TOMI: Transforming and Organizing Music Ideas for Multi-Track Compositions with Full-Song Structure

TL;DR

This paper introduces TOMI, a novel deep learning approach that uses instruction-tuned foundation LLMs to generate, transform, and organize multi-track electronic music with full-song structure, enabling interactive co-creation.

Contribution

The paper presents a new concept hierarchy framework for music generation and a TOMI-based model that integrates with digital audio workstations for improved music composition.

Findings

Produces higher-quality electronic music than baselines.

Generates music with stronger structural coherence.

Enables interactive human-AI co-creation.

Abstract

Hierarchical planning is a powerful approach to model long sequences structurally. Aside from considering hierarchies in the temporal structure of music, this paper explores an even more important aspect: concept hierarchy, which involves generating music ideas, transforming them, and ultimately organizing them--across musical time and space--into a complete composition. To this end, we introduce TOMI (Transforming and Organizing Music Ideas) as a novel approach in deep music generation and develop a TOMI-based model via instruction-tuned foundation LLM. Formally, we represent a multi-track composition process via a sparse, four-dimensional space characterized by clips (short audio or MIDI segments), sections (temporal positions), tracks (instrument layers), and transformations (elaboration methods). Our model is capable of generating multi-track electronic music with full-song structure, and we further integrate the TOMI-based model with the REAPER digital audio workstation, enabling interactive human-AI co-creation. Experimental results demonstrate that our approach produces higher-quality electronic music with stronger structural coherence compared to baselines.