Unmasking Choir Voices: Exploring Vocal Correctness in Collective Singing

Engaging in choir performances enables performers to collaborate and produce harmonious music, presenting a unified voice stemming from many sources. The process demands a deep understanding of music theory and extensive practice, with each singer contributing uniquely to construct a rich and complex performance. However, vulnerabilities to errors emerge within this harmonious collaboration, challenging even the most experienced conductors to pinpoint the origin of a specific voice amidst the collective singing. To overcome such hurdles, we developed an automated system capable of collecting, analyzing, and providing feedback to conductors as well as choir members. Our project, titled “Unmasking Choir Voices: Exploring Vocal Correctness in Collective Singing”, aims to address these challenges by developing a method to collect audio data from individual choir singers through a microphone, perform the necessary audio comparison with the reference (true) audio, and distinguish those individuals singing correctly and in tune from those needing improvement. We successfully implemented this system and presented the results through a user-friendly web application for the conductor to observe and assess individuals’ performances. The goal of the system was to assist conductors in identifying incorrect singing within a choir while also empowering individual singers to recognize and rectify their own mistakes during rehearsals. Our testing concluded that the system achieved an accuracy rate of 84.7%, significantly enhancing the overall quality of choir performances and the individual growth of choir members.

Project Details

[photo]