SVDD Challenge Dataset

该数据集主要关注真实和深度伪造的歌唱声音录音，旨在用于深度伪造检测。参与者需要提交测试集的得分和系统描述，这些信息将在挑战赛的网站上公布。挑战赛将根据等错误率（EER）来评估提交的成果，任务重点是歌唱声音的深度伪造检测。

This dataset primarily focuses on genuine and deepfake singing voice recordings, targeting deepfake detection tasks. Participants are required to submit their test set scores and system descriptions, which will be published on the challenge's official website. The challenge will evaluate all submitted entries using the Equal Error Rate (EER) as the evaluation metric, with the core task being deepfake detection for singing voices.