Abstract
The ASVspoof Dataset is one of the most established datasets for training and benchmarking systems designed for the detection of spoofed audio and audio deepfakes. However, we observe an uneven distribution of leading silence in dataset's training and test data, which hints at the target label: Bona-fide instances tend to have significantly longer leading silences than spoofed instances. This could be problematic, since a model may learn to only, or at least partially, base its decision on the length of the leading silence (similar to the issue with the Pascal VOC 2007 dataset, where all images of horses also contained a specific watermark). In this paper, we explore this phenomenon in depth. We train a number of networks on only a) the length of the leading silence and b) with and without leading silence. Results show that models trained on only the length of the leading silence perform suspiciously well: They achieve up to 85% percent accuracy and an equal error rate (EER) of 0.15 on the 'eval' split of the data. Conversely, when training strong models on the full audio files, we observe that removing leading silence during preprocessing dramatically worsens performance (EER increases from 0.05 to 0.2). This could indicate that previous work may, in part, have learned only to classify targets based on leading silence. We hope that by sharing these results, the ASV community can further evaluate this phenomenon.
Abstract (translated)
URL
https://arxiv.org/abs/2106.12914
