Upload
others
View
13
Download
0
Embed Size (px)
Citation preview
Enlightenment Approaches for Digital Absolutism.
Diplomatic Stepping-Stones Between the Real and the
Envisioned
Jörg REKITTKE and Philip PAAR
1 Introduction
The Lenné3D system is a prototype for a GIS-data based visualization tool for interactive exploration of three-dimensional landscapes. Lenné3D is primarily intended for use in landscape planning in order to enable scenarios and planned landscapes to be anticipated and explored by stakeholders and the general public. During preparation for the Lenné3D project in 1999, the development potential of three possible modes of representation was examined: a) photorealistic visualizations in the form of static and moving images; b) interactive visualizations with reduced graphic detail; c) interactive visualizations with sketchy rendition (COCONU ET AL., 2005). At the outset of the project in 2002, programmable graphics cards enabling real-time, photorealistic landscape visualizations on personal computers, seemed to represent the most promising approach. Accordingly, development of the Lenné3D software focused initially on the interactive simulation of landscapes with highly detailed, quasi-authentic representation of plants and vegetation. Technological progress has enabled us to render landscape scenes that, as a result of the level of detail of the simulation of vegetation, were seen to be convincing, even authentic. Nominal success of such elaborate visualizations is only attainable if the audience is presented with opportunities for mental participation in the digitally visualized scenario in order to explore the spectrum of possibilities between existing and potential realities. Authors such as TUFTE (1990) have illustrated the scope of visualization techniques for assisting in the comprehension of complex problems. What tends to be forgotten, however, is that if a picture can speak a thousand words, then there is a danger that a picture can cause more damage than can be repaired by a thousand words.
Landscape visualizations have, up until now, primarily been used to present, explain and market landscape planning scenarios (ORLAND, 1992). According to LANGE (1999), visu-alizations have usually fulfilled a decorative function, serving to “sell” a planning product, rather than being used to provide a meaningful contribution towards improving final re-sults. Visualizations from a first-person view come into conflict with the scale of conven-tional representations of community landscape plans, i.e. a greater degree of detail can call into question the accuracy of the basic data and the postulated planning scale of e.g. 1:10,000 (PAAR & REKITTKE, 2005). The following dilemma often arises in consequence: in order to recreate the impression of a photorealistic image, it is often unavoidable that
J. Rekittke and P. Paar
2
missing data must be fabricated or existing data manipulated in order to make them fit.
(ORLAND ET AL., 1997). SHEPPARD (1999) comments that sometimes correctly prepared landscape data is missing, but this cannot be discerned from the extremely realistic visuali-zation. Testing of the Lenné3D system during the Interactive Landscape Project Königslutter also drew attention to a further concern (WARREN-KRETZSCHMAR & V. HAAREN, 2005). It was found that users tended to focus on elements of the visualization that seemed inconsistent or lacking in detail, something which might have a negative effect on the plausibility of the entire visualization. This phenomenon whereby users focus on shortcomings has been de-scribed by Appleton and Lovett (2003) as the “smallest common denominator” effect.
There are differing views among specialists as to whether communication during planning can be most effectively supported by realistic visualizations that depict planned results as authentically as possible (OH, 1994), or whether more abstract representations are suitable (MUHAR, 1992). LEHMKÜHLER (1998) categorically rejects abstraction or sacrifice of indi-vidual elements in order to reduce visualizations to basic representations, stating that this would merely represent the subjective judgment of the creator of the visualized scenery. ZUBE ET AL. on the other hand, propose a differentiated method of approach, rebuffing the blanket rejection of any one mode of representation: “It is important to correlate the type of visualization to each project phase” (1987). ERVIN is clear about the fact that the means must necessarily be derived from the end goal: “[...] there is never a single correct answer to any of the many representational and abstraction problems [...], and so reference to the questions: “What is the purpose?”, and “What is the question?”, is an important touchstone for understanding visualization tasks and evaluating representations“ (2004). Visualization and media technology have one thing in common – in both cases, further development and innovation never lead to a complete replacement of a precursor but rather to its gradual marginalization or supplementation. This observation provides the basis for the three main objects of this study: a) The postulation of six theories for graphic reduction of planning visualizations; b) The presentation of graphic reduction methods specially developed for digital interactive representations of vegetation using the Lenné3D system; c) Recourse to tried and tested analogue graphics and communication techniques, imple-mentation of which in the digital tool box seems to make fundamental sense. All of the above points are considered in terms of the insights they provide for theoretical and conceptual development of the Lenné3D visualization system.
