arXiv:2408.15176v5 Announce Type: replace-cross 
Abstract: We present a unified framework for automatic multitrack music arrangement that enables a single pre-trained symbolic music model to handle diverse arrangement scenarios, including reinterpretation, simplification, and additive generation. At its core is a segment-level reconstruction objective operating on token-level disentangled content and style, allowing for flexible any-to-any instrumentation transformations at inference time. To support track-wise modeling, we introduce REMI-z, a structured tokenization scheme for multitrack symbolic music that enhances modeling efficiency and effectiveness for both arrangement tasks and unconditional generation. Our method outperforms task-specific state-of-the-art models on representative tasks in different arrangement scenarios -- band arrangement, piano reduction, and drum arrangement, in both objective metrics and perceptual evaluations. Taken together, our framework demonstrates strong generality and suggests broader applicability in symbolic music-to-music transformation.

تم تقديم إطار جديد لترتيب الموسيقى المتعددة المسارات تلقائيًا، مما يسمح لنموذج مدرب مسبقًا واحد بالتكيف مع سيناريوهات متنوعة مثل إعادة التفسير والتبسيط. هذه الابتكار مهم لأنه يتيح تحولات مرنة في الآلات الموسيقية، مما يسهل على المبدعين استكشاف أنماط وترتيبات موسيقية متنوعة.

Se ha presentado un nuevo marco para el arreglo musical multitrack automático, que permite a un solo modelo preentrenado adaptarse a diversos escenarios como la reinterpretación y la simplificación. Esta innovación es significativa, ya que permite transformaciones flexibles en la instrumentación musical, facilitando a los creadores la exploración de diversos estilos y arreglos musicales.

Un nouveau cadre pour l'arrangement musical multitrack automatique a été introduit, permettant à un seul modèle pré-entraîné de s'adapter à divers scénarios tels que la réinterprétation et la simplification. Cette innovation est importante car elle permet des transformations flexibles dans l'instrumentation musicale, facilitant ainsi l'exploration de styles et d'arrangements musicaux variés.

A new framework for automatic multitrack music arrangement has been introduced, allowing a single pre-trained model to adapt to various scenarios like reinterpretation and simplification. This innovation is significant as it enables flexible transformations in music instrumentation, making it easier for creators to explore diverse musical styles and arrangements.

Unifying Symbolic Music Arrangement: Track-Aware Reconstruction and Structured Tokenization

One More Thing in AI – Your Shortcut to AI Mastery

Unifying Symbolic Music Arrangement: Track-Aware Reconstruction and Structured Tokenization

Was this article worth reading? Share it

One More Thing in AI

Airparser

Music Maker AI

AI Powered Music Generator

Sound Of Meme

Summify

Ready to build your own newsroom?