Conference paper Open Access

Crowd-driven Music: Interactive and Generative Approaches using Machine Vision and Manhattan

Nash, Chris

Editor(s)
Michon, Romain; Schroeder, Franziska

This paper details technologies and artistic approaches to crowd-driven music, discussed in the context of a live public installation in which activity in a public space (a busy railway platform) is used to drive the automated composition and performance of music. The approach presented uses realtime machine vision applied to a live video feed of a scene, from which detected objects and people are fed into Manhattan (Nash, 2014), a digital music notation that integrates sequencing and programming to support the live creation of complex musical works that combine static, algorithmic, and interactive elements. The paper discusses the technical details of the system and artistic development of specific musical works, introducing novel techniques for mapping chaotic systems to musical expression and exploring issues of agency, aesthetic, accessibility and adaptability relating to composing interactive music for crowds and public spaces. In particular, performances as part of an installation for BBC Music Day 2018 are described. The paper subsequently details a practical workshop, delivered digitally, exploring the development of interactive performances in which the audience or general public actively or passively control live generation of a musical piece. Exercises support discussions on technical, aesthetic, and ontological issues arising from the identification and mapping of structure, order, and meaning in non-musical domains to analogous concepts in musical expression. Materials for the workshop are available freely with the Manhattan software.
Files (140.7 MB)
Name Size
nime2020_paper49.mp4
md5:d368c3693b69548233d0db4b8a395049
138.8 MB Download
nime2020_paper49.pdf
md5:5bb5414aa899b54d089b6a21fe917bcd
1.9 MB Download
nime2020_paper49.srt
md5:52412458f5e4786334d15675949a1ef4
15.2 kB Download
71
32
views
downloads
All versions This version
Views 7171
Downloads 3232
Data volume 470.5 MB470.5 MB
Unique views 5656
Unique downloads 2626

Share

Cite as