MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle Advisors: Mary Simoni William P. Birmingham
Dec 14, 2015
MusEn: Atonal Analysis Project
CARAT/Rackham Fellows Meeting10/17/01
Erica SchattleAdvisors: Mary SimoniWilliam P. Birmingham
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Presentation Overview
I. Project Overview
II. Introduction to Atonal Music
III. Introduction to Atonal Analysis
IV. Project Specifics
V. Project Status and Semester Goals
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Presentation Overview
I. Project Overview
II. Introduction to Atonal Music
III. Introduction to Atonal Analysis
IV. Project Specifics
V. Project Status and Semester Goals
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
What is MusEn
• Interdisciplinary research group that combines UofM School of Music and College of Engineering
• Aim is to use computers to both assist and independently execute many musical tasks
• Current projects include classification and retrieval by musical style
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Atonal Analysis Project Goals
• Automate processes that aid analysis of atonal music while preserving salient aspects of hand analysis
• Develop more sophisticated forms of music information retrieval
• Provide a framework for collaborative music analysis
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Presentation Overview
I. Project overview
II. Introduction to Atonal Music
III. Introduction to Atonal Analysis
IV. Project Specifics
V. Project Status and Semester Goals
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
What is atonal music?
• A genre of Western “classical” art music that developed in early 20th-century Europe
• New musical grammars, new sounds, new musical structures
• “Emancipation of the dissonance”
Webern String Quartet Op. 28, III (1936-8)
Wolpe Symphony No. 1, I (1955-6)
Beethoven String Quartet op. 18 no. 5 III (1798-1800)
Brahms Symphony No. 4, I (1884-5)
Tonal: Atonal:
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Presentation Overview
I. Project Overview
II. Introduction to Atonal Music
III. Introduction to Atonal Analysis
IV. Project Specifics
V. Project Status and Semester Goals
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Hand analysis of atonal music
• Primary methods to describe structures developed out of mathematical set theory
• Groups of notes are often related by transposition, inversion, complement, or complex relations
• In addition to notes, other elements of the music can make important contributions to the structure, e.g. dynamics, beaming
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Analysis as pattern searching
• Analysis can be seen as a series of queries sparked and reinforced by listening– For example: Find all chords that are the same
set class as the opening chord
• Segmentation—dividing a work into structural components—can be accomplished by diverse methods
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Segmentation Example
A few ways to segment:• rest delimited vertical sonorities• < > dynamics within one part• vertical sonorities with same duration• contiguous instances of fdyn.• notes grouped by slurs or phrase markings• groups of three contiguous notes with same rhythmic pattern
And a few concerns:• dynamics subject to interpretation• hard to find all valid query results, even in 6 measures • at this micro level, many more segmentation possibilities— scaling to 112 meas. is hard!• our segmentation criteria did not include segments of aural significant
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Segementation…. then what?
• Could determine set-class for each segment• Could determine interval content/generate interval
vector for each segment• Could compare segments aurally • Could transform segments by transposition,
inversion, retrograde, or retrograde inversion For all of these options, an analytic framework is
necessary to determine what information is of significance!
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Challenges of Hand Analysis
• Immediate hearing of specific parts at specific measures requires piano skills
• Playback is not heard on appropriate instrument• Only results considered significant would be
recorded on the score or elsewhere, often without context of why they seemed important
• Analytic framework or intent may change over time and necessitate repeating “lost” searches
• Many possible sources of error
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
How will this project help?
