Speech is Silver, Silence is Golden: What do ASVspoof-trained Models Really Learn?

by   Nicolas M. Müller, et al.

We present our analysis of a significant data artifact in the official 2019/2021 ASVspoof Challenge Dataset. We identify an uneven distribution of silence duration in the training and test splits, which tends to correlate with the target prediction label. Bonafide instances tend to have significantly longer leading and trailing silences than spoofed instances. In this paper, we explore this phenomenon and its impact in depth. We compare several types of models trained on a) only the duration of the leading silence and b) only on the duration of leading and trailing silence. Results show that models trained on only the duration of the leading silence perform particularly well, and achieve up to 85 between 0 and 1). At the same time, we observe that trimming silence during pre-processing and then training established antispoofing models using signal-based features leads to comparatively worse performance. In that case, EER increases from 0.03 (with silence) to 0.15 (trimmed silence). Our findings suggest that previous work may, in part, have inadvertently learned thespoof/bonafide distinction by relying on the duration of silence as it appears in the official challenge dataset. We discuss the potential consequences that this has for interpreting system scores in the challenge and discuss how the ASV community may further consider this issue.


JDI-T: Jointly trained Duration Informed Transformer for Text-To-Speech without Explicit Alignment

We propose Jointly trained Duration Informed Transformer (JDI-T), a feed...

Improving Event Duration Prediction via Time-aware Pre-training

End-to-end models in NLP rarely encode external world knowledge about le...

Expressive, Variable, and Controllable Duration Modelling in TTS

Duration modelling has become an important research problem once more wi...

The Microsoft System for VoxCeleb Speaker Recognition Challenge 2022

In this report, we describe our submitted system for track 2 of the VoxC...

DVR: Micro-Video Recommendation Optimizing Watch-Time-Gain under Duration Bias

Recommender systems are prone to be misled by biases in the data. Models...

Consonant gemination in Italian: the nasal and liquid case

All Italian consonants affected by gemination, that is affricates, frica...

Understanding Prediction Discrepancies in Machine Learning Classifiers

A multitude of classifiers can be trained on the same data to achieve si...

Please sign up or login with your details

Forgot password? Click here to reset