数据堂—1,796小时德语手机采集语音数据

1,796小时德语手机采集语音数据，时长1796小时，由3,442名德国本土发音人参与录制；录音文本由语言专家参与设计，涵盖通用、交互、车载、家居等多类别；文本经过人工校对，准确率高；本套德语手机采集语音数据可应用于语音识别、机器翻译、声纹识别

This German mobile-collected speech dataset has a total duration of 1,796 hours, recorded by 3,442 native German speakers. The recording texts were designed with the participation of language experts, covering multiple categories including general scenarios, interactive scenarios, in-vehicle scenarios and smart home scenarios. All texts have undergone manual proofreading with high accuracy. This dataset can be applied to speech recognition, machine translation and speaker verification.

搜集汇总

背景与挑战

背景概述

该数据集包含1,796小时的德语语音数据，通过移动设备采集自3,442名母语者，录音内容涵盖一般话题、交互场景和命令短语等类别，所有文本均经过人工校对以确保高精度。它适用于语音识别、机器翻译和说话人验证等任务。