Acoustic and Perceptual Comparison of Speech and Drum Sounds in the North Indian Tabla Tradition: An Empirical Study of Sound Symbolism Aniruddh D. Patel and John R. Iversen The Neurosciences Institute, 10640 John Jay Hopkins Dr., San Diego, CA 92121, USA E-mail: apatel@nsi.edu, iversen@nsi.edu ABSTRACT North Indian tabla drumming is an oral tradition which uses a system of nonsense syllables (vocables) to name drum sounds. We tested the hypothesis that the vocables are a case of sound symbolism (onomatopoeia). Eight vocables and their corresponding drum sounds were collected from 6 professional tabla players in India. Analysis revealed that acoustic properties of drum sounds were reflected by a variety of phonetic components of vocables: spectral centroid, rate of amplitude envelope decay, duration between the releases of consonants in a cluster, fundamental frequency, and the influence of aspiration on the balance of low vs. high frequency energy in a vowel. We also describe a perceptual experiment which demonstrates that naïve listeners can match vocables to their corresponding drum sounds. Taken together, these results provide strong evidence of sound symbolism in the North Indian tabla drumming tradition. 1. INTRODUCTION 1.1 GENERAL BACKGROUND Numerous cultures use nonsense syllables or ‘vocables’ in systematic ways to represent musical sounds [1,2,3,4]. A familiar Western example is solfège, in which the notes of the scale are represented by the syllables do, re, mi, etc. Two contrasting hypotheses can be proposed concerning vocables. The first is that their relationship to musical sounds is arbitrary, as appears to be the case in Western solfège [1]. The alternative hypothesis is that they acoustically resemble the musical sounds they represent, i.e. they are based on sound symbolism or onomatopoeia [5]. Recent empirical research has found evidence of sound symbolism in the vocables used to represent percussion instruments in Peking opera [3] and pitches in Japanese Noh flute music [1]. Thus the two hypotheses are not mutually exclusive; rather, the question is one of relative frequency. Another relevant question is how sound symbolism manifests itself in different cultures. Since phoneme inventories differ across languages, what similarities and differences emerge in the way people use the sounds of their language to represent musical sounds? 1.2 NORTH INDIAN TABLA DRUMMING The tabla is a percussion instrument used in North Indian music, consisting of two sealed membranophones with animal-skin heads: the smaller, wooden-shell ‘dayan’ (played with the right hand and tuned to a higher pitch) and the larger, metal-shell ‘bayan’ (played with the left hand). Teaching and composition on this instrument is based on oral tradition, and employs a linguistic syllabary in which particular drum strokes are named by particular Hindi syllables, known as ‘bols’ (for example, ‘Tin’, ‘Tun’, ‘Ta’, ‘Dha’, cf. Table 1 for IPA notation). The size of the bol inventory is between 10-20 items [2,6]. Past research has shown that a vocable may be pronounced with a different vowel or consonant depending on context, without altering the drum stroke [6]. Furthermore, bols and drum strokes vary somewhat between different schools of tabla (gharanas). Thus the mapping between syllables and drum sounds is not one-to-one. However, there are certain basic bols which are widely used and which we focus on in this study. To our knowledge, there are no previous empirical studies comparing tabla sounds to the spoken syllables used to signify them. The goal of the current study is to determine if the mapping between drum sounds and speech sounds in tabla music has an acoustic and perceptual basis, i.e. if it is an example of sound symbolism. 2. METHODS 2.1 DETAILS OF SPEECH AND DRUM SOUNDS Vocables and drum sounds were recorded by six professional tabla players (one female) in a sound recording studio in Bombay, India in April of 2002. After being asked to play all the basic strokes, a list of vocables (n = 11) was determined by the players in conference with each other. Each player produced 4-5 repetitions of each vocable and its corresponding drum sound in alternation, pausing long enough to avoid any overlap in sound. All players used the same set of drums; occasionally a player would re-tune the drum before recording. Sounds were sampled at 44,100 Hz and recorded directly to computer in .wav format. A digital video recording was also made. From this corpus, eight common vocables were chosen for this study, and are presented in pairs in order to focus on specific acoustic 15th ICPhS Barcelona 925 ISBN 1-876346-48-5 © 2003 UAB