D3DA – by Rasmus Lunding, CC, ARTS, Aarhus University.
Disclaimer: The following is a presentation and description in the making.
For now it is first and foremost a rudimentary means to information about the D3DA concept, the thoughts behind it and reflections with potentially diverging views in relation to topics relevant to this work. It will therefore eventually also include references and links to relevant literature and materials to substantiate the idea behind the concept.
D3DA — Dirty 3D Acousmonium
D3DA is a multi-channel speaker concept based on natural behaviour of sound in any given space/surrounding. It´s inspired partially by the concept of the Acousmonium as presented by INA – GRM (Francois Bayle) in the 1970ties.
Concept and Idea
The core of the D3DA concept is based on two notions plus reflections on how to realize those in actual physical soundsystems:
-Notion 1. Reflections on how to utilize natural sound propagation in any given, natural (= untreated) space/surrounding.
-Notion 2. Reflections on how to utilize everyday sound perception in any given, natural (= untreated) space/surrounding.
-Soft- and hard-ware solutions are by definition second of nature to the D3DA concept itself. However, meaningful, potentially flexible ways of controlling the programme materiale applied to the given space/surrounding are crucial. Also the relationship between materiale (source) and speaker (exciter) has to be considered in the light of fore-mentioned notions.
This relationship is based on the idea that source and exciter act as independent yet interconnected entities at the control level of the system. Yet is perceived as a continuous element, rather than separate entities, in the actual physical setup.
(For descriptions of software solutions in the different realizations based on D3DA, please see Realizations).
The grounding principles of what eventually would become the D3DA concept was established early 2014. The Audiodesign programme at the Digital Design education at CC, ARTS, AU (http://cc.au.dk/om-instituttet/afdelinger/digital-design-og-informationsvidenskab/) was asked to help realize a system for 3D Audio Propagation (Spatialization) to be installed in a new dome structure containing the tropical climate zone at the Greenhouses in Aarhus (http://sciencemuseerne.dk/en/botanical-garden/?fb_locale=fi_FI). A system, that among other things, had to be realized on a limited budget and within a limited period of time.
The Greenhouse staff wanted an expansion of the sensory experience a visit to the Greenhouses represents: the humidity and warmth of the climate (huge contrast to the natural climate in DK), the green that fills the eye – and potentially, a surrounding sonic habitat that reflects the actual geographical zones represented in the Greenhouses, yet of a designed and potentially interactive nature.
Therefore, the concept of D3DA was actually not a concept at first, more a process of mentioned notions and reflections to help facilitate an audio system based on very real-life practical aspects of appliance.
A small (16 speaker) rudimentary prototype presenting the core elements of what later would become D3DA, termed genericAcousmonium, was presented in late spring 2014 to the Greenhouse staff. Based on this prototype the production-team at CAVI produced a full scale version handling 100+ speakers totally. This version was premiered sept. 2014.
A short video presentation of the Greenhouse setup:
Should have been “quick and dirty”, however the “quick” was dropped for obvious reasons; D3DA is not a quick way to establish a multi-channel speaker setup. In a number of ways it is indeed very unpractical – and slow. It is however “dirty” in relation to other 3D audio-concepts and -formats. D3DA in many ways represent an opposite approach to how multi-dimensional sonic representation typically is perceived. It does not try to control the environment or calculate ways of representation designed to convince the ear/the perceiver of the spatial propagation of sound through space. Instead it, as explained further on, is part of the audible space it inhabits.
Sound is always and forever in 3D, no matter the source/sources, no matter the space, no matter the perceiver, no matter the sound. The moment the transduction from source to media happens (e.i. from loudspeaker to air, from door slam to air, from whale-mouth to water) sound can not help but propagate spherically. Sound is always 3D(*).
However, the general understanding of the term “3D sound” as we know it, is when a controlled way and a control environment for steering sound in more than one direction, is at hand. Thus, the term is normally referring to systems, formats or concepts that provide these options; Dolby Surround, Ambisonics, WaveSynthesis, IRCAM SPAT etc (Gerzon et all). -And, potentially D3DA.
(*This is true for any naturally occurring sound. In terms of transmitted sound, there is one exception; the natural spheric propagation of sound waves can be transduced into a planar propagation, as used in parabolic speaker constructions).
Taken it´s origin from Jacques Poullin´s “potentiomètre d`espace”, a hardware system designed to move sound sources across 4 loudspeakers in the early 1950ties, the Acousmonium was realized by head of INA – GRM, François Bayle, and engineer Jean-Claude Lallemand around 1975. The Acousmonium consists of 50-100 loudspeakers of varying shape, size and positioning. It uses spatialization of sound as an integral part (wiki).
The Acousmonium lends it´s name from the concept of Acousmatics – understanding the sound as a connected yet separat entity from it´s source of origin. The loudspeakers representing a “veil” between the sound and it´s source (…). D3DA lends the A from the Acousmonium because it in many ways lends its format from the idea of that concept.
