Please use this identifier to cite or link to this item:
|Title:||Multipurpose microphone array using Raspberry Pi and MATRIX Creator||Authors:||Chen, Zhe||Keywords:||DRNTU::Engineering::Computer science and engineering||Issue Date:||2017||Abstract:||A microphone array is a set of microphones operating in tandem. It is essential in the field of speech enhancement and can improve the overall robustness of the speech recognition system. With the popularity of Smart Home Assistant devices with a microphone array, such as Amazon Echo, microphone array has proved its usefulness in the field of speech enhancement. Although microphone array is an area of interest in research, the existing products with microphone array are either not easily customizable or lack flexibility in deployment. These obstacles hinder the research into the microphone array. Therefore, this project proposes a pioneering solution of a multipurpose microphone array using MATRIX Creator and Raspberry Pi, to aid the research of microphone array by being easy to customize and flexible to deploy, and to demonstrate its potential applications in speech recognition. For maximum flexibility, this project embraces a fully wireless architecture and allows the microphone array to stream multichannel audio to a remote PC besides saving internally. This project shall also develop real-time Automatic Speech Recognition capability using Google Cloud Speech as its potential application. Keeping the real world applications in mind, the functionalities developed in this project for the multipurpose microphone array are real-time capable. Moreover, multiple devices can record in synchronous through network-based synchronization. This project also developed a Control Terminal For efficient management of deployed devices. Numerous end-to-end tests have verified that the system works correctly in real time condition. Despite various challenges, this project has achieved its original objectives of developing a multipurpose microphone array that is easy to customize and flexible to deploy and demonstrated its potential of real-time Automatic Speech Recognition.||URI:||http://hdl.handle.net/10356/70444||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||SCSE Student Reports (FYP/IA/PA/PI)|
Updated on Jun 19, 2021
Updated on Jun 19, 2021
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.