Skip to yearly menu bar Skip to main content

Workshop: Machine Learning for Audio

The Song Describer Dataset: a Corpus of Audio Captions for Music-and-Language Evaluation

Ilaria Manco · Benno Weck · Seungheon Doh · Yixiao Zhang · Dmitry Bogdanov · Yusong Wu · Ke Chen · Philip Tovstogan · Emmanouil Benetos · Elio Quinton · George Fazekas · Juhan Nam · Minz Won


We introduce the Song Describer dataset (SDD), a new crowdsourced corpus of high-quality audio-caption pairs, designed for the evaluation of music-and-language models. The dataset consists of 1.1k natural language descriptions of 706 music recordings, all publicly accessible and released under Creative Common licenses. To showcase the use of our dataset, we benchmark popular models on three key music-and-language tasks (music captioning, text-to-music generation and music-language retrieval). Our experiments highlight the importance of cross-dataset evaluation and offer insights into how researchers can use SDD to gain a broader understanding of model performance.

Chat is not available.