MorpheuS: constraining structure in automatic music generation

Similar documents
TECHNOLOGIES for digital music have become increasingly

TECHNOLOGIES for digital music have become increasingly

TENSION RIBBONS: QUANTIFYING AND VISUALISING TONAL TENSION

Tension ribbons: Quantifying and visualising tonal tension

Musical Harmonization with Constraints: A Survey. Overview. Computers and Music. Tonal Music

Music Education. Test at a Glance. About this test

CPU Bach: An Automatic Chorale Harmonization System

Outline. Why do we classify? Audio Classification

Musical Creativity. Jukka Toivanen Introduction to Computational Creativity Dept. of Computer Science University of Helsinki

Analysis (MUSI 4211), Spring 2006 Professor: Andrew Davis ( )

Modeling Temporal Tonal Relations in Polyphonic Music Through Deep Networks with a Novel Image-Based Representation

69 A Functional Taxonomy of Music Generation Systems

Music Similarity and Cover Song Identification: The Case of Jazz

Computational Modelling of Harmony

Statistical Modeling and Retrieval of Polyphonic Music

Harmonic Generation based on Harmonicity Weightings

jsymbolic and ELVIS Cory McKay Marianopolis College Montreal, Canada

Palestrina Pal: A Grammar Checker for Music Compositions in the Style of Palestrina

Figured Bass and Tonality Recognition Jerome Barthélemy Ircam 1 Place Igor Stravinsky Paris France

Tristan Und Isolde In Full Score (Dover Music Scores) By Richard Wagner;Opera and Choral Scores READ ONLINE

Curriculum Standard One: The student will listen to and analyze music critically, using the vocabulary and language of music.

PLANE TESSELATION WITH MUSICAL-SCALE TILES AND BIDIMENSIONAL AUTOMATIC COMPOSITION

MUSIC THEORY CURRICULUM STANDARDS GRADES Students will sing, alone and with others, a varied repertoire of music.

Curriculum Standard One: The student will listen to and analyze music critically, using the vocabulary and language of music.

Building a Better Bach with Markov Chains

A Tabu Search Algorithm to Generate Piano Fingerings for Polyphonic Sheet Music

Generating structured music for bagana using quality metrics based on Markov models

A PERPLEXITY BASED COVER SONG MATCHING SYSTEM FOR SHORT LENGTH QUERIES

Texas State Solo & Ensemble Contest. May 25 & May 27, Theory Test Cover Sheet

Melodic Pattern Segmentation of Polyphonic Music as a Set Partitioning Problem

A multi-modal platform for semantic music analysis: visualizing audio- and score-based tension

Composing first species counterpoint with a variable neighbourhood search algorithm

CHOIR Grade 6. Benchmark 4: Students sing music written in two and three parts.

PKUES Grade 10 Music Pre-IB Curriculum Outline. (adapted from IB Music SL)

A Model of Musical Motifs

A Model of Musical Motifs

An Integrated Music Chromaticism Model

Sampling the extrema from statistical models of music with variable neighbourhood search

Non-chord Tone Identification

Week 14 Music Understanding and Classification

Unit 8 Practice Test

GRADUATE PLACEMENT EXAMINATIONS MUSIC THEORY

Minimax Disappointment Video Broadcasting

ORB COMPOSER Documentation 1.0.0

Curriculum Standard One: The student will listen to and analyze music critically, using the vocabulary and language of music.

Course Schedule 1 DATE TOPICS AND READING ASSIGNMENTS THEORY ASSIGNMENTS DUE

A PROBABILISTIC TOPIC MODEL FOR UNSUPERVISED LEARNING OF MUSICAL KEY-PROFILES

EE391 Special Report (Spring 2005) Automatic Chord Recognition Using A Summary Autocorrelation Function

1 Overview. 1.1 Nominal Project Requirements

Tristan Und Isolde In Full Score (Dover Music Scores) By Richard Wagner;Opera and Choral Scores

Automated extraction of motivic patterns and application to the analysis of Debussy s Syrinx

Computer Generation and Classification of Music through Operations Research Methods

Curriculum Standard One: The student will listen to and analyze music critically, using vocabulary and language of music.

