VIVAE

情感和情感语料库 (VIVAE) 的可变强烈发声由一组人类非言语情感发声组成。全套包括1085个音频文件，具有11个扬声器，表达三个积极 (成就/胜利，性快感和惊喜) 和三个消极 (愤怒，恐惧，身体疼痛) 的情感状态，每个参数从低到峰值的情感强度变化。较小的核心480文件集表示基于判断的真实性选择的全套 (6个情感x 4个强度x 10个扬声器x 2个项目) 的完全交叉子样本。

The Variable Intensity Emotional Vocalizations Corpus (VIVAE) consists of a set of human non-verbal emotional vocalizations. The full dataset includes 1,085 audio files from 11 speakers, expressing six emotional states: three positive ones (achievement/victory, sexual pleasure, and surprise) and three negative ones (anger, fear, and physical pain), with each emotion varying in intensity from low levels to peak values. The smaller core subset of 480 files represents a fully crossed subsample of the full dataset selected based on perceptual authenticity, following the design of 6 emotions × 4 intensity levels × 10 speakers × 2 items.

搜集汇总

背景与挑战

背景概述

VIVAE是一个情感和情感语料库数据集，包含1085个音频文件，涉及11个扬声器表达6种情感状态，每种情感状态有从低到峰值的情感强度变化。数据集由纽约大学·马克斯普朗克经验美学研究所发布，适用于情感分析和语音识别研究。