Controlled Singing Voice Deepfake Detection (CtrSVDD)

该数据集包含了日文和中文的唱歌样本，专门为唱歌声音样本而设计。关于详细信息，评估集的标签尚未公布，开发集则被用于测试。数据规模方面，没有提供具体的样本数量。该数据集的任务是针对唱歌声音的深度伪造检测。

This dataset contains singing voice samples in Japanese and Chinese, and is specifically tailored for singing voice audio tasks. Regarding dataset details, the labels of the evaluation split have not been publicly released, while the development split is used for testing. No specific number of samples is provided in terms of dataset scale. The core task of this dataset is deepfake detection targeting singing voice audio.