Lesson Week: August 17-19, 2016 Grade Level: 11 th & 12 th Subject: Advanced Placement Music Theory Prepared by: Aaron Williams Overview & Purpose:

fpa 147 Week 3 Up until the post war (WWII) period, electronic music tends to emerge as the technology to enable it does.

CHAPTER ONE TWO-PART COUNTERPOINT IN FIRST SPECIES (1:1)

MMS 8th Grade General Music Curriculum

Music 231 Motive Development Techniques, part 1

Curriculum Development In the Fairfield Public Schools FAIRFIELD PUBLIC SCHOOLS FAIRFIELD, CONNECTICUT MUSIC THEORY I

Take a Break, Bach! Let Machine Learning Harmonize That Chorale For You. Chris Lewis Stanford University

Analysis Worksheet Fauré Elegy

Week 5 Music Generation and Algorithmic Composition

Melody Extraction from Generic Audio Clips Thaminda Edirisooriya, Hansohl Kim, Connie Zeng

TEST SUMMARY AND FRAMEWORK TEST SUMMARY

Scheme of work: 2 years (AS and A-level)

AP Music Theory 2015 Free-Response Questions

ANALYSIS BY COMPRESSION: AUTOMATIC GENERATION OF COMPACT GEOMETRIC ENCODINGS OF MUSICAL OBJECTS

Homework 2 Key-finding algorithm

BAND Grade 7. NOTE: Throughout this document, learning target types are identified as knowledge ( K ), reasoning ( R ), skill ( S ), or product ( P ).

COURSE OUTLINE. Corequisites: None

A Novel Approach to Automatic Music Composing: Using Genetic Algorithm

ORCHESTRA Grade 5 Course Overview:

YEAR 5 AUTUMN 1. Working with pentatonic scales

Murrieta Valley Unified School District High School Course Outline February 2006

Notes on David Temperley s What s Key for Key? The Krumhansl-Schmuckler Key-Finding Algorithm Reconsidered By Carley Tanoue

Algorithms for melody search and transcription. Antti Laaksonen

AP/MUSIC THEORY Syllabus

Labelling. Friday 18th May. Goldsmiths, University of London. Bayesian Model Selection for Harmonic. Labelling. Christophe Rhodes.

Characteristics of Polyphonic Music Style and Markov Model of Pitch-Class Intervals

Course Overview. Assessments What are the essential elements and. aptitude and aural acuity? meaning and expression in music?

AP MUSIC THEORY 2015 SCORING GUIDELINES

Chords not required: Incorporating horizontal and vertical aspects independently in a computer improvisation algorithm

Scheme of work: Co-teaching AS and A- level (1 year)

Acoustic and musical foundations of the speech/song illusion

DETECTION OF KEY CHANGE IN CLASSICAL PIANO MUSIC

USING HARMONIC AND MELODIC ANALYSES TO AUTOMATE THE INITIAL STAGES OF SCHENKERIAN ANALYSIS

Chord Recognition. Aspects of Music. Musical Chords. Harmony: The Basis of Music. Musical Chords. Musical Chords. Music Processing.

Rise of the Drumpreneur Unleashing Creativity Through Drum Circle Facilitation

Texas State Solo & Ensemble Contest. May 26 & May 28, Theory Test Cover Sheet

School of Church Music Southwestern Baptist Theological Seminary

Grade Level 5-12 Subject Area: Vocal and Instrumental Music

Teaching Resource COMPOSITION. 7 Activities for Teaching Composition

Leaving Certificate 2017: Music Marking Scheme Composing Higher level - Core. SECTION A MELODY COMPOSITION (40 marks) Q Descriptors Mark

1a.51 Harmonic Seconds and Fifths WB2 1A_51ABCDEFGHIJ.WAV 1a.52 Identifying and Notating Seconds and All WB2 1A_52ABCDEFGHIJ.WAV

Course Outcome Summary

Contents INTRODUCTION CONTEXT PART ONE

SAN FRANCISCO CONSERVATORY OF MUSIC

NUMBER OF TIMES COURSE MAY BE TAKEN FOR CREDIT: One.