2 Catharsis
Since ancient times, the visual arts have been concerned with depicting realism. In ancient Greece from about the 4th century BC, illusionist effects were considered the most impor-tant criteria for assessment of a painter’s skill. The art of the ancient world was predomi-nantly based on the ideal of mimesis – Plato’s critiques on the artworks of the day may be understood as a critical reaction to the dominance of the mimetic (BOEHM, 1994). Plato was
Enlightenment Approaches for Digital Absolutism
3
of the opinion that pictures reproduce only the surface of objects, i.e. their external image, however much they seek to suggest the presence or essence of the object depicted. Plato considered the scientific or philosophical understanding of objects to be a more important task than their external depiction. Since its origins in the 1960s, the computer graphics industry has been concerned with generating synthetic images of simulated 3D environments that seemed almost to be photo-graphic snapshots of reality – the axiom of this digital wizardry was: photo-realism. As a result, the concept of photo-realism moved from being a purely artistic concept to becom-ing an object of research and development in the field of computer graphics, for example in the visualization of virtual worlds. Both paradoxical and fascinating is the fact that an im-age may be de facto non-realistic, yet at the same time photorealistic. The depicted reality may appear verisimilar without necessarily having any correlation in the physical world. Photorealistic computer simulations fundamentally entail exactness, perfection and authen-ticity – yet these aspects may be purely superficial (SHEPPARD, 2001). BUHMANN (1994) therefore concludes that the complaisance and verisimilitude of computer simulations cre-ates a burden of responsibility for planners. Digital landscape visualization methods must be oriented around both what is technologi-cally possible in the field of landscape modeling, computer graphics and around traditional representational techniques. Computer-generated images that are intended to be photoreal-istic tend to have a short aesthetic half-life. Only the pioneering early achievements of the computer graphics industry manage to retain lasting charm. In retrospect, these early achievements may be deemed the unintentional prototype of today’s highly current “non-photorealistic rendering”. Since the end of the 1990s, research in computer graphics has increasingly focused on “non-photorealistic representations” (NPR). The term “non-photorealistic” is today applied to computer-graphic representations, the individual ele-ments of which may contain realistic illustrative aspects, but whose representation clearly diverges from reality in terms of the depiction of aspects such as form, color, structure, shading, light, shadow and perspective. The aim of NPR is to communicate and transport complex information more quickly through imitation of manual representational tech-niques. Let us recall: new media and representational techniques do not replace their pre-cursors, they supplement them. With the realization that photorealistic images may not always be the most suitable method for communicating visual information, a process of collective reform has begun. If, until now, the aim of high art was to reproduce reality in the manner of photography – i.e. to appear photorealistic – then it now seems time to depart from this goal. To apply an anal-ogy from economics: the market for photorealistic images has become saturated. Saturation occurs when the marginal utility reaches zero – in other words, when no further utility can be attained despite all efforts: on the contrary; beyond the point of saturation, all effects are negative. Today it has never been easier to create pictures that lie using commercially available, reasonably priced tools and systems. In the days when there was no alternative to analogue photography, the “retouching” of pictures constituted a real art and was employed for beautifying pictures or creating propaganda (Fig. 1). Nowadays, the oversupply of
J. Rekittke and P. Paar
4
desktop publishing tools and filters beguiles the user into manipulating almost everything – be it press photos, job application photos or private holiday snaps. This inflationary prac-tice of manipulating images and the accompanying inflationary misuse has irreparably damaged the credibility of professional and technically sophisticated digital visualizations – people no longer always believe what they see. For who does not remember large-scale architectural projects from the 1990s that were promoted through the use of what were at that time impressive computer visualizations? Now, comparing the final result with those images, one cannot help but smile – the lesson has been learned; digital fantasies can quickly evaporate.
a) b)
Fig. 1: Pictures that lie. a: Mussolini with the „Sword of Islam“ in Tripolis, 1942, b: Stylization of the „Duce“.1
It is particularly important for those fields that traditionally rely on visualizations of future scenarios to consider the consequences of this fundamental loss of credibility (SHEPPARD, 2001) – in other words, for all planning and design disciplines that rely on visual commu-nication of their projects. This paper focuses on visualization practices in landscape plan-ning and landscape architecture – professions whose task it is to turn planned landscapes and spatial designs into reality. These landscape professionals, working as they do with images of the future, must proceed more diplomatically – i.e. shrewdly and subtly – than ever. Visible change often gives rise to mistrust and doubt, a reflex that can and must be countered by systematic reassurance and guarantees of control. Unquestionably, photoreal-istic images can provide a valuable service in communicating landscape planning and de-signs, e.g. they can depict future visual experiences. That said, the photo-realism of today must be critically viewed as an anticipatory gesture, in light of its power of suggestion, its potential to offer seductive, visual ersatz-realities and that fact that it is symptomatic of the inflationary flood of images. Photorealistic simulations elevate visual representations to the level of perfect “as-if” scenarios, in which fact and fiction converge (BOEHM, 1994) to the
1 Picture credits: Stiftung Haus der Geschichte (eds.) (1998): Bilder, die lügen. Bonn: Bouvier Ver-
lag: 12.
