EnvSDD-Development

This is development set of Environment Sound Deepfake Detection (EnvSDD) dataset. Audio generation systems now create very realistic soundscapes that can enhance media production, but also pose potential risks. Several studies have examined deepfakes in speech or singing voice. However, environmental sounds have different characteristics, which may make methods for detecting speech and singing deepfakes less effective for real-world sounds. In addition, existing datasets for environmental sound deepfake detection are limited in scale and audio types. To address this gap, we introduce EnvSDD, the first large-scale curated dataset designed for this task, consisting of 45.25 hours of real and 316.74 hours of fake audio. The test set includes diverse conditions to evaluate the generalizability, such as unseen generation models and unseen datasets. We also propose an audio deepfake detection system, based on a pre-trained audio foundation model. Results on EnvSDD show that our proposed system outperforms the state-of-the-art systems from speech and singing domains. Please cite our paper if you want to use the data: Han Yin, et al. "EnvSDD: Benchmarking Environmental Sound Deepfake Detection." Interspeech. 2025.本数据集为环境声音深度伪造检测（Environment Sound Deepfake Detection, EnvSDD）数据集的开发集。 当前音频生成系统可生成极具真实感的音景，既能为媒体制作赋能，也潜藏着潜在风险。已有多项研究针对语音或歌唱类深度伪造开展了相关探索，但环境声音具备独特的声学特性，使得针对语音与歌唱深度伪造的检测方法，在真实环境声音的检测场景中往往难以取得理想效果。此外，现有环境声音深度伪造检测数据集在规模与音频类型覆盖范围上均存在明显局限。 为填补这一研究空白，我们推出了EnvSDD——本领域首个精心打造的大规模专用数据集，其包含45.25小时真实音频与316.74小时伪造音频。该数据集的测试集设置了多样化的评估条件，以检验模型的泛化能力，涵盖未知生成模型与未知数据集等场景。我们还提出了一种基于预训练音频基础模型的音频深度伪造检测系统。在EnvSDD数据集上的实验结果表明，我们提出的系统性能优于语音与歌唱领域的现有顶尖检测系统。 若需使用本数据集，请引用如下文献：Han Yin, et al. "EnvSDD: Benchmarking Environmental Sound Deepfake Detection". Interspeech. 2025.