4th International Conference on Integrating GIS and Environmental Modeling (GIS/EM4):
Problems, Prospects and Research Needs. Banff, Alberta, Canada, September 2 - 8, 2000.

Virtual Reality as an analyzer of the urban "ambients" parameters

GIS/EM4 No. 233

Dominique Follut
Philippe Woloszyn

Abstract

Urban and architectural spaces generates, in interaction with environmental parameters, such as lighting or sonic, ambients with identifiable characteristics. This notion of ambients favors largely the human being position notably to breadth of the phenomenon of perception. Virtual reality, with its principles and linked techniques, seemed to be a relevant way to develop a particular instrumentation to our research. Associated with more classic methods of simulation, it allows us to integrate relevant perceptive principles to position our model in the architectural and urban ambients research fields. This methodology we present creates a new notion called "Ambianscape" and is actually developed in the Cerma laboratory.

Introduction

Since many years, architectural and urban ambients are the main subject study of the Cerma Laboratory, dealing with the urban built shapes interactions. Classical results are predictive numerical 3-D models, which provides parameters related to the ambients physical factors. The data structuration we obtain by associating 3-D geometrical elements and physical or textual datas is similar to GIS 3-D structuration, so we can claim that the Virtual Reality approach is full of promises for integrating man into the scientific knowledge of physical processes of urban environment. In addition, virtual technologies has emerged where the user can become immersed in multidimensional data sets. This allows the possibility of an ultimate visualization system, including the space (town), the process (environmental physics), and the actor (human). So, the powerful multidimensional toolbox offered with Virtual Reality (VR) can be relied with the main " classical " topics of our laboratory research works, concerning the architectural and urban environment, including thermal, aerodynamical, acoustical or lighting processes.

Notion of ambients

In this way, the architectural and urban ambients notion is related to three main interacting parameters [FOL98], which are :

• The specified built form, which shape and material characteristics are able to generate an identified ambients condition, physically speaking. Main examples can illustrate this type of interaction, as the aeraulic urban flow, the acoustic diffusion, or the albedo factor of the built surfaces, which influences the thermal and lighting reflections.
• The various specific urban physical factors, which covers a large spectrum of environmental phenomenon, such as wind, temperature, sun exposure, acoustical propagation or pollution dispersion.
• Finally, the third element of the urban environment system is the human being, identified at the same time as an modificative actor of the environment, and as an ambients perceptor.

Figure 1. Ambients scheme

 

Space and physics : the rule of scale interactions

Considering the interaction between space and ambients phenomenon, it is useful, in a first step, to segregate them into elementary physical manifestations, which involves specific parameters. Each parameter describes, during its manifestation, a particular spectrum, which is a function of each phenomenon's particular scale, called characteristic scale [WOL97b]. Each of them is referring to a specific urban space analysis model, as follows :

Figure 2. The physical ambients phenomenon and their associated scales.

The example of the sound wavelength, which corresponds to architectural and urban scales (from millimeter to meter), opposed to the light wavelength (micrometer), or to the microclimatic phenomenon (kilometer), is, at this aim, absolutely eloquent.
In addition, those spatial characteristic scales are complemented with the phenomenon temporal scales, which have to be taken into account in order to implement the physical parameters into the urban built geometry, as following :

Figure 3. Spatial and temporal scales applied to the major ambients phenomenon

The reference to various system analysis will be necessary. The most currently used methods, outer the direct phenomenon parameters measuring, implies a physical model for process evaluation. This multidimensional and plurifactorial characterization affords us to organize the urban space as a field of data [WOL98] relied to ambients physical parameters. Nevertheless, the 3-D data structuration doesn't allow an easy exploitation of those data fields. This justifies the use of Virtual Reality as a powerful and flexible 3-D multisensor visualization tool for the complex geometrical-physical data.

The place of human interaction

The use of VR for multisensor visualization of environmental physical parameters doesn't afford a resolution about ambients representation without integrating the sensitive aspect of phenomenal perception. In that way, we can consider this aspect as a sensation vector from subject (human being) to the ambients complex. Considering this subjective implementation in the virtual simulation process, we are aiming a hybrid ambients representation model, rooted to our researches in the both domains of spatial analysis and ambients perception. The final model we propose would be able to "re-feel" the ambiant parameters at each step of the perceptor's wandering [FOL99b], and, simultaneously, to preview the ambients configuration for the whole spatial wandering system.

Figure 4. Integration of the perceptive dimension in structural multi-phenomenon scheme.

