Mandarin Conversational Speech Data by Mobile Phone and Voice Recorder - 1,351 Hours

1950 speakers participated in the recording, and conducted face-to-face communication in a natural way. They had free discussion on a number of given topics, with a wide range of fields. The voice was natural and fluent, in line with the actual dialogue scene. Text is transcribed manually, with high accuracy.Format：Mobile phone: 16kHz, 16bit, mono channel, .wav; Voice recorder: 44.1kHz, 16bit, dual channel, .wavRecording Environment：quiet indoor environment, without echoRecording content：dozens of topics are specified, and the speakers make dialogue under those topics while the recording is performedDemographics：1,950 people; 66% speakers of all are in the age group of 16-25; 962 speakers of them spoke in groups of two speakers, 312 speakers of them spoke in groups of three speakers, 396 speakers of them spoke in groups of four speakers, and the other 280 speakers spoke in groups of five speakersAnnotation：annotating for the transcription text, speaker identification and genderDevice：mobile phone and voice recorderLanguage：MandarinApplication scenarios：speech recognition; voiceprint recognitionAccuracy rate：97%

本数据集由1950位参与者录制，他们以自然的方式进行面对面交流。参与者就众多指定主题展开自由讨论，涵盖广泛领域。录音中的语音自然流畅，与实际对话场景相契合。文本内容系人工转录，确保了高精度。录音格式：手机录音为16kHz，16位，单声道，.wav格式；录音设备为44.1kHz，16位，双声道，.wav格式。录音环境为安静室内环境，无回声。录音内容：指定了数十个主题，参与者在这些主题下进行对话。人口统计学特征：共1950人参与，其中66%的参与者年龄在16至25岁之间；962位参与者以两人一组进行对话，312位参与者以三人一组进行对话，396位参与者以四人一组进行对话，其余280位参与者以五人一组进行对话。标注内容：对转录文本、说话人识别和性别进行标注。设备：手机和录音设备。语言：普通话。应用场景：语音识别；声纹识别。准确率：97%。