Libri-adhoc40

Libri-adhoc40是一个同步语音语料库，它通过实际办公环境中40个强同步分布式节点的ad-hoc麦克风阵列从扬声器中收集重播的Librispeech数据。此外，为了为语音前端处理和其他应用提供评估目标，它还在消声室中记录了重播的Librispeech数据。

Libri-adhoc40 is a synchronized speech corpus. It collects replayed Librispeech data from speakers via an ad-hoc microphone array with 40 highly synchronized distributed nodes in real office environments. Additionally, to provide evaluation targets for speech front-end processing and other applications, this corpus also records replayed Librispeech data in an anechoic chamber.

搜集汇总

背景与挑战

背景概述

Libri-adhoc40是一个同步语音语料库，通过40个分布式节点的ad-hoc麦克风阵列在办公环境中收集重播的Librispeech数据。它还在消声室中记录相同数据，为语音前端处理等应用提供评估目标。