2,657小时普通话语音数据_对话（手机）【数据堂】

普通话语音数据_对话（手机），由录音人自由交谈并进行录制。此数据集标注了文本内容、句时间戳、说话人身份、性别等多种属性，由4,000余名来自不同地域的中国本土人录制，准确性高，为语音识别相关研究及应用提供了丰富的资源，经多家AI公司验证：有助于模型面对真实世界的多样性时能够表现出色。我们严格遵循数据保护法规和隐私规定，确保数据采集、存储和使用的过程中维护用户的隐私和合法权益，所有数据均遵循GDPR, CCPA, PIPL

Mandarin conversational speech dataset (collected via mobile phones): All data was recorded by speakers engaging in free spontaneous conversations. This dataset is annotated with multiple attributes including transcript text, sentence-level timestamps, speaker identity, gender, etc. It was recorded by over 4,000 native Chinese speakers from diverse regions, featuring high annotation accuracy and providing abundant resources for speech recognition-related research and applications. Verified by multiple AI companies, this dataset helps models deliver outstanding performance when confronted with the diversity of real-world scenarios. We strictly comply with data protection regulations and privacy rules to safeguard users' privacy and legitimate rights and interests throughout the entire process of data collection, storage and usage. All datasets adhere to GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含2,657小时的普通话对话语音，由4,491名中国本土人员使用手机在安静环境中自然录制而成。数据标注了文本内容、时间戳和说话人身份，准确率达97%，专为语音识别和声纹识别研究设计，但仅限企业合作购买使用。