The article ‘Are words enough?’ explores the semantic conditioning of affective music generation. It discusses the growing use of natural language processing and deep learning models in music generation, conditioned by emotions. The authors review two main paradigms used in automatic music generation: rules-based and machine-learning models, particularly deep learning architectures that generate music from textual descriptions. The article argues that despite the challenges, such as language ambiguity and the subjectivity of emotions, the use of deep learning with natural language can potentially impact the creative industries by providing powerful tools for prompting and generating new musical works.
Publication date: 8 Nov 2023
Project Page: https://aimc2023.pubpub.org/pub/9z68g7d2
Paper: https://arxiv.org/pdf/2311.03624
