Skip navigation
Please use this identifier to cite or link to this item:
Title: HMM Beats
Authors: Ryan, Daniel
Advisors: Pritchard, David
Department: Computer Science
Class Year: 2013
Abstract: This paper discusses the motivation, design, implementation, and evaluation of a new electronic beat sequencing system called HMM Beats. HMM Beats is inspired by recent work in machine learning on user drum beat prediction. HMM Beats is an interactive system through which a user composes beat loops through an iterative process with a learned computer agent. Based upon the beats of the drums that the user has already composed, the agent can be queried as to what the beats of the another drum should be. The agent's knowledge is stored in Hidden Markov Models (HMMs) built from a statistical analysis of preexisting midi drum loops. Through the use of this system, a user can freely compose while at the same time leverage the compositional aid of the computer. The user and computer agent thus iterate back and forth in realtime, building interesting electronic beat structures. The physical interface of HMM Beats is the QuNeo 3D, a multi-touch customizable midi controller. The main control unit of HMM Beats is a Max/MSP patch. The machine learning and beat queries take place in a python script utilizing the General Hidden Markov Model (GHMM) library.
Extent: 50 pages
Access Restrictions: Walk-in Access. This thesis can only be viewed on computer terminals at the Mudd Manuscript Library.
Type of Material: Princeton University Senior Theses
Language: en_US
Appears in Collections:Computer Science, 1988-2016

Files in This Item:
File SizeFormat 
Daniel Ryan.pdf3.91 MBAdobe PDF    Request a copy

Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.