Comparing Self-Supervised Learning Models Pre-Trained on Human Speech and Animal Vocalizations for Bioacoustics Processing

TL;DR

This study compares self-supervised learning models pre-trained on human speech and animal vocalizations to determine their effectiveness in bioacoustic tasks, finding minimal performance differences and highlighting the robustness of speech-pretrained models.

Contribution

It provides a comprehensive comparison of SSL models pre-trained on speech versus animal sounds for bioacoustic applications, showing that speech models are generally sufficient.

Findings

Bioacoustic pre-training offers marginal improvements over speech pre-training.

Fine-tuning on ASR tasks has mixed effects on bioacoustic performance.

Speech-pretrained SSL models are robust and often adequate without extensive fine-tuning.

Abstract

Self-supervised learning (SSL) foundation models have emerged as powerful, domain-agnostic, general-purpose feature extractors applicable to a wide range of tasks. Such models pre-trained on human speech have demonstrated high transferability for bioacoustic processing. This paper investigates (i) whether SSL models pre-trained directly on animal vocalizations offer a significant advantage over those pre-trained on speech, and (ii) whether fine-tuning speech-pretrained models on automatic speech recognition (ASR) tasks can enhance bioacoustic classification. We conduct a comparative analysis using three diverse bioacoustic datasets and two different bioacoustic tasks. Results indicate that pre-training on bioacoustic data provides only marginal improvements over speech-pretrained models, with comparable performance in most scenarios. Fine-tuning on ASR tasks yields mixed outcomes, suggesting that the general-purpose representations learned during SSL pre-training are already well-suited for bioacoustic tasks. These findings highlight the robustness of speech-pretrained SSL models for bioacoustics and imply that extensive fine-tuning may not be necessary for optimal performance.