172小时美式英语分轨自然对话手机采集语音数据

美式英语分轨自然对话手机采集语音数据，基于常见主题来模拟录制。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，由来自不同地域和文化背景的美国本土人录制，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。

American English multi-track natural conversational speech data collected via mobile phones, recorded by simulating natural conversations centered on common daily themes. This dataset is annotated with multiple attributes including text transcripts, sentence-level timestamps, speaker identity, gender and other relevant information. It was recorded by native American speakers from diverse regional and cultural backgrounds, ensuring high data accuracy. This dataset serves as a rich resource for research and applications related to speech recognition. Verified by multiple AI companies, it helps models achieve excellent performance when facing real-world diversity. We strictly comply with data protection regulations and privacy rules, ensuring that user privacy and legitimate rights and interests are safeguarded throughout the entire process of data collection, storage and usage. All data adheres to the requirements of GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含172小时美式英语自然对话语音，通过手机采集并基于常见主题模拟录制，标注了文本、时间戳、说话人身份等多重属性。数据由来自不同地域的美国本土人提供，准确性高，适用于语音识别研究及多人对话、会议转写等应用场景，且遵循严格的数据保护法规。