数据堂—2,028小时普通话手机采集语音数据

2,028小时普通话手机采集语音数据是由4,787名中国本土发音人参与录制，覆盖多个省份，男女均衡。语音均采用标准普通话录制。录音内容丰富，涵盖手机语音助手交互、智能家居命令、车载命令词、数字等多种类别。2,028小时普通话手机采集语音数据可以精准匹配智能家居、智能车载等实际应用场景。

The 2,028-hour Mandarin speech dataset collected via mobile devices is recorded by 4,787 native Chinese speakers, covering multiple provinces with a balanced gender ratio. All recordings are made in standard Mandarin. The content of the recordings is diverse, covering various categories including mobile voice assistant interactions, smart home commands, in-vehicle command terms, digits, and more. This dataset can accurately match practical application scenarios such as smart home and intelligent in-vehicle systems.

搜集汇总

背景与挑战

背景概述

该数据集包含2,028小时通过手机采集的标准普通话语音数据，由4,787名母语说话者录制，性别分布均衡，内容涵盖语音助手交互、智能家居命令等多种实际应用场景。数据格式为16kHz、16位未压缩的.wav单声道文件，标注准确率不低于95%，适用于测试标准普通话语音识别模型，版权归DataTang所有，为商业数据。