Enlightenment Approaches for Digital Absolutism
5
point that they can no longer be held apart. Communication in planning must, however, systematically guard against the dissemination of misunderstanding or false statements. It is not difficult to evoke images; it is virtually impossible, however, to retract images once they have been evoked. Communication during planning must therefore issue constant reminders of the intrinsically tenuous nature of landscape development prognoses (MUHAR 1992).
3 Trust-Building Measures
Once this prevailing uncertainty on the part of “image consumers” has been recognized and accepted, planning professions can begin to work at systematically integrating trust-building measures into planning visualizations and creating “diplomatic stepping-stones” or “footholds” for participants. Target groups must be given the feeling of being in control and in a position to differentiate between truth and speculation within digitally generated visualizations. The keywords in this process of regaining visual control are: graphic reduc-
tion and abstraction. Reduction refers to the decrease or removal of elements, whilst ab-straction involves relinquishment and generalization, describing things that are not neces-sarily objective or directly correlated to reality. The use of maximum level of detail creates highly complex representations that are intellectually hard to decode — there is no way of differentiating and hence comparing what is fact and what is fiction, what is extant and what is proposed. The logical consequence of this has been the above-mentioned loss of credibility and the necessary response must be the systematic de-precision, reduction and abstraction of the flood of visual information. ERVIN (2001, p. 62) writing about abstraction, states that “(...) we landscape modelers must also remember the valuable roles of abstraction in both cognition and communication, and not believe that ‘photo-realism’ – or even ‘physical realism’ – is the be-all, end-all of digital modeling. We make models to make explorations or to convey messages, and the infinite variety of explorations and messages will surely yield an equally boundless variety of digital landscape models”. Something else that speaks against photo-realism as the ultimate solution for landscape visualizations is the fact that methods of representation for planning purposes are required to highlight and depict intended, i.e. future changes. If this does not happen, observers are no longer in a position to judge what is important. Planning necessarily involves building in a certain amount of room for maneuver; without this there can be no success. Prognoses of planned changes can never be one hundred per-cent accurate – strict adherence to intended yet unrealizable developments leads to inflexi-bility and lack of compromise. Leeway and tolerance between the cogs of complex ma-chinery is vital in order to avoid failure and frustration. The necessity of building in leeway to counteract chance occurrences can be seen, for example, in public transport timetables that are calculated on the basis of ideal conditions – but which are continually thrown out by unforeseen circumstances such as the weather, accidents etc. This is a particularly nega-tive example of poor planning due to the considerable annoyance and frustration such
J. Rekittke and P. Paar
6
mismanagement provokes for the travelers – in this case the target group. Greater leeway would lead to greater understanding and tolerance on the part of customers, as well as in-creasing the chances of smooth operations and the credibility both of the timetables and of the service provider. Landscape and spatial planning is a process with an infinite number of variables. From vegetation – which, in contrast to architecture, is constantly growing and changing – through to political “weather conditions”, most factors cannot be precisely predicted and calculated during the planning process. Acceptance of this fact must be documented through the knowingly vague representation of the predicted or planned landscape change.
4 Six Theses for the Graphic Reduction of Visualizations in
Planning
Graphic reduction refers here to the particular approach of the Lenné3D system, by which the reduction process starts from a highly detailed visualization in order to not to waste the high modeling effort required to generate photorealistic images. Simplified representations are thus able to profit effectively from the capacity for detail of the high-performance soft-ware. The graphic reduction procedure outlined here and illustrated through examples in the following chapter is distinguished by the fact that it fundamentally rejects any use of artistic license – something, which can often lead to huge misinterpretations during plan-ning. The simplified depictions of graphically reduced visualization objects are not the product of a stroke of genius but are based on a scientifically demonstrable process of re-duction. This method of approach is consistent with the self-perception of landscape plan-ners and architects that they are doctors, in a figurative sense; though they do not perform operations on people, they must nonetheless intervene in the process of nature and land-scape with the dexterity and precision of a surgeon. (REKITTKE, 2002). Digitally generated, graphically reduced visualizations for planning possess inherent advantages over their photorealistic cousins, advantages that are often ignored in current practice. This stance may initially seem trivial, yet it is true to say that the range of digital tools available and the opportunities that have been created as a result have marginalized the advantages of graphically reduced forms of visualization in the eyes of users. Because the obvious is often overlooked, in the following, the aims and means of using graphic reductions in the context of digital visualizations for planning are examined as working hypotheses. 1. Graphic reduction as a voluntary control mechanism.
The less concrete the planned scenario, the more graphic reduction can provide a safe-
guard against planners having to invent details or use too much artistic license.
The generalization of visual information in planning should lead to an increase in the suc-cess rate of planning projects and help individual planning participants to identify with the
Enlightenment Approaches for Digital Absolutism
7
intended results. However, graphic reduction does not eliminate the need to represent varia-tions, or alternative possible courses of development.