UNDERGRADUATE MUSIC THEORY COURSES INDIANA UNIVERSITY JACOBS SCHOOL OF MUSIC

MUSIC CURRICULM MAP: KEY STAGE THREE:

Transcription:

MorpheuS: constraining structure in automatic music generation Dorien Herremans & Elaine Chew Center for Digital Music (C4DM) Queen Mary University, London Dagstuhl Seminar, Stimulus talk, 29 February 4 March 2016 D. Herremans (C4DM, QMUL, London) Structure & generation 1 / 24

Automatically generating music? Why don t we listen to automatically generated music? D. Herremans (C4DM, QMUL, London) Structure & generation 2 / 24

Automatically generating music? Why don t we listen to automatically generated music? Long-term structure D. Herremans (C4DM, QMUL, London) Structure & generation 3 / 24

Automatically generating music? Composing music = combinatorial optimization problem Decide on which notes Objective: fit a style/structure Solved by metaheuristic such as variable neighbourhood search D. Herremans (C4DM, QMUL, London) Structure & generation 4 / 24

How do we evaluate music? Human evaluation Music theory Machine learning objective function D. Herremans (C4DM, QMUL, London) Structure & generation 5 / 24

Outline 1 Structural constraints Global structure Tension profile Pattern detection D. Herremans (C4DM, QMUL, London) Structure & generation 6 / 24

Outline 1 Structural constraints Global structure Tension profile Pattern detection D. Herremans (C4DM, QMUL, London) Structure & generation 7 / 24

Global structure Bagana: Ethiopian Lyre Match an expectancy profile (Markov model) Fixed structure (A 1 A 2 A 1 A 2 A 3 A 1 A 3 A 1 ) Hard constraint during optimization 8 A1 4 3 2 2 2 4 A2 5 4 2 2 2 2 A3 4 2 3 3 1 1 5 1 1 3 2 2 4 A2 5 4 2 2 2 2 D. Herremans (C4DM, QMUL, London) Structure & generation 8 / 24

Outline 1 Structural constraints Global structure Tension profile Pattern detection D. Herremans (C4DM, QMUL, London) Structure & generation 9 / 24

Tension profile Complex composite concept Tonal, melodic, harmonic, rhythmic, expressive,... Useful for complete compositions & video/game music T R Tonal tension model based on the spiral array (Chew, 2000) D. Herremans (C4DM, QMUL, London) Structure & generation 10 / 24

Spiral array, 3D mathematical model for tonality D. Herremans (C4DM, QMUL, London) Structure & generation 11 / 24

3 aspects of tonal tension Cloud diameter Cloud momentum D. Herremans (C4DM, QMUL, London) Structure & generation 12 / 24

3 aspects of tonal tension Tensile strain (distance to key) D. Herremans (C4DM, QMUL, London) Structure & generation 13 / 24

Tonal tension tristan chord Wagner s opera Tristan und Isolde Bass note and augmented 4th, 6th and 9th D. Herremans (C4DM, QMUL, London) Structure & generation 14 / 24

Tonal tension tristan chord D. Herremans (C4DM, QMUL, London) Structure & generation 15 / 24

Tension profiles soft constraints D. Herremans (C4DM, QMUL, London) Structure & generation 16 / 24

Outline 1 Structural constraints Global structure Tension profile Pattern detection D. Herremans (C4DM, QMUL, London) Structure & generation 17 / 24

Pattern detection Compression algorithm: COSIATEC (Meredith, 2013) Point set representation of a piece Computes a compressed encoding of the piece maximal translatable patterns D. Herremans (C4DM, QMUL, London) Structure & generation 18 / 24

Polyphonic example Bach prelude 20 (book 2) D. Herremans (C4DM, QMUL, London) Structure & generation 19 / 24

Putting it all together: MorpheuS Problem: find pitches Objective: match tension profile to template Hard constraint: detected patterns Test: Bach 1st Prelude: D. Herremans (C4DM, QMUL, London) Structure & generation 20 / 24

Preliminary results: pattern detection D. Herremans (C4DM, QMUL, London) Structure & generation 21 / 24

Preliminary results: random starting piece D. Herremans (C4DM, QMUL, London) Structure & generation 22 / 24

Preliminary results: fit in tension profile & patterns D. Herremans (C4DM, QMUL, London) Structure & generation 23 / 24

MorpheuS: constraining structure in automatic music generation Dorien Herremans & Elaine Chew Center for Digital Music (C4DM) Queen Mary University, London Dagstuhl Seminar, Stimulus talk, 29 February 4 March 2016 D. Herremans (C4DM, QMUL, London) Structure & generation 24 / 24