Brass dataset for AI-Generated Music
The Brass dataset is a compilation of compositions that are carefully annotated, providing insights into chord progressions, instrumentation choices, key signatures, tempo variations, and precise timestamps. Such annotations serve as invaluable resources for training models in generative AI music, enabling the creation of algorithmically generated compositions that faithfully replicate the power and grandeur of live brass performances.

It serves as a cornerstone for Music Information Retrieval (MIR), facilitating tasks such as instrument recognition, harmonic analysis, and melody extraction. Detailed metadata accompanying each audio track empowers researchers to develop algorithms capable of dissecting complex brass arrangements with precision, thereby advancing the frontiers of music analysis and synthesis. Whether exploring generative music, delving into MIR tasks, or pioneering source separation techniques, researchers can leverage this dataset to unlock the transformative potential of brass instruments in the digital age.
Dataset Highlights

Diverse Brass Ensemble
The "Brass" dataset features a diverse array of brass instruments, including trumpets, trombones, French horns, and tubas, offering a comprehensive resource for composers, arrangers, and music producers.

Authentic Performance Dynamics
With meticulously captured performances, the dataset includes a range of articulations, dynamics, and playing techniques specific to each brass instrument, ensuring realism and depth in musical compositions.

Ethically Sourced and Copyright-Cleared
Committed to ethical standards, all content in the dataset is responsibly sourced and copyright-cleared, providing users with peace of mind while upholding ethical practices in music creation.

Premium Audio Quality
Each recording in the dataset meets stringent quality standards, offering pristine sound quality and fidelity, empowering composers to create professional-grade compositions with confidence and ease.
