RawBoost: A Raw Data Boosting and Augmentation Method applied to Automatic Speaker Verification Anti-Spoofing

View on arXiv ← Back to list

Authors: Hemlata Tak, Madhu Kamble, Jose Patino, Massimiliano Todisco, Nicholas Evans

Published: 2021-11-08 12:50:51+00:00

AI Summary

RawBoost is a data augmentation method for improving spoofing detection in automatic speaker verification, operating directly on raw waveforms without requiring additional data sources. It enhances a state-of-the-art system by 27% relative performance on the ASVspoof 2021 logical access database, outperformed only by methods using external data or model-level interventions.

Abstract

This paper introduces RawBoost, a data boosting and augmentation method for the design of more reliable spoofing detection solutions which operate directly upon raw waveform inputs. While RawBoost requires no additional data sources, e.g. noise recordings or impulse responses and is data, application and model agnostic, it is designed for telephony scenarios. Based upon the combination of linear and non-linear convolutive noise, impulsive signal-dependent additive noise and stationary signal-independent additive noise, RawBoost models nuisance variability stemming from, e.g., encoding, transmission, microphones and amplifiers, and both linear and non-linear distortion. Experiments performed using the ASVspoof 2021 logical access database show that RawBoost improves the performance of a state-of-the-art raw end-to-end baseline system by 27% relative and is only outperformed by solutions that either depend on external data or that require additional intervention at the model level.

Key findings

RawBoost significantly improved the performance of a state-of-the-art raw waveform-based spoofing detection system by 27% relative. The best results were achieved by combining linear/non-linear convolutive noise and impulsive signal-dependent additive noise. The method is competitive with other state-of-the-art techniques while requiring no external data or model modifications.

Approach

RawBoost augments audio data by adding linear and non-linear convolutive noise, impulsive signal-dependent additive noise, and stationary signal-independent additive noise to simulate real-world telephony variability. This is done directly on raw waveforms without external datasets or model modifications.

Datasets

ASVspoof 2019 Logical Access (LA) dataset (training and development partitions) for training; ASVspoof 2021 LA dataset for evaluation.

Model(s)

RawNet2 (an end-to-end system)

Author countries

France

← Previous