数据堂—500小时德语自然对话手机采集语音数据

500小时德语自然对话手机采集语音数据由约750名德国本土人参与录制，录音人男女比例均衡，500小时德语自然对话手机采集语音数据无预设语料，为确保对话的流畅自然，由录音人根据自己熟悉的话题展开对话并录制

This dataset consists of 500 hours of German natural conversational speech data collected via mobile phones. Recorded by approximately 750 native German speakers with a balanced gender ratio, the dialogues were not pre-scripted. To ensure the fluency and naturalness of the conversations, participants initiated and recorded talks based on topics familiar to them.

搜集汇总

背景与挑战

背景概述

该数据集包含500小时的德语自然对话语音数据，由约750名母语者通过手机采集，对话基于参与者熟悉的话题以确保流畅自然。数据格式为16kHz/16bit单声道WAV，主要用于测试德语自然对话语音数据模型。