A Unix-based speech data collection platform

Abstract

It is highly desirable to collect speech data from the telephone network via a digital interface. This avoids an additional A/D conversion normally required by analog telephone data collection hardware. A popular solution to this problem is the use of a T1 line which offers 24 digital phone lines. The leading T1 interface for Sun workstations is a system developed by Linkon Corporation. Using the Linkon framework, we have developed a fully-expandable, robust environment for platform-independent collection of telephone speech data. The object-oriented software libraries and intuitive GUI provide powerful tools with which even a novice user can efficiently prototype complex applications. The system is currently being deployed by the Linguistic Data Consortium to collect part of the next Switchboard Corpus.