This cumulative approach doesn't discriminate phenomenon, but considers perception as a single object. This allows sensation synthesis through modeling discreet scalable and measurable objects, which systemic organization constitutes the ambients entity, placed in space and time.

Those scalar physical phenomenon are perceived with specific "human sensors", which accuse sensorial disparities between ambients phenomenon. In fact, the laws of human perception, considered as a "sensorial filter" of the reality, are specific for each physical domain, relatively to their spatial and temporal scaling qualification. Otherwise, psychological perception studies points out a noticeable difference between individual perceptors, which "ambients sensitivity" is depending at the same time on the local context of perception situation and on the inner disposition of the human being.

Actually, we won't integrate this type of parameter in the modelisation process, but their importance for the perceptive process stay at our mind. A further evolution of the model will take those parameters into account, by referring largely on psychoperception studies [DUB93].

Following those constatations, we can afford to recall this complex interactions as :

Figure 5. Burdea's VR scheme.

This known Burdea's VR symbolization [BUR93] is based on the three major principles of Imagination, Immersion and Interaction.

The Immersion principle [SLA93] places human in an equivalent position of a "classical" urban pedestrian, with a height situated between 1.6 and 2 meters. The effect of this positioning is the simulation of physics and spaces interaction very near from the reality. This realistic interaction principle is conditioning the user's freedom degree in the system.

An immersion with a total interaction which permits a free moving and a dynamic parameter modification is not possible at the moment, but it will allow a real total interaction between each side of the Burdea's triangle in a near future. This freedom degree of moving and sensing is important for the space conditioning, and, consequently, for the exploration acknowledgement. At this time, the method we use predefines a exploration way for our "robot", which can also walk in the space, exploring the ensembles of simulated data .

Imagination characterizes the meaning which can be given to our VR experimentation. Our answer to this point is the proposal of description tools specific to our research topics. Those description tools are related to the physical phenomenon, with replacing them in the perceptual context. Several known methods can be applied at this aim, and the results will present a "zoning ratio" for each perceptual characterization, accessible from the stating point of the perceptor, and relating its evolution during a walktime.

Technically, our methodology consists to a "feedbak" between a 3-D geometrical model of the town and the associated physical data : after the simulation process, the geometrical model is informed with the physical data, in order to allow an interrogation of the so-constituted database with the known techniques of VR, including the human being .

Figure 6. Ambients synthetizing scheme.

This model constitutes a first answer to the double question of psychophysical representation and immersion into ambients. The ambients 3-D scheme used in immersion process relates synchronously physical dimensions to perception shapes, into a spatio-temporal dynamic. This double indexation allows to traduce the physical shape of a phenomenon into a perceptual feedback through the use of our hybrid RV model.

The psychophysical representation space usable for such an adequation suppose the construction of an homeomorphism between temporal and spatial figures of the phenomenon, and their correspondence with perceptual space-time "feedbaking".

Considering the ambients definition and VR principles we exposed, we propose a new triangular scheme, synthesizing this methodological convergence between Virtual Reality area and ambients restitution :

Figure 7. Ambients acknowledgement scheme.

- Here, the human factor, relied to the immersion item corresponds to a realistic positioning in the space, associating to a phenomenon perceptive situation.

- Psycho-physical parameters are directly linked with the imagination principle used in the virtual world through perception's specific descriptors.

- Finally, the interaction principle acts on the built environment through a wander permitting its discovering.

Application: Towards the notion of "Ambianscape"

The method presented here constitutes an exploration tool for ambients. Actually developed in the Cerma Laboratory, this tool is based on a principle of coupling between visual [HUB94] and auditive perception. It supposes therefore a lighting and acoustic ambients study.

Ecological validity

So we would solve the topic for ambients representation both with a realistic human placement in the 3-D town numerical model and with the use of pertinent perceptive descriptors.

For this purpose, we will associate the different modes of shape/phenomenon relationships we described previously, taking into account the structural organization of the two main stimuli which built the perceptual representation of ambients, vision and audition.

This leads to the question of ecological validity of the restitution conditions. This ecological validity is the guarantee of a real functioning referential illusion, with respect of the initial conditions of the reference situation (the reality) for the confection of stimuli from which we constitutes the matter of our descriptors. Those "ecological conditions" would be modeled on the background studies concerning relationships between auditive and visual perception.

The interaction effect between those two main perception domains have been described on concert hall studies, showing a noticeable influence from vision on auditive perception [NAT97]. Nevertheless, our systemic model would be able to reduce this influence by a convenient separation of visual and sonic data on the perception scheme.

