ACappellaSet: A Multilingual A Cappella Dataset for Source Separation and AI-assisted Rehearsal Tools

Authors: Ting-Yu Pan, Phyllis Ju, Hao-Wen Dong

A cappella music presents unique challenges for source separation due to its diverse vocal styles and the presence of vocal percussion. Current a cappella datasets are limited in size and diversity, hindering the development of robust source separation models. We are building ACappellaSet, a collection of 55 professionally recorded a cappella songs performed by three professional groups. Preliminary experiment results showed that fine-tuning Demucs on ACappellaSet substantially improves vocal percussion (VP) separation, raising VP SDR from 5.22 dB to 7.62 dB. We envision that ACappellaSet can be further strengthened through AI-driven dataset augmentation. ACappellaSet has the potential for a cappella / choir source separation and supporting tools for asynchronous a cappella rehearsals.

ACappellaSet Paper AcaMate Paper

VP Separation Demo

Start (SATB + VP, Korean, Dry)

Original Mixture

Ground Truth SATB

Separated SATB (Demucs, Pre-trained)

Separated SATB (Demucs, Fine-tuned on ACappellaSet)

Ground Truth VP

Separated VP (Demucs, Pre-trained)

Separated VP (Demucs, Fine-tuned on ACappellaSet)

頭前 (SATB + VP, Haka, Dry)

Original Mixture

Ground Truth SATB

Separated SATB (Demucs, Pre-trained)

Separated SATB (Demucs, Fine-tuned on ACappellaSet)

Ground Truth VP

Separated VP (Demucs, Pre-trained)

Separated VP (Demucs, Fine-tuned on ACappellaSet)

AcaMate: Supporting Novice A Cappella Singers in Iterative Individual Practice

Novice singers in collegiate A cappella groups often struggle with individual practice due to limited guidance and the challenges of asynchronous rehearsal. We propose AcaMate, a system designed to support individual practice by integrating group recordings, visualizing musical patterns across voice parts, and providing intuitive feedback to guide iterative and deliberate practice.

A pipeline of AcaMate The interface of AcaMate