Text Conditioned Symbolic Drumbeat Generation using Latent Diffusion Models

TL;DR

This paper presents a novel text-conditioned drumbeat generation method using Latent Diffusion Models, multimodal contrastive learning, and a multi-resolution LSTM, producing diverse, prompt-appropriate, and high-quality drumbeats comparable to human compositions.

Contribution

It introduces a new approach combining LDMs, contrastive learning, and a multi-resolution LSTM for text-conditioned drumbeat generation, enhancing diversity and quality.

Findings

Generated drumbeats are diverse and align well with prompts.

The method produces drumbeats comparable in quality to human-created ones.

The approach speeds up generation using latent space diffusion.

Abstract

This study introduces a text-conditioned approach to generating drumbeats with Latent Diffusion Models (LDMs). It uses informative conditioning text extracted from training data filenames. By pretraining a text and drumbeat encoder through contrastive learning within a multimodal network, aligned following CLIP, we align the modalities of text and music closely. Additionally, we examine an alternative text encoder based on multihot text encodings. Inspired by musics multi-resolution nature, we propose a novel LSTM variant, MultiResolutionLSTM, designed to operate at various resolutions independently. In common with recent LDMs in the image space, it speeds up the generation process by running diffusion in a latent space provided by a pretrained unconditional autoencoder. We demonstrate the originality and variety of the generated drumbeats by measuring distance (both over binary pianorolls and in the latent space) versus the training dataset and among the generated drumbeats. We also assess the generated drumbeats through a listening test focused on questions of quality, aptness for the prompt text, and novelty. We show that the generated drumbeats are novel and apt to the prompt text, and comparable in quality to those created by human musicians.