All English Audio Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/SaSs7/Dataset
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含英语真实语音和模仿语音的音频样本，专门为分类伪造语音而设计。该数据集共有1517个样本，分为训练样本1383个和测试样本134个，样本命名规则能明确指示语音类型（真实或伪造）。此外，该数据集的规模为1860个音频样本，其任务是对语音进行检测，以区分真实语音与模仿语音。
This dataset contains audio samples of authentic English speech and imitated (forged) speech, specifically designed for forged speech classification tasks. It comprises 1517 samples in total, which are split into 1383 training samples and 134 test samples. The naming convention for the samples clearly indicates the speech category (authentic or forged). Additionally, the dataset is stated to have a total of 1860 audio samples, with its core task being speech detection to distinguish between authentic and imitated/forged speech.
提供机构：
Authors of the paper
