Benchmarking Sub-Genre Classification For Mainstage Dance Music
This addresses a domain-specific problem for music information retrieval applications like recommendation and DJ curation, but it is incremental as it builds on existing classification tasks with a new dataset and baseline.
The paper tackled the lack of datasets and methods for sub-genre classification in mainstage dance music by introducing a new benchmark with a dataset and baseline, achieving high accuracy with specialized models while state-of-the-art MLLMs struggled.
Music classification, a cornerstone of music information retrieval, supports a wide array of applications. To address the lack of comprehensive datasets and effective methods for sub-genre classification in mainstage dance music, we introduce a novel benchmark featuring a new dataset and baseline. Our dataset expands the scope of sub-genres to reflect the diversity of recent mainstage live sets performed by leading DJs at global music festivals, capturing the vibrant and rapidly evolving electronic dance music (EDM) scene that engages millions of fans worldwide. We employ a continuous soft labeling approach to accommodate tracks blending multiple sub-genres, preserving their inherent complexity. Experiments demonstrate that even state-of-the-art multimodal large language models (MLLMs) struggle with this task, while our specialized baseline models achieve high accuracy. This benchmark supports applications such as music recommendation, DJ set curation, and interactive multimedia systems, with video demos provided. Our code and data are all open-sourced at https://github.com/Gariscat/housex-v2.git.