SLIDE 1
From Sounds to Music: Learning the Bohlen-Pierce Scale Psyche Loui - - PowerPoint PPT Presentation
From Sounds to Music: Learning the Bohlen-Pierce Scale Psyche Loui - - PowerPoint PPT Presentation
From Sounds to Music: Learning the Bohlen-Pierce Scale Psyche Loui Beth Israel Deaconess Medical Center Harvard Medical School Bohlen-Pierce Scale Symposium March 7, 2010 The world knows and loves music Whence musical knowledge?
SLIDE 2
SLIDE 3
Whence musical knowledge?
Perspectives:
Developmental studies Cross-cultural studies Artificial system
Bohlen-Pierce scale
SLIDE 4
The tritave as a musical system
200 300 400 500 600 700 1 2 3 4 5 6 7 8 9 10 11 12 13
increments (n) frequency (Hz)
F = 220 * 3 n/13
Bohlen-Pierce
3 : 5 : 7
SLIDE 5
Composing in the Bohlen-Pierce scale
10 7 10 10 6 4 7 6 3 F = 220 * 3 n/13
Krumhansl, 1987 Loui & Wessel, 2008
SLIDE 6
Composing melody from harmony – applying a finite-state grammar
10 7 10 10 6 4 7 6 3
Loui & Wessel, 2008
SLIDE 7
Melody: 10 10 4 7 6 10 10 7 10 10 6 4 7 6 3
Loui & Wessel, 2008
Composing melody from harmony – applying a finite-state grammar
SLIDE 8
Can we learn the B-P scale?
General design of behavioral studies:
1.
PRE-TEST
assess baseline
2.
EXPOSURE to melodies in one grammar
~30 minutes
3.
POST-TESTS
assess learning
SLIDE 9
Learning a musical system: basic questions
Can we recognize old melodies?
2-AFC test of recognition
Can we generalize to new melodies?
2-AFC test of generalization
Can we learn to like new melodies?
Preference ratings
SLIDE 10
Double dissociation between grammar learning and preference change
- No. of melodies
1 27 40 100
- No. of repetitions
5 10 15 400 40% 50% 60% 70% 80% 90% 100% Percent Correct 0.2 0.4 0.6 0.8 1 1.2 Difference in rating (familiar - unfamiliar) recognition generalization preference change
Loui & Wessel, 2008 Loui, Wessel & Hudson Kam, in press.
SLIDE 11
Learning a new musical system: more questions
Can we learn to expect frequent tones? Probe tone ratings test
Probe tone profiles reflect frequencies of
compositions
Krumhansl, 1990
SLIDE 12
Testing for expectation for frequencies
Probe tone ratings test (Krumhansl, 1990)
- Melody tone
- Task: rate how well the tone fits the melody
- Scale of 1 through 7
- Tests conducted both pre- and post-
exposure
SLIDE 13
Pre-exposure probe tone ratings
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9 10 11 12 Probe tone Rating 200 400 600 800 1000 1200 Rating Exposure Frequency of exposure
F = 220* 3n/13
Loui, Wessel & Hudson Kam, in press.
SLIDE 14
Post-exposure probe tone ratings
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9 10 11 12 Probe tone Rating 200 400 600 800 1000 1200 Rating Exposure Frequency of exposure Loui, Wessel & Hudson Kam, in press.
SLIDE 15
Correlating ratings with exposure
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Pre Correlation (r) Post Exposure
Loui, Wessel & Hudson Kam, in press.
SLIDE 16
Sounds give rise to implicit learning of music
Music Sounds
Can we observe implicit learning in real time? with Event-Related Potentials:
Yes
SLIDE 17
Event-Related Potentials can measure brain activity – Western music
Loui et al, 2005
5mV +
- 1000ms
Fza Deviant Standard Early Anterior Negativity
150-200ms
Late Negativity
500-550ms
SLIDE 18
Event-Related Potentials can measure brain activity in the Bohlen Pierce scale
Experiment design:
Chord progressions:
Standard 70%
Deviant 20%
Fadeout 10%
Amplitude change detection task
Attending to auditory stimuli but not to harmony
Dissociating perception from decision-making
Loui, Wu, Wessel, & Knight, 2009.
SLIDE 19
ERP responses to Bohlen-Pierce scale
2
- 2
[µV] 500 [ms]
AFz
Standard Deviant
ERPs for improbable chords in B-P scale elicit EAN and LN.
500 – 550ms 150 – 210ms
Early Anterior Negativity Late Negativity
Loui, Wu, Wessel, & Knight, 2009.
SLIDE 20
Effects driven by probability?
2
- 2
[µV] 500
AFz
[ms]
Equal probability
Probability of exposure vs. surface features of stimulus
Loui, Wu, Wessel, & Knight, 2009.
SLIDE 21
2
- 2
[µV] 500 [ms]
Fz
2
- 2
[µV] 500 [ms] 2
- 2
[µV] 500 [ms]
Learning probability during exposure
Fz Fz
Standard Early Standard Late Deviant Early Deviant Late
Early Late
- 5
- 4
- 3
- 2
- 1
Early Late amplitude (mV)
EAN peak amplitude
Loui, Wu, Wessel, & Knight, 2009.
SLIDE 22
ERP amplitude reflects individual differences
R = 0.75
- 1
- 0.5
0.5 1 1.5 2 2.5 3 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Generalization (proportion correct) EAN amplitude (mV)
Loui, Wu, Wessel, & Knight, 2009.
SLIDE 23
Music Sounds
+
- Statistics of sounds give rise to musical knowledge
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9 10 11 12
Statistics and acoustics constrain music by limiting what we can learn.
SLIDE 24
Sound spectrum constrains knowledge in music
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9 10 11 12 Probe tone Rating 200 400 600 800 1000 1200 Frequency of exposure
SLIDE 25
Sound spectrum constrains knowledge in music
SLIDE 26
Sound spectrum constrains knowledge in speech and language?
Wordle.net
SLIDE 27
Acknowledgements
David Wessel
Erv Hafter
Carla Hudson Kam
Bob Knight
Marty Woldorff (Duke)
Carol Krumhansl (Cornell)
Center for New Music & Audio Technologies
Auditory Perception Lab
Language & Learning Lab
Knight Lab
UC Berkeley Psychology