1. The Communication of Musical Expression Roger A. Kendall & Edward C. Carterette University of California, Los Angeles Reviewed by Erdem Unal

2. Introduction Result: both musicians and non-musicians can discern expressive levels of instruments Experimental methods:  Categorization, matching and rating  Acoustical analyses of timing and amplitude Focuses on Performer-Musical Chain of musical communication…  BASIC QUESTION: “How does the performer convey his ideas to the listener?”

3. Background Grouping and parsing elementary thought units are basic mental capability of humans Music is conceived as basing on a system of sequences of states such as sound states, acoustic states, symbol states, mental states… At the transitions between these states, human can mentally organize a pattern.

4. Experimental Rationale Focus: To what extent are the performer’s expressive intentions conveyed to the listener. categorization: assign an instance to a model matching: compels the listener to remember rating: listener assigns values

5. Experiment Data: 5 instruments: piano, clarinet, oboe, violin and trumpet. Expert pianist performed selected pieces in three expressive levels:  No expression (mechanic)  Appropriate  Exaggerated (too much) Expert players ( clarinet, oboe, violin, trumpet) listened to these recorded performances (each expression only once) and asked to perform the intended expression level once each…

6. Experiment I (Categorization under piano performances) Results: (9 musicians) ClarinetOboe ViolinTrumpet Exp Lvl P1 P2 P3 P1 P2 P3 P1 P2 P3 P1 P2 P3 E1 8 1 0 8 1 0 9 0 0 8 1 0 E2 1 6 2 1 6 2 0 2 7 1 6 2 E3 0 1 8 0 3 6 0 7 2 0 2 7 Hit: 81 74 48 78 Results: (10 non-musicians) ClarinetOboe ViolinTrumpet Exp Lvl P1 P2 P3 P1 P2 P3 P1 P2 P3 P1 P2 P3 E1 6 3 1 6 2 2 8 1 1 8 0 2 E2 1 7 2 1 6 3 1 2 7 2 6 2 E3 2 2 6 2 2 6 1 6 3 2 3 5 Hit 63 6043 70

7. Experiment II (Categorization under other instruments) Results: (10 musicians –two sets) ClarinetOboe ViolinTrumpet Exp Lvl P1 P2 P3 P1 P2 P3 P1 P2 P3 P1 P2 P3 E1 11 7 2 10 6 4 14 2 4 14 2 4 E2 4 14 2 6 13 1 3 10 7 7 6 7 E3 2 9 9 3 5 12 1 10 9 5 2 13 Hit: 57 58 55 55 Discussion: Results are sufficient (over 33%)  First experiment gave better results

8. Experiment III (Time and Amplitude Deviation) Expressive levels are analyzed with acoustical signal processing software. Main focus at this point is timing and RMS of amp.: Time and amplitude derivations of the piano onsets and the four instruments’ expressive performances are observed. Artificial performance: synthesis of notes with the experimental values of amplitude and time

9. Exp. III cont… Correlations of Time Deviations Piano1 Oboe1 Trumpet Violin1Clarinet1 Piano11.000 Oboe1.658 1.000 Trumpet1.167.443 1.000 Violin1.038.374.222 1.000 Clarinet1.007.136.371.441 1.000

10. Exp III cont… Results: (12 non-musicians) ClarinetOboe ViolinTrumpet Exp Lvl P1 P2 P3 P1 P2 P3 P1 P2 P3 P1 P2 P3 E1 7 3 2 7 0 5 5 5 2 8 2 2 E2 4 8 0 2 8 2 2 2 8 2 6 4 E3 1 1 10 3 4 5 5 5 2 2 4 6 Hit: 69 56 33 56 Same Problem (!!!) at E2-E3  Violin Conclusion: Timing and amplitude are sufficient for categorization…

11. Experiment IV (Feature added: Rating) 10 musicians asked to rate a set of melodies that includes random selections of artificial and natural performances between 1 and 100. The mean ratings are observed:  Expressive levels are clearly shown by ratings  As a result, natural performances are more expressive…

12. Conclusion: Detailed statistical data proves: Both musicians and non-musicians can discern expressive intent along natural and artificial performances.