arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2410.23005cs.SDeess.AS
收藏

通过扩散Transformer改进音乐伴奏协同创作

Improving Musical Accompaniment Co-creation via Diffusion Transformers

Sony Computer Science Laboratories(索尼计算机科学实验室);Queen Mary University of London(伦敦玛丽女王大学)

Javier Nistal, Marco Pasini, Stefan Lattner

AI总结:

本研究基于Diff-A-Riff潜扩散模型,通过升级立体声自编码器、替换扩散Transformer架构、优化文本提示跨模态映射及引入一致性训练加速推理,全面提升了音乐伴奏生成的质量、多样性、推理速度与文本控制能力。

摘要:

我们以用于乐器伴奏生成的潜扩散模型Diff-A-Riff(即Diff-A-Riff潜扩散模型)为基础,针对生成质量、多样性、推理速度以及文本驱动控制能力提出了一系列改进方案。首先,我们将底层自编码器升级为支持立体声的模型,以获得更优的保真度,同时用Diffusion Transformer(扩散Transformer)替代潜空间U-Net。此外,我们通过训练一个跨模态预测网络,将文本衍生的CLAP嵌入转换为音频衍生的CLAP嵌入,以此优化文本提示效果。最后,我们采用一致性框架训练潜模型以提升推理速度,在更少去噪步数下仍能达到有竞争力的生成质量。我们在消融实验中使用客观指标将所提模型与原始Diff-A-Riff变体进行对比,结果表明模型在所有目标改进方向上均取得了可观的进展。音频示例可访问:https://sonycslparis.github.io/improved_dar/。

英文摘要:

Building upon Diff-A-Riff, a latent diffusion model for musical instrument accompaniment generation, we present a series of improvements targeting quality, diversity, inference speed, and text-driven control. First, we upgrade the underlying autoencoder to a stereo-capable model with superior fidelity and replace the latent U-Net with a Diffusion Transformer. Additionally, we refine text prompting by training a cross-modality predictive network to translate text-derived CLAP embeddings to audio-derived CLAP embeddings. Finally, we improve inference speed by training the latent model using a consistency framework, achieving competitive quality with fewer denoising steps. Our model is evaluated against the original Diff-A-Riff variant using objective metrics in ablation experiments, demonstrating promising advancements in all targeted areas. Sound examples are available at: https://sonycslparis.github.io/improved_dar/.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