At this aim, an ambients taxonomy, actually developed between the Cerma and the Musical Acoustic Laboratory [MAF97], would be able to describe the knowledge organization of a complex multisensorial townscape [FOL99a] and therefore to separate visual and sonic mental representations.

Ambients predictive system

Furthermore, our interest for ambients characterized by solar luminous and acoustics phenomena is guided by a twice interest:

• Concerning environmental parameters, sun and, more generally, lighting are constantly modeling our built spaces. So, they noticeably influence our practice of the urban space. Elsewhere, fundamental principles of architecture integrates the relationship between shape and light.
• Similarly, the importance of sound in urban spaces is denounced through the sonic pollution increasing. This fact has made up the capital rule of acoustic parameters for our environment. It is through these terms that architects have been interested in sound propagation phenomena in built spaces.
  

In order to estimate physical phenomena linked to these environmental parameters, we dispose of numerical simulation tools that allow us to predict their intensity and location :

- the lighting software application Solène developed by the Cerma laboratory [GRO95].

Figure 8. Lighting simulation in built environment by Solène .

- the acoustics software Orphea [WOL97a] developed jointly by the Cerma, the Acoustics Laboratory of the Maine University and the Laboratoire Central des Ponts et Chaussées.

Figure 9. Orphea's modulation principle .

 

At a more technical point of view, as our method lies on the urban void, it imports to define a calculation support there. Physical information resulting of the different simulations are applied on a meshed model. So we can obtain an informed numerical model for VR ambients processing.

Ambientscape

In order to structure our analysis method on this virtual meshed space, we defined a new notion we call Ambientscape. This notion is refereed to the different studies about landscape, mainly on the Murray Schäfer neologism Soundscape [SCH81].

Figure 10. The sun superposed to the built landscape .

For the pedestrian that circulates in the urban space, the sun superposed to the urban landscape is traducing a solar landscape. Therefore, we can define the Ambientscape as the totality of phenomenon distributed around a place, creating an atmosphere perceptible for any pedestrian situated in this space. It cannot be resumed to urban space volumes, it is revealing a new geometry of the city, modifying its shape. This ambients landscape reveals its dynamics at different levels:

• Even when freezing time to an instant t, walking in the new perspective city reveals new phenomena that punctuate space by distribution.
• Even without displacement, we can observe the continuously modifying Ambientscape during time, sun evolution and usage modifications. This induced a constant new phenomena distribution, interacting with the urban frameworks.

So we can develop two complementary interrogation methods concerning our space :

- An absolute method that describes the individual perception from a given position in the analysis space.

Figure11. The absolute method principle.

- A relative method that tries to approach closer the individual perception by a differential position analysis. This method is refereed to perceptible event intervening during the walk and contributes to its sensitive qualification.

Figure 12. The relative method operations.

The result can be represented with a cartography informing a point in the space (center of gravity of a patch), or with a graph where the two axes figures the intensity of the descriptor on one hand, and the instant of the walk on the other hand. This method allows the study of simultaneous events.

The "Ambianscape" descriptors

The descriptors are presented here around three poles : urban space geometry, lighting and acoustics.

Geometrically, space morphology description is made from a specific point of view. Length of view informs us on the vision depth of the field, indicating the urban built density. Clear sky factor indicator informs us on the visibility importance of the sky since a viewpoint ; it indicates also the importance of the urban built masking. The ratio indicator viewed built surface / total viewed surfaces informs us on the urban built pregnancy of a place. Furthermore, these two last descriptors are describing the confinement notion.

For lighting, we describe the perceived lighting state for an human being placed in this space. The percentile of visible brightened faces will express a notion not far from direct perception : is the space sunny or not sunny? This sunny space notion implies at least two parameters : is the vision point directly brightened? Is the visible space around from this point brightened? If the first parameter is directly clarified by the simulation, the second point needs specific descriptors such as solar spot surface views. This indicator renders a visible lighting state, the solar spot distribution in the vision field (are solar spot amalgamated in a well-defined zone or are they distributed in all the observation-point surrounding space?), and the average duration of lighting view of the place.

In acoustics, we'll talk about phonicity, representing the homeomorphism measure between the perceived sonic picture and its evocation [MOL90]. This notion describes both the quality of the identification of all components of a Soundscape by listening, and the capacity for the listener to find there a sense, that means to recognize a natural and relevant organization of its constitutive elements for identification.

