An End-to-End Approach for Chord-Conditioned Song Generation
This work addresses the challenge of constrained control in AI-generated music for applications in creative industries, representing an incremental improvement over prior methods like Jukebox.
The paper tackles the problem of generating songs with vocals and accompaniment from lyrics by introducing chord conditioning to improve control and musical performance, resulting in a model that outperforms existing approaches in these aspects.
The Song Generation task aims to synthesize music composed of vocals and accompaniment from given lyrics. While the existing method, Jukebox, has explored this task, its constrained control over the generations often leads to deficiency in music performance. To mitigate the issue, we introduce an important concept from music composition, namely chords, to song generation networks. Chords form the foundation of accompaniment and provide vocal melody with associated harmony. Given the inaccuracy of automatic chord extractors, we devise a robust cross-attention mechanism augmented with dynamic weight sequence to integrate extracted chord information into song generations and reduce frame-level flaws, and propose a novel model termed Chord-Conditioned Song Generator (CSG) based on it. Experimental evidence demonstrates our proposed method outperforms other approaches in terms of musical performance and control precision of generated songs.