99小时普什图语语音数据_对话（手机）【数据堂】

普什图语自然对话手机采集语音数据，基于约二十个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156名来自不同地域和文化背景的阿富汗本土人录制，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

Pashto natural conversational speech data collected via mobile devices, simulated and recorded based on approximately twenty common topics. This dataset is annotated with multiple attributes including text transcripts, sentence-level timestamps, speaker identity, gender and other relevant information. It was recorded by 156 native Afghan speakers from diverse regional and cultural backgrounds, featuring high accuracy, and serves as a rich resource for speech recognition-related research and applications. Verified by multiple AI companies, this dataset enables models to perform excellently when addressing the diversity of real-world scenarios. We strictly adhere to data protection laws and privacy regulations to safeguard users' privacy and legitimate rights and interests throughout the entire process of data collection, storage and utilization. All data is fully compliant with GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含99小时的普什图语自然对话语音，由156名阿富汗本土人通过手机录制，基于约二十个常见主题模拟交谈，并标注了文本内容、时间戳、说话人身份和性别等属性。音频格式为16kHz、16bit未压缩wav单声道，词准确率达95%，适用于语音识别研究，但为商业数据仅限企业合作购买。