2. Differentiating between reality and fiction.
Reality may be represented through realistic forms of representation (“photo-realism”).
Fiction, on the other hand, must be distinguished through abstract forms of representation.
As self-evident as this principle may seem, it is often ignored in planning and visualization practice. Not being able to see the wood for the trees is a phenomenon that occurs in plan-ning communication when existing elements, which are often familiar to the planning par-ticipants, can no longer be clearly distinguished from new design elements. 3. Congruence of importance, and visual clarity.
The various levels of importance of proposed planning measures must be represented
clearly through a corresponding level of visual clarity and various stages of graphic reduc-
tion. A picture can speak a thousand words, but not everybody will interpret an image in the same way. Visual communication must be characterized by systematic guidance through the visualized scenario, in order that the important aspects might be emphasized and the less important aspects relativized. Visual clarity is not the same as visual perfection – the intended level of importance must be subtly balanced through various levels of graphic reduction. 4. Graphic gradation of successive procedural steps.
In addition to the various levels of importance of different planning elements, the temporal
sequence of the realization is also significant for successful planning and must be illus-
trated.
In the presentation of landscape planning projects, it is often only the final intended results that are presented. In order for implementation of the project to be successful however, the temporal, hierarchical or pragmatic sequence of the individual steps is also relevant and should be represented accordingly using various graphic techniques. Tedious phases of development and construction that will impose a burden on both people and the environ-ment, as well as intermediate stages of the process must all be represented in temporal cross-sections. In this sense, the planner may seem to assume the role of the non-fiction or documentary writer, who presents pictures of changing landscapes in time-lapse form in a didactic manner. The computer-graphic visualization of intermediate phases of landscape change is expanded on in HUMPHRY REPTON’S concept of “before and after” pictures (1803), which made both his Red Books famous and his landscape designs successful.
5. Making the invisible visible.
Landscape planning as such is essentially imperceptible, and is only ultimately manifested
once planned measures have been realized, or requirements have been complied with.
Furthermore, landscape plans create effects that may not be visually perceived, or which
are outside the scope of the general imagination. Yet these effects should also be visually
tagged during the planning process, in a procedure that we name “inverse abstraction”.
J. Rekittke and P. Paar
8
Landscape planning projects tend to have a greater sphere of effect than their ground plan would indicate. This can be positive – for example when a pleasant aura is exuded; how-ever, it can be detrimental, spreading negative effects or creating encumbrances. These factors are important for decision-making during the planning process, yet their impact can only be assessed if they are illustrated or visualized. There is much unexplored terrain for using visualizations in this area of planning, and much potential that requires investigation. Interestingly, that which cannot be seen in reality cannot be represented through realism – non photo-realism, however, will certainly play an important role in this area of planning in the future. 6. Graphic asceticism instead of media hedonism.
The more comprehensive the digital graphics toolbox becomes, the greater the challenge
and necessity of practicing graphic minimalism in planning communication.
The catalogue of media possibilities must be simplified, and clear information content distilled prior to presentation of the project to laypersons. The half-life of digital effects is short – initial enthusiasm quickly turns to frustration in most cases. Graphic variables (BER-
TIN, 1982) such as color, form etc. should therefore only be used in suitable measure.
5 Graphic Reduction Methods for the Interactive Representa-
tion of Vegetation
The above hypotheses for abstract visualizations in planning provide the basis for devel-opment and utilization of graphic reduction methods which, using the Lenné3D system, should help enable planned future landscapes be more easily conceived and envisioned. These methods were developed with a view to creating “diplomatic stepping-stones” be-tween real and envisioned landscapes. In the following pictorial examples all reductions are derived from color, high-resolution, photorealistic (i.e. reality-based) visualization objects. The main protagonist here and the thematic object of interest is the 3D visualization of vegetation – the aspect that has traditionally proved to be one of the greatest challenges for landscape representations. We therefore propose a turn from photorealistic representation towards “embedded” connotation, i.e. the visualization of planning information applying graphic variables on the plant objects.
The X-ray corner – behind the scenes
Even the most perfect computer image is incapable of mirroring reality completely. In order to illustrate this important fact, at the bottom right-hand corner of every visualization and image generated by the Lenné3D system there appears a section of the wire-netting model that exists beneath the surface (Fig. 2). This X-ray view leaves no room for doubt that these are not photographs but artificial landscapes. We are demanding clarity and we are trying to fulfill this criterion.
