数据堂—849小时普通话交互类手机采集语音数据

中国普通话语音数据_交互场景（手机），在真实的家居场景下进行录制，录音内容以家居指令、功能性助手和唤醒词为主，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, PIPLChinese Mandarin Speech Dataset - Interactive Scenarios (Mobile Phones). Recorded in real home scenarios, the recorded content mainly focuses on home commands, functional assistant utterances and wake words, with high accuracy. This dataset provides abundant resources for speech recognition-related research and applications. Verified by multiple AI companies, it helps models achieve excellent performance when dealing with the diversity of the real world. We strictly abide by data protection laws, regulations and privacy rules to ensure the protection of user privacy and legitimate rights and interests during the entire process of data collection, storage and usage. All data complies with GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含849小时通过手机采集的普通话交互语音，专门用于普通话关键词检测模型的测试任务。数据来自998名参与者，每人录制约800句，涵盖家庭场景命令、功能助手和唤醒词等内容，采用48kHz、16bit单通道WAV格式。