WoW-Bench: Evaluating Fine-Grained Acoustic Perception in Audio-Language Models via Marine Mammal Vocalizations
This addresses the need for better auditory grounding in audio-language models for real-world tasks involving unfamiliar sounds, though it is incremental as it focuses on benchmarking rather than model improvement.
The paper tackled the problem of evaluating low-level auditory perception in large audio-language models by introducing WoW-Bench, a benchmark using marine mammal vocalizations, and found that state-of-the-art models perform far below human levels.
Large audio language models (LALMs) extend language understanding into the auditory domain, yet their ability to perform low-level listening, such as pitch and duration detection, remains underexplored. However, low-level listening is critical for real-world, out-of-distribution tasks where models must reason about unfamiliar sounds based on fine-grained acoustic cues. To address this gap, we introduce the World-of-Whale benchmark (WoW-Bench) to evaluate low-level auditory perception and cognition using marine mammal vocalizations. WoW-bench is composed of a Perception benchmark for categorizing novel sounds and a Cognition benchmark, inspired by Bloom's taxonomy, to assess the abilities to remember, understand, apply, and analyze sound events. For the Cognition benchmark, we additionally introduce distractor questions to evaluate whether models are truly solving problems through listening rather than relying on other heuristics. Experiments with state-of-the-art LALMs show performance far below human levels, indicating a need for stronger auditory grounding in LALMs.