Enlightenment Approaches for Digital Absolutism
9
Fig. 2: The X-ray corner
Unstylized, abstract representations
We seek to create abstractions of complex representations without imposing or generating an individual characteristic style. The graphic reduction methods used by the Lenné3D system are always derived from visualization of objects with the maximum level of detail; therefore it uses the high modeling effort necessary in order to create photorealistic repre-sentation of trees and other plant life. The graphically reduced representation of landscapes and especially trees is not understood as a creative task but rather as an unstylized mini-malization of detail. Representation of vegetation in particular can only guarantee sure identification of individual genera and types through maximum botanical authenticity. Even in abstract form, specific plant features should not be idealized artistically but should be instantly recognizable from the representation. With each passing year, nature demon-strates how visual reductions similar to those described above function effectively: when trees shed their leaves their appearance is dramatically reduced. From a distance, the loss of color means that they appear as near-black silhouettes. Yet the trees do not lose their par-ticular character and defining characteristics during this process, nor do they lose any of their spatial presence or capacity for orientation (REKITTKE ET AL., 2004). Grading the level of detail and representational form in congruence with planning
stages
Lenné3D’s graphic techniques can be used to represent objects at various stages of the planning project in different ways. The level of detail and the form of representation are graded in logical correlation to the stage of the project. In the following example, five modes of representation of both chronological stages and planning stages are presented using the example of a planned tree (Fig. 3). A. “Silhouette” – Silhouette representation for the drawing-board phase
In the initial phase of planning it is sufficient to work with a very rough model of the plant. A silhouette model allows a first impression of the plant’s size, habitat and its spatial effect
J. Rekittke and P. Paar
10
on the surroundings. The silhouette may be moved around the scene freely, and enables attention to be focused on fundamental issues in this early planning stage. B. “Blur” – rough adjustment of location
Once the initial draft has been completed there comes the task of arranging the elements, carrying out preliminary checks and roughly adjusting intended locations. During this phase it is important that the planner already has a more exact idea of the plant – this is why at this stage details are included in color, even though the plant form is indistinct. This lack of focus indicates on the one hand the fact that the element is not yet firmly anchored and that there is still flexibility, on the other hand this indistinctness is the product of an effect that harks back to analogue methods of visualization and which is particularly effec-tive: in order to gain an overall impression of a design it helps to narrow your eyes. Though the focus is blurred, it enables a good impression of the spatial and visual ensemble. Nar-rowing the eyes also helps when looking at a screen or projection – in this case, the unfo-cussed tree is then on the same level as the rest of the visualization. Just try it and see. C. “Transparent” – Fine tuning of the location
Once the elements have been brought into some semblance of order, the fine-tuning of subtleties and correlations can begin. In this phase, the plant is depicted as transparent, but in color with a high level of detail. This allows the planner to assess how the plant will look in the chosen surroundings, as well as see what will be behind the plant. In this manner, the intended effect may be calculated with a high degree of precision. D. “Grayscale” – preliminary fixing
Once the fine-tuning has been carried out, the object and its location may be provisionally fixed. Partial color reduction towards a grayscale representation allows clear visual parameters to be generated without relinquishing the high resolution and level of detail. Existing elements remain in color, while planned elements are depicted in grayscale. E. “Perfect” – a photo of reality
In contrast to a color portrayal, the grayscale representation of the previous phase clearly signals that there is still something missing, the image is unfinished. The maximum level of representation is deliberately held back in order to emphasize the room for maneuver that still exists at this stage of the planning process. Only at the end of the planning process — if desired during presentation of the project — should photorealistic representations be employed, thereby allowing the planned future landscape to be represented visually as lifelike as possible.
Enlightenment Approaches for Digital Absolutism
11
a) b)
c) d)
Fig. 3: Example of graphic gradation of successive procedural steps. a: “Silhouette”, b: “Blur”, c: “Transparent”, d/e: “Grayscale/Perfect”
Contour representation for depicting disappearance and planned removal
Landscape planning entails not just creating new elements but also the removal of existing landscape elements. It is often necessary to create new space, while other landscape ele-ments may disappear unplanned. Elements that must be removed or which will otherwise disappear, in this case plants, are depicted as outlines. The residual contour allows for a free view of the remaining landscape scene.
Color reduction to denote flashbacks
Today, moving (motion) color pictures are standard practice. If whole sequences are shown in black and white, the viewer almost automatically associates them with temporal flashbacks – an effect which can be utilized systematically. This retrospective effect is amplified through use of sepia colors. Reduction of color provokes a reflex, based on vis-ual conventions, that evokes an association with the past. Black and white films belong to the past; color represents the present.
J. Rekittke and P. Paar
12
Sketches as a basis for and as extracts of realistic visualizations
Basic sketches are suitable for developing complex scenes. AL-KODMANY (1999, 2002) and WARREN-KRETZSCHMAR & TIEDTKE (2005) describe participatory planning processes that utilize a combination of artists’ freehand sketches and information technology. Even the most complicated special effects in Hollywood films are still developed in sketch form with hand-drawn storybooks. These provide valuable examples for computer-generated graphics in planning. The Lenné3D system is able to reverse the chronology of sketches and final images and generate sketches (Fig 4, left) from photorealistic visualization objects (Fig. 4, right), thereby facilitating structural assessment of the visualized objects. Even in sketch mode, the simplified images are derived from highly complex computer models and can be devoid of any artistic stylizations.