Except for two distinctive differences; One is that D3DA primarily is based on (so far) small uniform loudspeaker-units. The second difference refers to the understanding of the sound, similar to the Acousmonium, separated from it´s source. But in D3DA not from the “veil”.
More precisely put:
The space between the actual loudspeaker unit – the exciter of the sound, and the receiving ear can be viewed as a continuum no matter what materials and media this space consist of: loudspeakerunit-materials, cabinet-materials, air, reflecting surroundings etc. The sum of all these elements constitutes the actual vibrant sounding entity (aural architecture?). So D3DA does not conceptually distinguish between the sound, it´s exciter and the surroundings as such. This puts a number of paradigms at stake concerning the optimization of the surroundings, the technological solutions, the understanding of what the term “sound quality” refers to and ultimately; the role of the listener (see Distinctions).
Explanation (On what´s happening when experiencing D3DA)
Hearing vs Listening:
In the general everyday environment – be it indoor or outdoors, we subconsciously treat the sounds that surrounds us; other people talking, passing vehicles, sounds of nature etc. etc., in a one-to-one relation, meaning; we perceive and recognize the sounds without considering or reflecting that what we perceive as a distinct voice or other sound coming solely from a specific source, in fact mostly consists of reflections from the environment surrounding that source. On an average basis 2/3 of what we hear is the sound of the given surroundings, not sound-sources (….). We instinctively register these acoustic properties, but do not connotate them as content. In short, we don´t really listen to what we hear. That is; becomes aware of what we hear.
There is a good reason for that. As we shall see. But before continuing it might be worth it to note that there are obviously many theories, explanations, concepts and terms relating to and concerned with aural perception. Among others the concepts of the Soundscape (…), Deep Listening (…) and related aesthetics. Similarly with psychological concepts like Everyday Listening (Gaver). As well as concepts in the areas of psychoacoustics and acoustic engineering. (Stanford: Auditory Perception, …, …)
The approach that D3DA represents is however, at least somewhat, off axis to these different areas of listening and aural perception. And, it potentially merges several of them.
What is important to emphasize first and foremost is the fact – a fact that can be tested by all humans gifted with the sense of hearing;
We hear always, it is however not possible to listen always. Furthermore what we hear we connotate, in a number of ways depending on the information present in the sound;
When we hear - not listen, but hear, aesthetic and other related connotations are secondary. As mentioned, for good reason, one realizes when lending a reflecting thought to evolutionary progress (…). There is a purposefulness connected to the fact that our hearing sense in terms of both processing speed, sensitivity and spatiality in tests and experiments again and again prove superior to the in other ways know “first sense”; the visual sense (…). An evolutionary purpose based on the fact that hearing follow us from before we are born and through all the hours of each and every day and night till we die. We can shut off visual impact at will but we literally can not escape hearing unless employing external paraphernalia. Even more interesting; we rarely want to engage such (unless they provide sound; headphones). Obviously, again for evolutionary reasons; Our hearing is one of the primary elements that connects us to our given environment, to our sense of being. Hearing is so to speak “hardwired” to our instinctive self (referred to by some as “body-cognition”) (…), and can – if needed, make us react before “proper” cognitive connotation takes place. In other words, the evaluation of what is heard is of second nature, for instance if we hear a loud noise, we react first, then realize (“understand” in the connotative sense) what was heard secondly.
Listening however, is a much more profound process that take up much more cognitive processing, take more time and demands much more energy. And not the least; also demands a conscious effort. -Like, for instance; listening to music.
So, what has this to do with loudspeaker systems, one might ask.
In short, D3DA is about trying to engage and benefit from the relationship between sound and the reflecting surroundings, and how we as humans process what is heard (not what is listened to).
When considering the instinctive activity of hearing and the fact that a lot of what is heard in reality is (also) information concerning the surrounding environment: it´s size and proportions, relations between mass, density and viscosity of the materials and surfaces constituting objects in and the environment as a whole, and also distances between soundsources and other objects in the environment. We realize that when hearing, we constantly and inevitably (re)create acoustic models of any given surrounding we are in. And we constantly update those -literally at the speed of sound…. We do so to connect to the surroundings, to understand the conditions of the environment we are in. Also, we constantly monitor given variations and changes in that environment, for instinctive and primal purpose. (…)
In other words; it is actually quite remarkable how precisely the human aural perceptual system (when we are deprived from other inputs (eyes shut etc,)) (…) in a splitsecond can monitor and detect the constitution of any given environment entered: The size, it´s materials, who/what inhabits the environment, where in the environment is this who/what; near/far away etc. In fact, this inner acoustic model is allready established before we even start to consider what we, cognitively, perceive.
The paradox is that, traditionally when envisioning, designing and realizing 3D audio systems, we do what we can to deprive the setting/the space these systems are meant to inhabit any information about the space (the environment) itself for our aural perceptual system to process. Also, the technical equipment we design as “transparently” as possible. Traditionally, in audio systems in general, we do so in the quest for “sound quality”. Here understood in terms of linear frequency response and transparent, neutral representation of programme materiale (…). In relation to 3D audio systems, we do so in order to create a foundation that enable us to (potentially) create another space than the space the audio system and the listener(s) actually inhabit; a virtual acoustic 3D space. Another acoustic model.