• Automated searching• Automated assignment of pc-set, interval vectors,
etc. for given segments• Automated MIDI generation/transposition etc.• Automatic generation of search history and option
to annotate search results• Results presented visually (on score) and aurally
(MIDI)
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Presentation Overview
I. Project Overview
II. Introduction to Atonal Music
III. Introduction to Atonal Analysis
IV. Project Specifics
V. Project Status and Semester Goals
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
System Architecture
• Built upon two music software programs for analysis and display
• Humdrum – for encoding, manipulating, and analyzing music
• Mup – for music publication
• Basic Humdrum to Mup translator developed by Humdrum research group
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
System Architecture II
Generate Query Interpret Results Browse History
Need to: • view results in context of score• listen to MIDI• annotate with text or drawings• easily reformat query
Browse by:• query title• query author• piece title/ composer/ mvmt.• measure numbers• instruments• date and time• MIDI content• comments/ annotations
Provide for:• well-formed queries• different levels of specificity
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Example of kern Encoding
**dyn **kern **dyn **kern **dyn **kern **dyn **kern*staff1 *staff1 *staff2 *staff2 *staff3 *staff3 *staff4 *staff4*clefG2 *clefG2 *clefG2 *clefG2 *clefC3 *clefC3 *clefF4 *clefF4*M2/8 *M2/8 *M2/8 *M2/8 *M2/8 *M2/8 *M2/8 *M2/8=1 =1 =1 =1 =1 =1 =1 =1. 8r . 4r . 4r . 4rp<[ 8d-~ . . . . . .=2 =2 =2 =2 =2 =2 =2 =2. 16r . 4r . 8r p< 4B-~. (16ccnL . . . . [ .>] 8ee-J) . . p<[ 8gn~ >] .*M3/8 *M3/8 *M3/8 *M3/8 *M3/8 *M3/8 *M3/8 *M3/8* * * * * * * *clefC4=3 =3 =3 =3 =3 =3 =3 =3. 8r p< 4en~ . 16r . 8r. . [ . . (16F#L . .p 8ddn~ > . > 8AnJ) > (8bnL. 8r ) 8r ) 8r ] 16g#J) . ] . ] . . 16r
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Example of Mup Encoding
music
1 1: 8r; [with "\(leg)"] 8d&4;2 1: mr;3 1: mr;4 1: mr;boldital (12) below 1: 1.75 "p";< below 1: 1.9 til 2.5;
bar reh "2"1 1: 16r; 16cn5 bm; 8e&5 ebm;2 1: mr;3 1: 8r; [with "\(leg)"] 8gn4;4 1: [with "\(leg)"] 4b&3;phrase 1: 1.5 til 2;boldital (12) below 4: 1 "p";> below 1: 2 til 3;< below 3: 2.3 til 2.8;< below 4: 1.2 til 1.8;> below 4: 1.9 til 2.5;bar reh "3"
boldital (12) below 3: 2 "p";score time=3/8 beamstyle=8,8,8
music1 1: 8r; [with "\(leg)"] 8dn5; 8r;2 1: [with "\(leg)"] 4en4; 8r;3 1: 16r; 16f#3 bm; 8an3 ebm; 8r;4 1: 8r; 8bn4 bm; 16g#4 ebm; 16r;phrase 3: 1.5 til 2;phrase 4: 2 til 3;boldital (12) below 1: 2 "p";boldital (12) below 2: 1 "p";< below 2: 1 til 1.5;> below 2: 2 til 3;> below 3: 2.3 til 3;> below 4: 2 til 3;
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
A Sample Query
What are the intervals (non-compound) between the violin I and violin II parts in measure 5?
P4 P5 m6 M6
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Query: Interface Prototype Format
What are the intervals (non-compound) between the violin I and violin II parts in measure 5?
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Query: Command Line View
What are the intervals (non-compound) between the violin I and violin II parts in measure 5?
extract –f 2,4 283d.krn > 283d_ex.krn
ditto –s ^= 283d_ex.krn > 283d_ex.ditto
hint –cu –s r 283d_ex.ditto > 283d_ex.hnt
paste 283d_ex.hnt 283d.krn > 283d_hnt.krn
yank –n ^= -r 5 283d_hnt.krn > 283d_hnt5.krn
ms 283d_hnt5.krn > 283d_hnt5.ps
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Presentation Overview
I. Project Overview
II. Introduction to Atonal Music
III. Introduction to Atonal Analysis
IV. Project Specifics
V. Project Status and Semester Goals
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Victory and challenge
• Victory: Finding an extensible ASCII representation that captures significant aspects of the score and generates MIDI output
• Challenge: Hand encoding of scores is labor intensive—need an automated process that converts scores (sheet music) to Humdrum representation
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Victory and Challenge II
• Victory: System preserves salient aspects of hand analysis and presents them in a multimedia setting
• Challenge: Make the system even better! – Direct manipulation of the score– Embed levels of intelligence into the system:
“Find me things that sound like … and why”– Improve MIDI options
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Status
• Working on a web prototype that replaces command line interface with graphic interface
• Developing a protocol for collaborative analysis project using Worktools
• Modifying encoding scheme for humdrum dynamics and analysis to improve display
MusEn: Atonal Analysis Project CARAT/Rackham Fellows Meeting 10/17/01 Erica Schattle
Semester Goals:
• Develop and refine web interface
• Conduct experimental group analyses
• Evaluate usability and functionality
• Explore issues raised by collaborative analysis
• Explore compositional and educational applications
Any Questions?