Fig. 4: Mixed representation (left): combination of graphic realism (representing reality) and graphic reduction (representing fiction)2; full graphic realism (right)
6 The Good Old Days
In order to develop reduced, simplified forms of representation that benefit communication during planning, it is worth making use of traditional visual conventions. The so-called “new media” merely constitute additions to the existing repertoire of visual representation techniques; as such they should operate in conjunction with and exploit the existing vo-cabulary or visual conventions – there is no need to reinvent the wheel. The development of the human race is accompanied by gestures, symbols and signs that are based on conven-tions and that therefore have a broad, if not universal sphere of effect. For example, no-where in the world does a red traffic light mean anything other than “stop”. In this case, abstraction provides the greatest possible clarity.
2 Provisional example (Screenshot 3D-Player, L. Coconu, 04/2005)
Enlightenment Approaches for Digital Absolutism
13
Now that so much time and energy has been put into exploring all the possibilities of photorealistic simulations, the opportunity arises for critical reflection of this “hype”. The fundamental issues to be addressed with every visualization in planning come to the fore: Whom is the visualization aimed at? What should be shown? What is fundamentally im-portant? What is less relevant? Photorealistic forms of visualization are a blind alley, from which the only exit is to retrace one’s steps. On the return path one encounters old ac-quaintances, ones that had possibly not been sufficiently appreciated or valued on the out-ward journey. Tried-and-tested methods should be integrated into new visualization tech-niques – the way forward for communication in planning lies in the combination of old and new approaches. Some possible methods for combining the two are presented below. The human factor and telepresence
If proportions of scale in perspective representations are generally hard enough to judge, then the perspective representation of landscapes is harder still. The Lenné3D system ad-dresses this problem by means of the interactive scale function “family with child” which can be turned on or off as required. Experience has shown that, in the free navigation through 3D computer landscapes, it helps the navigators to have an illustration of themselves as subjects – this doesn’t have to be the obligatory handgun of the so-called ego-shooters. For the walk-through perspective this illustrative reference – termed “telepresence” (DRAPER ET AL., 1998) – could be provided by footsteps, for example. Saving or freezing the footsteps of previous navigators in the landscape could provide orientation and a reference of paths already trodden for subse-quent explorers. It is also possible that, over time, certain preferred paths might emerge, providing a useful indicator of common use and areas of interest. Layer fading and highlighting
In order to steer the user’s eye towards certain landscape elements or structures, objects and layers of vegetation can be faded in or out. Application of this layering technique also allows differentiation between different types of measures. Highlighting of certain elements – much in the manner of representations of biblical halos – also helps to make marked elements an object of focus. Sketch function “6b”
Sketching paper and soft pencils – a few quick strokes of the “6b” by the art teacher (6b refers to the thickness and softness of the pencil lead) will stand out in the memory of every one-time student as the quickest and most effective method of correcting any draft. We propose placing the “6b” in the hands of the users of interactive 3D visualization systems, so that they might intervene in the image without any ado. Use of Text
Many aspects and interconnections that cannot be unequivocally represented visually may profit from linguistic descriptions. The abstraction level “text” allows room for interpreta-tion and imagination in the mind of the spectator. Where the limits of visualization tech-niques are reached, conventional forms of planning communication should not be forgot-ten.
J. Rekittke and P. Paar
14
Stickies
Users are given the opportunity to react to the presentation of moving landscape scenes by placing annotated post-it® notes next to objects (Fig. 5). Thoughts and ideas are quickly forgotten – the label, easy to hand allows them to be noted and subsequently discussed with other participants in the planning process. The conventional notepad form systematically counteracts any misplaced reverence for digital graphics, and emphasizes the fact that the scene is not yet set in stone. This interactive tool encourages and helps planners to collect local knowledge, suggestions or criticism on an ad-hoc basis from those involved with or affected by the planning initiative, and to frame the information within a 3D geo-referenced object space.
Fig. 5: Interactive annotations. Logbook
A logbook is used to record the succession of visual presentations, and allows the digital occurrences to be recalled at will much in the manner of a flight recorder. In the simplest case, screen shots document the chronology of communication during planning. The pattern chart effect, or visual palimpsest
Finally, one last example to provide food for thought and impetus for further ideas. Planning uses theoretical and graphic layering – levels are added to existing surfaces. Each new planning level leaves behind traces that are taken over in the next level. This chain of processes can be comprehended if representation of the various levels are able to be turned on or off both individually and simultaneously – often this results in complex, even confus-ing pictures which, however, clearly reflect the complexity of the planning process. The image of a pattern chart provides a clear idea of the nature and complexity of such forms of representation – each planning project contains numerous possibilities for variation.