Perspectives

This method is positioning the study in the field of ambients. It constitutes an overlay for simulation or measure data analysis. The possibilities offered by VR have a multiplicator effect on the ambients research field. It allows to consider this particular interaction with urban built environment in silico.

As mention before, the perception of the urban form results from a superposition of different interactive layers. The result of these interactions gives birth to a complex entity called Ambientscape. What is presented here for the lighting and acoustic could have equally been imagined for others environmental factors (wind, thermal,…).

The future way for our research is to increment the interactive function for the perceptor's wander and to improve the specific descriptors. The possibility offered by multi-agent system seems to be an interesting direction to approach those factors.

References used

[BUR93] Burdea G.; Coiffet Ph. La réalité virtuelle, Hermès, Paris, 1993

[DUB93] Dubois, D. Sémantique et cognition - catégorisation, prototypes, typicalité. (Sciences du langage, CNRS Editions, 1993)

[FOL98] Follut D. Ambiances et réalité virtuelle, application à l'analyse diachronique de la micro-climatologie urbaine Nantaise. In : Arévi 98 " Réalité virtuelle et projet urbain", Ecole Nationale d'Ingénieurs de Brest, Brest juin 1998

[FOL99a] Follut D., Groleau D.; The "Solarscape" or the sun as a revealer of urbans forms,in. sixth International Seminar On Urban Form, ISUF 1999, Florence Italie, 23-26 juillet 1999

[FOL99b] Follut D., Groleau D; "Ambianscope": An Immersion method to analyze urban environments; in Greenwich 2000 International Symposium Digital Creativity: Architecture. Landscape. Design; Greenwich, January 2000; Proposal for the symposium Computers and the urban environment and Digital aspects of Landscape environments accepted

[GRO95] Groleau D., Marenne C., Environmental specificities of the urban built forms, Rebuild - Rebuilding the european city, Corfou, 29-30 juillet 1995

[HUB94] Hubel D., L'œil, le cerveau et la vision, Pour la Science, Paris, 1994

[MAF97] Maffiolo V. (1997). Méthodes d'approche de l'environnement sonore urbain. Mairie de Paris, Direction de la Protection de l'Environnement, SPAAS, Paris, France.

[MOL90] Moles A., Art et Ordinateur, Blusson, Paris, 1990

[NAT97] Nathanail C., Lavandier C. Polack J.D. & Warusfel O.(1997). Influence of visual information on auditory perception. Consequences on the objective characterization of room acoustics. Proc. ASVA 97 (International Symposium on Simulation, Visualisation and Auralisation for Acoustic Research and Education), Tokyo, Japan (2-4 April 1997), 285-290.

[SCH81] Schäfer M. : " Le paysage sonore ", Ed. J. C. Lattès, 1981, 309 p. (éd. originale : " The tuning of the world ", A. Knopf Inc., New York) · SLA93] Slater M., Usoh M.; Representations systems, perceptual position and presence in immersive virtual environments; Presence, vol. 2, n°3, MIT Press;1993

[WOL98] Woloszyn P. Caractérisation dimensionnelle de la diffusivité des formes architecturales et urbaines. Nantes : Université de Nantes, 1998, 269p.+[5p.] Thèse de Doctorat Science de l'Ingénieur option Architecture

[WOL97a] Woloszyn P., Raymond F., Picaut J. Morphological parametering of a diffusive acoustic model for complex urban areas. in : Wesprac'97 6th Western Pacific regional acoustic conference, Hong-Kong, 19-21 novembre 1997. Hong-Kong : Hong Kong Institute of Acoustics, 1997, pp. 231-236

[WOL97b] Woloszyn P. Mesures multiéchelles du tissu urbain et paramétrage d'un modèle de diffusion acoustique en milieu construit. in : Symposium Saint-Venant : analyse multiéchelle et systèmes physiques couplés, Marne-la-Vallée, 28-29 août 1997. Paris : Presses de l'Ecole Nationale des Ponts et Chausées, 1997, pp. 89-96

Authors

Dominique FOLLUT, Laboratoire CERMA, UMR CNRS 1563,

Ecole d'Architecture de Nantes, Rue Massenet, 44300 Nantes, France.
follut@cerma.archi.fr.

Philippe WOLOSZYN, Laboratoire CERMA, UMR CNRS 1563,

Ecole d'Architecture de Nantes, Rue Massenet, 44300 Nantes, France.
woloszyn@cerma.archi.fr.