Academic journal article Yearbook for Traditional Music

Computer Sound Analysis of Traditional Music of Transcaucasia and Central Asia

Academic journal article Yearbook for Traditional Music

Computer Sound Analysis of Traditional Music of Transcaucasia and Central Asia

Article excerpt

(ProQuest: ... denotes non-US-ASCII text omitted.)

This article deals with two examples of a large number of computer-assisted investigations of Central Asian and Caucasian traditional music conducted over a twentyyear period at the Moscow P. I. Tchaikovsky Conservatory. The main objectives of these investigations have included: (1) analysis of sound spectrum structure and timbre; (2) melody extraction and the subsequent analysis of its pitches and intervals; and (3) analysis of performing style. Today, the study of performance style is common in both modem musicology and ethnomusicology. In oral traditions, performance style and individual interpretation are of primary importance because they form the final product. Usually, descriptions of performance of traditional music contain some colourful but inexact terms that try to portray the style of performance, the timbral palette, and the pitches and intervals. A properly designed computer programme, on the other hand, can provide accurate and exact measurements and indications of basic features of the sound.

For our analyses we used computer software called SPAX, which Alexander Kharuto, head of the Department of Musical Informatics of the Moscow P. I. Tchaikovsky Conservatory, developed between 1993 and 2005. This work culminated in the creation of SPAX for Windows (Kharuto 2005), with updates every year until the present. Using this software (and its earlier DOS versions), many studies of different traditional singing techniques and also of instruments and instrumental music have been conducted, including analyses of Tuvan throat singing, Peking opera voices, Kalmyk and Korean traditional musical instruments, Azerbaijani classical music, and Tajik, Uzbek, and Kazakh instrumental music.1 The computer analysis presented in this article involves the study of professional music in the oral tradition collected and recorded for the last three decades by Violetta Yunusova in Azerbaijan and Tajikistan. We also used more recent samples of music collected by Saule Utegalieva. Utegalieva trained with us on the use of the SPAX programme, and an article by her along similar lines is also included in this issue (see also Utegalieva 2013; Utegalieva and Kharuto 2013a, 2013b).

One of the problems that the SPAX programme helps us to solve is the extraction of the melodic line with its associated pitches and intervals. For ethnomusicological investigations, the exactness of pitch estimation is about five cents, which corresponds to the biological capabilities of human hearing. Through the use of three well-known methods of pitch extraction, the required accuracy was achieved with sufficient signal-to-noise ratio (10 dB or more) and a suitable form and size of the "analysis window" in the time domain (Kharuto 2015). The SPAX programme applies three different mathematical approaches to the problem: (1) an auto-correlation function of sound oscillations (Boersma 2001); (2) a calculation of weighted quadratic difference with further numerical minimization of the supposed period, as seen in the YIN algorithm (de Cheveigné and Kawahara 2002); or (3) the cepstrum calculation method, which proposes two Fourier transforms with logarithmic transformations of the spectrum after the first step, which approximates human perception. In the SPAX software, all these methods exist, and the researcher can choose one of them or compare results derived by different methods, which often differ in their details.

Using data from the classical music traditions of Transcaucasia (Azerbaijani mugham) and Central Asia (Tajik maqom), we studied changes in the system of modes and in musicians' hearing between different generations of performers. We found that microtones were more prevalent in the performances of older-generation traditional musicians trained in the oral tradition than in those of musicians trained in conservatories (Yunusova and Kharuto 2014, 2015). We uncovered some features of performance style by measuring two long-necked plucked lutes with tied-on frets: the eleven-string Azerbaijani tar (which has two pairs of melody strings, two pairs of high-pitched strings acting as a high bourdon, and three bass strings) and the three-stringed Tajik tanbur (with one melody string and two drone strings). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.