Fusing Audio and Metadata Embeddings Improves Language-based Audio Retrieval

TL;DR

This paper presents a hybrid audio retrieval system that combines audio content with metadata embeddings, significantly enhancing language-based audio retrieval accuracy on benchmark datasets.

Contribution

It introduces a novel fusion approach using metadata and explores late and mid-level strategies, improving retrieval performance over content-only methods.

Findings

Metadata fusion improves retrieval accuracy by up to 3.69 percentage points.

Late fusion strategy outperforms other fusion methods.

The approach achieves state-of-the-art results on ClothoV2 and AudioCaps datasets.

Abstract

Matching raw audio signals with textual descriptions requires understanding the audio's content and the description's semantics and then drawing connections between the two modalities. This paper investigates a hybrid retrieval system that utilizes audio metadata as an additional clue to understand the content of audio signals before matching them with textual queries. We experimented with metadata often attached to audio recordings, such as keywords and natural-language descriptions, and we investigated late and mid-level fusion strategies to merge audio and metadata. Our hybrid approach with keyword metadata and late fusion improved the retrieval performance over a content-based baseline by 2.36 and 3.69 pp. mAP@10 on the ClothoV2 and AudioCaps benchmarks, respectively.