50小时美国英文情感语音数据_朗读（麦克风）【数据堂】

美国英文情感语音数据_朗读（麦克风），基于给定的脚本朗读并模拟录制，每人约2,100句。录音文本覆盖愤怒、高兴、悲伤等10种情感。录音均采用高保真麦克风录制，语音质量高。本数据适用于情感语音的分析检测，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, PIPLAmerican English Emotional Speech Data - Read Speech (Microphone-Recorded). Collected by having speakers read provided scripts under simulated real-world recording conditions, with approximately 2,100 sentences per speaker. The recorded speech samples cover 10 emotional categories including anger, joy, sadness and others. All recordings are captured using high-fidelity microphones, ensuring high-quality speech. This dataset is suitable for emotional speech analysis and detection. Verified by multiple AI companies, it helps models achieve robust performance when facing real-world diversities. We strictly comply with data protection regulations and privacy rules, ensuring the protection of participants' 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50小时的美国英文情感语音，基于脚本朗读录制，覆盖愤怒、高兴、悲伤等10种情感。数据由20名美国人（男女各半）使用高保真麦克风在安静室内录制，格式为44.1kHz、16bit、未压缩wav单声道，句准确率达95%，适用于情感语音分析检测，但仅限企业合作购买。