EAED

EAED is an Egyptian-Arabic emotional speech dataset. The dataset is a semi-natural one as it was collected from five well-known Egyptian TV series. Each audio file ranged in length from 1 to 8 seconds depending on the completion time of the given sentence. The dataset contains six different emotions: happy, sad, angry, neutral, surprised, and fearful. All audio files were recorded using the open source Audacity Software at sampling rate 44.1KHz. Four different human labelers were assigned to hear the recorded audio files in order to annotate/label them. Then, a fifth labeler was assigned for the task of tie-breaking. The number of speakers in the dataset is 79 including 37 males and 42 females.

EAED是一款埃及阿拉伯语情感语音数据集（Egyptian-Arabic emotional speech dataset）。该数据集属于半自然语音数据集，其数据采集自五部知名埃及电视剧集。每条音频文件的时长介于1至8秒之间，具体时长取决于对应语句的完整时长。该数据集涵盖六种情感类别：喜悦、悲伤、愤怒、中性、惊讶与恐惧。所有音频文件均通过开源Audacity软件录制，采样率为44.1kHz。本次标注共配备四名人工标注员，负责对录制的音频文件进行情感标注；随后增设第五名标注员，负责处理标注分歧。该数据集共包含79名说话人，其中男性37名，女性42名。

搜集汇总

背景与挑战

背景概述

EAED是一个埃及阿拉伯语情感语音数据集，包含3,614个音频文件，这些文件从五部埃及电视连续剧中收集，属于半自然数据集。数据集涵盖快乐、悲伤、愤怒、中性、惊讶和恐惧六种情感，由79名说话者（37名男性和42名女性）录制，并经过多标注员标注以确保准确性。该数据集主要用于情感识别研究，适用于人工智能和信号处理领域。