Gamelan pelog shows a statistically significant bias toward the interval steps of a 9-tone equal temperament (9ET)
Investigated question: Can the bias of the pelog scale toward 9ET, which seems obvious in the data, be a result of chance distribution (null hypothesis)?
Method: Considering the 22 interval types, the largest difference between a mean interval and a 9ET interval is 36 Cent (see diagram for the interval nem-barang: mean interval = 169 Cent, 9ET interval = 133 Cent). This maximum-difference range (0-36 Cent) was divided into a close-range quarter (0-8 Cent) and a distant-range three-quarter (9-36 Cent).
Expected chance distribution of the 22 interval types on the two ranges: 5.4 / 16.6
Measured distribution of the 22 interval types on the two ranges: 10 / 12
The deviation of the measured from the expected data was tested by a chi-square goodness-of-fit test. The result was P = 0.02.
Conclusion: We can say with a certainty of 98 % that the pelog tunings of Central Java use intervals that are based on the 9ET scale.
Back to: gamelan page