EmoFilm (Emotional speech from Films)

EmoFilm是多语言情感语音语料库，包括以英语、意大利语和西班牙语生成的1115音频实例。音频片段 (平均长度为3.5秒和std 1.2秒) 以波格式 (未压缩，单声道，48 khz采样率和16位) 从43部电影 (英文原件及其配音意大利语和西班牙语版本) 中提取。考虑了喜剧，戏剧，恐怖和惊悚片等类型; 考虑了愤怒，鄙视，幸福，恐惧和悲伤的情绪状态。

EmoFilm is a multilingual emotional speech corpus containing 1,115 audio instances generated in English, Italian and Spanish. These audio clips, with an average duration of 3.5 seconds and a standard deviation (std) of 1.2 seconds, are extracted from 43 films including their original English versions and dubbed Italian and Spanish counterparts, and stored in WAV format (uncompressed, mono channel, 48 kHz sampling rate and 16-bit depth). The corpus covers film genres such as comedy, drama, horror and thriller, as well as five emotional states: anger, disgust, happiness, fear and sadness.

搜集汇总

背景与挑战

背景概述

EmoFilm是一个多语言情感语音数据集，包含1115个来自电影的音频片段，覆盖三种语言和五种情绪状态，适用于情感语音分析研究。