WhAM: Towards A Translative Model of Sperm Whale Vocalization

TL;DR

This paper introduces WhAM, a transformer-based model that generates realistic sperm whale codas from audio prompts, advancing marine bioacoustics and vocalization modeling.

Contribution

WhAM is the first transformer-based model capable of synthesizing sperm whale codas from audio prompts, trained on extensive real-world data and evaluated with multiple metrics.

Findings

WhAM produces high-fidelity synthetic codas preserving acoustic features.

WhAM's representations perform well on classification tasks.

The model is effective for bioacoustic research and vocalization analysis.

Abstract

Sperm whales communicate in short sequences of clicks known as codas. We present WhAM (Whale Acoustics Model), the first transformer-based model capable of generating synthetic sperm whale codas from any audio prompt. WhAM is built by finetuning VampNet, a masked acoustic token model pretrained on musical audio, using 10k coda recordings collected over the past two decades. Through iterative masked token prediction, WhAM generates high-fidelity synthetic codas that preserve key acoustic features of the source recordings. We evaluate WhAM's synthetic codas using Fr\'echet Audio Distance and through perceptual studies with expert marine biologists. On downstream classification tasks including rhythm, social unit, and vowel classification, WhAM's learned representations achieve strong performance, despite being trained for generation rather than classification. Our code is available at https://github.com/Project-CETI/wham