Constructing terminology for musical timbre: A semantic mapping of acoustic features in Turkish language

Abstract

Timbre is a multidimensional sensory attribute of sounds, directly linked to the perception of the identity of sound sources. Recognition and categorization of sound sources and events occur largely by detection of spectro-temporal acoustic cues attributed to timbre. Furthermore, quality judgements of musical tones are largely made on the basis of timbre. Thus, exploring its perceptual representations to better exploit its allowances is important for communication-based practices in music and sound fields. Timbre perception has been studied through perceptual multi-dimensional spaces (MDS) which demonstrate dissimilarity-based relationships between sounds. Furthermore, various semantic tasks such as semantic differential (SD) and verbal attribute magnitude estimation (VAME) aimed to find perceptual relationships between timbres through their verbal descriptions. Because verbal labels reflect conceptual projections of sounds, exploring linguistic descriptions of sensory qualities brings about its various advantages to sound practice, especially in composition, instrumentation, sound art and design areas. However, due to the multi-dimensional and perceptually subjective quality of the concept, arriving at definitive measures is notoriously problematic. In this study, perception of timbral qualities of musical instruments were investigated through their verbal descriptions. Because there has not been a previous study concerning timbre semantics in the Turkish language, the test was designed in the form of a checklist procedure for observing preference data of descriptions in an unbiased manner. The preparation step of the experiment consisted of collecting adjectives from previous semantic studies of timbre and providing them with Turkish translations. 84 adjectives were chosen from a larger pool by academics/professionals in the sound engineering and design field. Test participants consisting mostly of music students, chose adjectives from a list of 83 adjectives as they listened to 24 instrument samples consisting mostly of orchestral instruments in equalized pitch and loudness. From the analysis, frequency data gave 27 adjectives which showed correlations most strongly with the acoustic parameters of pitch strength, harmonic energy distribution, signal bandwidth, noisiness, attack centroid and spectral flux.