数据堂—215小时美式英语手机采集语音数据

215小时美式英语手机采集语音数据,采集349名美式英语发音人，均为美国本地人员。录音环境安静，录音内容丰富，覆盖经济，娱乐，新闻，口语等多个领域。本套美式英语手机采集语音数据_朗读由人工精准转写，并标注语音起止时间点

This dataset contains 215 hours of American English speech data collected via mobile devices. The data was gathered from 349 American English speakers, all of whom are native residents of the United States. The recording environments were quiet, and the recording content is diverse, covering multiple fields including economy, entertainment, news, and daily spoken English. The read speech segments in this mobile-collected American English speech dataset have been accurately manually transcribed, with the start and end timestamps of each speech segment annotated.

搜集汇总

背景与挑战

背景概述

该数据集包含215小时美式英语语音，通过手机采集自349名美国本土说话者，录音环境安静，涵盖经济、娱乐等多主题，并包含人工转录和时间戳。它主要用于美式英语语音识别模型的评估任务。