Wav2Aug improves speech AI tasks with versatile waveform augmentation
Wav2Aug: A Drop-In Speech Augmentation Policy
Sound
Summary
People use data tricks called augmentations to help machines understand speech better. The common method, SpecAugment, works on sound images but can't help for all speech tasks. The authors created Wav2Aug, which changes the raw sound wave in random ways that work well for many speech tasks without extra tuning. Tests on different challenges show Wav2Aug matches or beats SpecAugment, especially beyond speech recognition. Training with two versions of the altered sound at once also helps boost performance efficiently.
What this means in practice
- •For speech system developers: Improve training of voice recognition and other speech AI systems using Wav2Aug as a plug-in data augmentation module.
- •For language technology teams: Enhance performance across varied speech processing tasks by integrating a uniform, pre-tuned waveform augmentation method that avoids task-specific changes.
Authors
Gianfranco Dumoulin Bertucci, Mirco Ravanelli
Abstract
SpecAugment is the dominant speech augmentation policy, but researchers have studied it primarily for automatic speech recognition (ASR), and it operates on spectrograms, limiting its applicability to other speech tasks and waveform models. We propose Wav2Aug, a drop-in waveform augmentation policy that randomly samples two augmentations from a calibrated pool of waveform transformations. We calibrate the augmentation hyperparameters once and reuse them across tasks and architectures, eliminating task-specific tuning. Experiments on ten speech benchmarks show that Wav2Aug achieves ASR performance comparable to SpecAugment while improving over it on average across non-ASR tasks, yielding an average relative error reduction of 12.8% over the strongest compatible SpecAugment baseline. We further show that multi-view training provides additional gains, with two augmented views offering the best tradeoff between performance and computational cost.