This means two things (at least); One has to do to with the creation of this virtual acoustic 3D model. This will be discussed further on (see Distinctions).
The other has to do with the clash between the two mentioned acoustic models; the virtual 3D model and the constantly updated aural perceptual model each person carries, that inevitably are bound to happen.
D3DA are concerned with getting rid of this clash, it does so based on the simple notions presented in the beginning of this presentation, -at the sacrifice of “sound quality” in the traditional meaning of the term. Potentially it bypasses the idea of “sound quality” in the traditional meaning of the term, at least when dealing with experiencing sound in 3D audio systems. Because it feeds on the listeners own acoustic hearing models instead of trying to create a virtual acoustic 3D model;
If we can not help but instinctively address any given acoustic surrounding via instant perceptual models. Models that even include a lot of information about what constitutes the acoustic surrounding. Information about, among other things; how sound propagate through and leaves reflecting copies of itself in the surroundings, precise registration of the spatial properties of the surroundings, and of how the materials constituting the surroundings and the objects inhabiting it filters and modulates the sound propagating through the surrounding environment. – Why not then use this information instead of trying to eliminate it?
Especially because, when engaging our perceptual acoustic model, we actually do not concern ourselves with the quality of the sounds themselves! (…)
We concern ourselves with the qualities of the propagation of the sounds! Do they make sense in relation to the overall perceptual acoustic model or not?
Simply, but precisely put: Does the tree-frog up in that tree over there sound how it is supposed to sound, when heard from behind this rock and these bushes 25 meters away and with a couple of chainsaws giving sound from what sounds like a couple of kilometers away partially (dynamically as well as frequency-wise) disguising the frog´s distinctive croaking?
This perceptual report is as good an explanation as any on how we experience our acoustic environment in general terms; We are not concerned with proper frequential representation or the like, only secondarily.
So, how does D3DA address this? Quite frankly, in an almost embarrassingly simple manner, at least hardware-wise (For a more precise description of the hardware and software involved in the different realizations of the system, please see Realizations);
D3DA places small, cheap, easily manageable speaker-units in the given untreated environment, basically.
Now, how simple this may sound, there is a number of considerations to dwell on, most is dealt with in Distinctions and Realizations. In relation to the overall concept it is though of interest to point out a couple of them here:
Positioning multiple small speaker-units more or less randomly distributed in the environment gives one clear advantage; one basically position the sound where one want´s it to appear. The main advantage is however, as indirectly addressed above; the propagation of the sound happens in the given surrounding, and as a result is “coloured” by the acoustic properties of that same surrounding, making potentially any sound sound like it has a natural presence, thus a natural occurrence, in the given environment.
As experienced in the Greenhouses in Aarhus, concerns with the distribution of a proper frequential representation becomes of second nature. To a certain degree. There are limits, these are first and foremost tied to the programme materiale presented in D3DA, this will be discussed in Applications.
These limits is however not experienced in the Greenhouses, due to the close relationship between the environment and the sounds (the programme material) put in it; sounds of natural habitat.
The example with the tree-frog is applicable here. It is easily reproduced using D3DA, event sounds of birds or monkeys moving through the tree tops coming from D3DA has a natural occurrence. The birds seem present, even when you know they are not. The sounds “fit” our inner perceptual acoustic model. Listening to the same sound in one isolated small speaker-unit sitting on a desk, and listening to it in that same speaker-unit positioned in a tree, are experiences worlds apart.
Furthermore, movement and certain acoustic properties can be enhanced in the software that executes and distribute the sound material in D3DA. This will be discussed in detail in Realizations.
Distinctions (From other Systems; Acousmonium, Ambisonics, Wavefield Synthesis, SurroundSound, Andre formater (Envelop, 4D etc.))
-Three constructions so far (væksthuset, cavi (several sub-constructions), schön)
-Software (gaming(avatar)-concept mapped to sonic execution, ICST, SPAT, …..) (CAVI: Unity-based realization for installation purpose) (D3DA: MaxMSP-based realization based on compositional and realtime DSP processing perspectives).
-Hardware (mic principles used on speakers and other…..)
Applications (Programme Material)
-What programme material benefits from D3DA and what does not (“tape”, live etc.)
Future Experimentation and Appliance
-Realtime gesture control
So far to: Morten Breinbjerg, Ole Caprani, Kasper Fangel Skov, Morten Riis and ICST (*)
(*While D3DA is not based on the ambisonics (gpl) paradigm. In one of it´s MaxMSP versions it however uses a couple of ambisonics-based tools for graphic representation of sound and speakers; The ICST graphics tool “ambimonitor” and the ICST trajectory tool “ambicontrol (copyrighted by Philippe Kocher, www.icst.net)).
In progress, Rasmus Lunding firstname.lastname@example.org, nov/dec 2015