Towards a Music Language Model for Audio Analysis
Publisher
Publisher URL
Metadata
Show full item recordAbstract
Polyphonic Automatic Music Transcription remains a challenging problem. Many studies focus on the extraction of features from audio signals; we focus here on Music Language Models that help turn those features into a musically meaningful piano-roll representation. We review related work on music language models, and propose ways to improve it.