Real-Time Control of Greek Chant Synthesis

Publication Type:

Conference Paper

Source:

SMC Conference 2006 (2006)

URL:

files/proceedings/2006/6-Zannos-PhonoChantSMC06.pdf

Abstract:

This paper we report on an interdisciplinary project for modeling Greek chant with real-time vocal synthesis. Building on previous research, we employ a hybrid musical instrument: Phonodeon (Georgaki et al. 2005), consisting of a MIDI-accordeon coupled to a real-time algorithmic interaction and vocal synthesis engine. The synthesis is based on data provided by the AOIDOS program developed in the Department of the Computer science of the University of Athens, investigating Greek liturgical chant compared to bel canto singing. Phonodeon controls expressive vocal synthesis models based on formant synthesis and concatenated filtered samples. Its bellows serve as hardware control device that is physically analogous to the human breathing mechanism [georgaki, 1998a], while the buttons of the right hand can serve multiple functions. On the level of pitch structure, this paper focuses on a particular aspect of control, namely that of playing in the traditional non-tempered and flexible interval structure of Greek modes (ήχοι: echoi) while using the 12- semitone piano-type keyboard of the left hand. This enables the musical exploration of the relationship between the spectral structure of the vocal timbre of Greek chant and characteristic intervals occuring in the modal structure of the chant. To implement that, we developed techniques for superimposing interval patterns of the modes on the keyboard of the phonodeon. The work is the first comprehensive interactive model of antique, medieval and modern near-eastern tunings. The techniques developed can be combined with techniques for other control aspects, such as timbre and vocal expression control, phoneme or (expressive/ornamental/melodic pattern, inflection) sequence recall and combination, data record on/off, or others, which form part of the phonodeon project. On the level of timbre and expression, we make use of data obtained by analysis from audio samples of chanting as control sources for synthesis by concatenation of control data, thereby providing an example of realtime application of Diphone techniques (Rodet and Levevre). This research can find applications in many computer music fields such as algorithmically controlled improvisation, microtonal music, music theory and notation of (algorithmic/computerized) real-time performance, and computer modeling of experimental or non-western musical styles.