A minimum of contextual information is provided besides the auralisation of the György Ligeti hall in order to explain the intention of our proposal to use the same fixed loudspeaker configuration for all works. The contextual information may be skipped (e.g. in order to avoid any bias) as the auralisation – which is presented separately – is supposed to be self-sufficient.
The György Ligeti hall
The György Ligeti hall is the main hall of the MUMUTH, the House of Music and Music Drama of the University of Music and Performing Arts Graz (KUG). It is a multi-purpose hall built for concerts and opera performances. It offers an internationally unique combination of modern performance infrastructure components, including height-adjustable floor elements throughout the entire hall, a variable room acoustics system, a state-of-the-art lighting system, and an audio system with 105 built-in loudspeakers of which 33 can be positioned automatically. For Mind the Gap we will be using one fixed configuration of these 33 speakers, which are mounted on motorized lifts making it possible to change the height, azimuth and tilt of each speaker separately. This particular configuration was developed and used extensively in the artistic research project The Choreography of Sound (CoS).
The Choreography of Sound
The aim of the project is to propose alternatives to established ways of conceiving of sonic space in electro-acoustic music composition. By introducing and developing the notion of sound choreography as an alternative to sound spatialisation, the project experiments with new ways of conceiving sonic objects and textures. An important methodological constraint of the project is that all inquiry is based on the particular features of the György Ligeti hall, especially with respect to its acoustics and loudspeaker infrastructure. Instead of attempting to solve the problem of composing sonic space independently of a particular performance venue (which is one of the utopian ventures of spatialisation research), the project explores the idiosyncrasies of a unique setup through a technique called on-site composition – an approach blending composition and installation practices. In on-site composition, experiencing the work at hand in the same environment in which it will eventually be performed is central to the compositional process. In this approach the design of the loudspeaker configuration itself is an integral part of composition. For further details about the project please refer to the project website.
The Loudspeaker Configuration
When conceiving a loudspeaker configuration usually one has to obey a large number of theoretical and practical constraints. Particular sound spatialisation techniques require certain speaker layouts (e.g. stereo, 5.1 surround sound, or more specialized techniques used in computer music such as Ambisonics, VBAP, DBAP or Wave Field Synthesis). At a particular venue, speakers can only be mounted at certain locations, due to mechanical, visual or security constraints. Mounting speakers and especially experimenting with different configurations is a very labour-intensive process. It can take several hours or even days to set up a larger speaker configuration. This is one of the reasons why the György Ligeti hall has been equipped with an automatised solution to speaker positioning, relaxing many of the constraints mentioned while introducing new ones.
The positions at which the loudspeaker lifts are fixed to the ceiling of the hall have been chosen to easily realize hemispherical loudspeaker configurations optimized for sound spatialisation with the Ambisonics technique. Hence, although the computer-controlled positioning system enables an experimental exploration of a completely new space of possibilities, these possibilities are constrained by a certain way of thinking about spatialisation. Part of this thinking is the idea of assembling loudspeakers into arrays which are used as displays for rendering sound sources at positions where there is no loudspeaker. The simplest array of this kind is a pair of stereo speakers. Another important aspect of this thinking is that the speakers are usually oriented towards the listener, in an attempt to minimize the influence of the room acoustics on the sound rendering.
The loudspeaker configuration chosen for the Mind the Gap experiment is the result of a process aimed at avoiding any biases we might be subject to when conceiving a configuration based on an existing set of technical constraints. A number of randomly generated configurations were analyzed theoretically and explored practically by composing with and for them. One of them turned out to be particularly useful in the sense that it combines several features, without being systematic about it. It exhibits a certain overall balance of directly (no. 3, 4, 7, 9, 15, 16, 17, 18, 21, 22, 26, 29, E1, E2, E3) and indirectly (no. 1, 2, 5, 6, 8, 10, 11, 12, 13, 14, 23, 24, 25, 27, 28, E4) projecting speakers, while providing very unusual constellations one would probably not come up with in any other way. There are also two speakers (no. 4 and 15) which happen to be positioned at ear level, i.e. they can be approached by a freely moving audience, allowing the audience to listen to a speaker at close proximity. Some speakers project directly against a wall (no. 17, 19, 20), so there are larger areas in the hall where these speakers are experienced as mirror sound sources (appear to be located behind the wall). The random process used explored the full range of possible positions (height) and orientations (azimuth and tilt). The numerical details of the configuration can be found in the data sheet. The selected configuration can be considered as an objet trouvé in the fund of the CoS project.
The Room Acoustics Model
In order to enable us to work with the on-site composition approach in the context of the CoS project at times without access to the hall (which is shared by the 17 institutes of the university), a binaural auralisation technique is used based on room impulse response measurements. These measurements are performed with a dummy head, which is a stereo microphone shaped like a human head and torso, and microphone capsules placed at the positions of the eardrums. The acoustic transformation that a loudspeaker signal is subjected to on its way from the speaker membrane to the two eardrums is measured. When the transformation measured is applied to a sound signal by means of digital signal processing (with the transformation-measurement data functioning as a particular kind of filter) and the signal is listened to via headphones, it sounds almost as if we listened to the sound projected through the loudspeaker in the hall. In this kind of model, the measured data represent the apparent location of the speaker, and the acoustic signature of the room as experienced from a fixed listening position is applied to the signal. Performing such a measurement for each of the 33 speakers of the chosen configuration allows us to form a model of the configuration representing certain aspects of the auditory experience at one listening position. Switching between models allows us to virtually move (actually jump) around in the room and listen from different positions in order to compare them.
There are several drawbacks to this technique, a main one being the impossibility to improve localisation with head movements, something we do unconsciously all the time. There are many more conceptual, phenomenological and technical issues with this technique, which constrain its applicability and which cannot be discussed in the necessary detail here. In the context of CoS, it could be shown that certain kinds of compositional work can be performed successfully with the room acoustics model, especially if the person using it has experienced the modeled configuration directly in the hall. Then the model is definitely capable of conjuring up the memory of the experience, which, together with the aspects that the model represents well, results in an experience rich enough for one to continue working with headphones after having left the hall or when preparing for a future work session there. For the Mind the Gap experiment, room acoustic models for several listening positions are made available to be used in the preparation of the works to be presented at the symposium. A program accepting 33 mono signals as input and producing a stereo signal for headphone playback as output (VirtualMUMUTH) can be used to "prelisten" to the configuration before experiencing it in the hall. The auralisation page has been produced with the VirtualMUMUTH program.