这一限制：磁性（由 **Soundskrit领导）和光学读数（由uOttawa的St-Gelais实验室领导）。我们预计光学方法将产生 ** 最高的性能（对于大多数微系统表征技术来说），但也是最难 ** 小型化的。我们合作的目标是同时开发这两种方法，并将它们相互比较，以找到最有前途的一种。

Emerging voice applications, such as automatic speech recognition, are driving the need for directional**microphones that can pick up sound from a specific direction while being immune to ambient noise from other**sources. Current solutions for directional sound capture essentially rely on triangulation: a spatial array of**omnidirectional microphones achieves directional sensitivity by measuring time delay between multiple array**elements. These arrays are inconveniently large and are only directional at frequencies for which the acoustic**wavelength (>10 meters at low frequencies) is smaller than the array dimension.**Soundskrit is developing a new type of miniaturized directional on-chip microphone based on recent**progress emerging from fundamental acoustic research. They use nanofibers that are extremely soft, such that**they can move following the intensity and-most importantly-the direction of acoustic pressure waves. However,**the extreme softness of these nanofibers makes measuring their displacement (i.e., recording sound) extremely**challenging. Extreme softness makes the fiber incompatible with conventional electrical displacement**measurement transducers such as electrical strain gauges (which would add unacceptable weight to the fiber) or**capacitive readouts (which would create a parasitic electrostatic force much greater than the force exerted by**acoustic waves).**Soundskrit is investigating two different approaches for solving this limitation: magnetic (lead by**Soundskrit) and optical readouts (lead by St-Gelais lab at uOttawa). We expect the optical approach to yield the**highest performance (as for most microsystems characterization techniques), but to be the hardest to**miniaturize. The goal or our collaboration is to develop both approaches in parallel and to benchmark them**against one another to find the most promising one.