Advancing Natural-language Based Audio Retrieval With Passt And Large Audio-caption Data Sets

Abstract

This work presents a text-to-audio-retrieval system based on pre-trained text and spectrogram transformers. Our method projects recordings and textual descriptions into a shared audio-caption space in which related examples from different modalities are close. Through a systematic analysis, we examine how each component of the system influences retrieval performance. As a result, we identify two key components that play a crucial role in driving performance: the self-attention-based audio encoder for audio embedding and the utilization of additional human-generated and synthetic data sets during pre-training. We further experimented with augmenting ClothoV2 captions with available keywords to increase their variety; however, this only led to marginal improvements. Our system ranked first in the 2023's DCASE Challenge, and it outperforms the current state of the art on the ClothoV2 benchmark by 5.6 pp. mAP@10.

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyprimus2023advancing

Related papers

• Improving Natural-language-based Audio Retrieval With Transfer Learning And Audio & Text Augmentations (2022)0.00
• Matching Text And Audio Embeddings: Exploring Transfer-learning Strategies For Language-based Audio Retrieval (2022)0.00
• Audio Retrieval With Natural Language Queries: A Benchmark Study (2021)16.29
• Learning Audio-video Modalities From Image Captions (2022)12.54
• Dissecting Temporal Understanding In Text-to-audio Retrieval (2024)3.58
• Estimated Audio-caption Correspondences Improve Language-based Audio Retrieval (2024)0.00
• CLASP: Contrastive Language-speech Pretraining For Multilingual Multimodal Information Retrieval (2024)0.00
• CLIPS: An Enhanced CLIP Framework For Learning With Synthetic Captions (2024)0.00