Enlightenment Approaches for Digital Absolutism
15
7 Conclusions & Outlook
What use is a photograph to a mechanic if he already has the real thing in front of him (LANSDOWN & SCHOFIELD, 1995)? This is the issue faced by landscape planners. If spe-cific aspects or structures of a landscape are to be seen and understood, it is better to have a graphically reduced or figurative representation than a photorealistic one. The graphic reduction methods for the interactive representation of vegetation presented above complement ERVIN’S (2004) definition of four abstraction levels. While ERVIN
works in concept pairs, linking geo-specific with realistic, and geo-typical with illustrative,
the methods shown here allow both a geo-specific and an illustrative representation. The drive towards photorealistic perfection should not be eschewed – it is, after all, a hu-man characteristic to seek perfection wherever possible. An illustrative reproduction of a landscape impression that is as close to reality as possible is of particular relevance for planning and scientific analysis in areas such as landscape scenery evaluations, visual im-pact or visibility analysis, or as stimuli in experiments. The fascination with perfect images
is unquestionable and the skill necessary to create them laudable. The fact remains, how-ever, that depicting realism or photo-realism is just one of many methods of visual repre-sentation or communication using graphics. Just as no method of representation has ever succeeded in completely replacing or quelling any other, it is clear that reduced graphic visualizations will never completely replace photo-realism. However those whose job it is to represent landscape and urban planning projects and make them palatable must be aware that the success of their project rests largely on the extent to which they focus on deciding factors and create room for individual interpretation and imagination. Everyone agrees that “less is more” – yet this truism is often thrown overboard amidst the concert of technical possibilities. Computer-generated landscape visualizations have the potential to be developed into a perceptually efficacious “user interface” in landscape planning processes. The precondition thereof is to provide participants in the planning process with stepping stones with which to negotiate the “vacuum” between existing and potential realities, and which give them op-portunities for mental access in digitally visualized scenarios. There are no limits to the range of such “diplomatic” tools; further research is required to catalogue such options systematically, and also to demonstrate that planners take their work seriously and are prepared to learn from their mistakes.
8 Acknowledgements
We would like to thank our colleagues from the Lenné3D team for their support. In particular, we would like to thank Liviu Coconu for discussing the graphical methods from a computer graphics point of view and seeking to implement them in the Lenné3D-Player. The German Federal Environmental Foundation (DBU) sponsors the “Lenné3D” research project.
J. Rekittke and P. Paar
16
9 References
Al-Kodmany, K. (1999): Combining Artistry and Technology in Participatory Community
Planning. Berkeley Planning Journal, 13: 28-36. Al-Kodmany, K. (2002): Visualization Tools and Methods in Community Planning: From
Freehand Sketches to Virtual Reality. Journal of Planning Literature, 17: 27-35. Appleton, K. & A. Lovett (2003): GIS-based visualisation of rural landscapes: defining
“sufficient” realism for environmental decision-making. Landscape and Urban Plan-ning, 65: 117-131.
Bertin, J. (1982): Graphische Darstellungen und die graphische Weiterverarbeitung der
Information. Berlin, New York: Walter de Gruyter. Boehm, G. (1994): Die Wiederkehr der Bilder. In: Boehm, G. (Hrsg.): Was ist ein Bild?
München, Fink: 11-38. Buhmann, E. (1994): Technische Möglichkeiten - EDV in der Landschaftsbildsimulation.
Garten + Landschaft, 10: 31-32. Coconu, L., C. Colditz, H.-Ch. Hege & O. Deussen (2005): Seamless Integration of Styl-
ized Renditions in Computer-Generated Landscape Visualization, this volume. Draper, J.V., D.B. Kaber & J.M. Usher (1998): Telepresence. Human Factors, 40: 354-375. Ervin, S.M. (2004): Landscape Visualization: Progress and Prospects. ESRI 2004 UC
Proc. URL: http://gis.esri.com/library/userconf/proc04/docs/pap1647.pdf Ervin, S.M. (2001): Digital landscape modeling and visualization: a research agenda.
Landscape and Urban Planning 54 (1-4): 49-62. Lansdown, J. & S. Schofield (1995): Expressive Rendering: A Review of Nonphotorealistic
Techniques. IEEE Computer Graphics and Applications, 15(3): 29-37. Lange, E. (1999): Realität und computergestützte visuelle Simulation. Eine empirische
Untersuchung über den Realitätsgrad virtueller Landschaften am Beispiel des Tal-
raums Brunnen/ Schwyz. ORL-Berichte, 106. Zürich: VDF. Lehmkühler, S. (1998): VRML – 3D-Standard des Web - Chance für die Raumplanung.
CORP 98 Computergestützte Raumplanung, TU Wien. Muhar, A. (1992): EDV-Anwendungen in Landschaftsplanung und Freiraumgestaltung.
Stuttgart: Ulmer. Oh, K. (1994): A Perceptual Evaluation of Computer-based Landscape Simulations. Land-
scape and Urban Planning, 28: 201-216. Orland, B. (1992): Data visualization techniques in environmental management: a consoli-
dated research agenda. Landscape and Urban Planning, 21: 241-244. Orland, B., C. Ogleby, H. Campbell & P. Yates (1997): Multi-Media Approaches to Visu-
alization of Ecosystem Dynamics. In: Proceedings, ASPRS/ACSM/ RT’97, Seattle, American Society for Photogrammetry and Remote Sensing, Washington, DC., vol. 4: 224-236.
Paar, P. & J. Rekittke (2005, to appear): Lenné3D - Walk-through Visualization of Planned
Landscapes. In: Bishop, I.D. & E. Lange (eds.), Visualization in landscape and envi-ronmental planning, Spon Press, London: 152-162.
Rekittke, J., P. Paar & L. Coconu (2004): Dogma 3D. Grundsätze der non-
photorealistischen Landschaftsvisualisierung. Stadt+Grün, 7: 15–21.
Enlightenment Approaches for Digital Absolutism
17
Rekittke, J. (2002): Drag and Drop – The Compatibility of Existing Landscape Theories
and New Virtual Landscapes. In: E. Buhmann, U. Nothhelfer & M. Pietsch (eds), Trends in GIS and Virtualization in Environmental Planning and Design, Proc. at An-halt University of Applied Sciences 2002. Heidelberg: Herbert Wichmann Verlag: 110-123.
Repton, H. (1803): Observations on the theory and practice of landscape gardening: in-
cluding some remarks on Grecian and Gothic architecture. Taylor, London; Phaidon, Oxfort (facs.).
Sheppard, S.R.-J. (2001): Guidance for crystal ball gazers: developing a code of ethics for
landscape visualization. Landscape and Urban Planning 54 (1-4): 183-199. Sheppard, S.R.-J. (1999): Regeln für die Nutzung der digitalen Kristallkugel. Garten +
Landschaft, 11: 28-32. Warren-Kretzschmar, B. & C. v. Haaren (2005, to appear): Online landscape planning –
What does it take. A case study in Königslutter am Elm. In: E. Buhmann (ed.), Online landscape architecture, Proc. at Anhalt University of Applied Sciences. Heidelberg: Wichmann.
Warren-Kretzschmar, B. & S. Tiedke (2005): What Role Does Visualization Play in Com-munication with Citizens? – A Field Study from the Interactive Landscape Plan, this volume.
Tufte, E.R. (1990): Envisioning Information. Cheshire: Graphics Press. Zube, E.H., D.E. Simcox, & C.S. Law (1987): Perceptual landscape simulations: history
and prospect. Landscape Journal, 6 (1): 62-80.
Pictorial Quality and Communication of Information in a Rural
Landscape Context
Pernette MESSAGER BELVEZE and David MILLER
1 Introduction
Cartographic semiology considers the different visual variables that are used in the construction
of maps. Based upon on several constituents, representations of territory or localizable
phenomenon are guided by cartographic language and rules. Such a graphical system of
representation defines the transcription of data (BERTIN, 1964), within a process where the
interpretation (mainly statistical), leads to the emergence of the key elements and mechanisms
required for representing information suitable for its communication. The language used has to
be visual, obeying the general rules of visual perception, such that they are understandable,
clear and coherent to the intended audience. The outputs from the research should provide
guidance on the understanding and meaning of visualisations (BEGUIN & PUMAIN, 1994).
The paper presents the preliminary results of a test of visual semiology, in the context of a
management plan, the aim of which is to increase levels of biodiversity. The hypothesis is that
qualitative information is required for communication of the content and nature of a landscape,
and quantitative information is required for ease of identification of features (TUFTE, 2001).
2 Background to Cartographic and Visual Semiology
2.1 Semiology
Semiology, or ‘semiotics’, is a science grouping a range of themes: languages, sounds and
pictures. It is a study of the use of signage used in different systems of perception, to define
‘keys’ to unlock interpretation of features and the meaning being communicated by the
medium. Primarily based on the linguistic semiology sign system (F. de SAUSSURE, 1916)
pictorial semiology concepts are used to explain polysemic pictures for communicating several
levels of message (BARTHES, 1964; ECO, 1976). Therefore, the meaning is produced at the
source (the encoder) and received by the receptor (the decoder).
The relationship between reality and representation includes a mono-linkage based on
arbitrariness and understanding by cultural convention. Pierce (In ECO, 1984) has criticized
this approach as being restrictive and established a semiotic system based upon three signs:
icons (similarity to the object), indices (indication of physical presence) and symbols (arbitrary
link). Such a system is always under development, with new signs being created to fulfil new
P. Messager Belveze and D. Miller
2
needs. In visual communication, the system is adapted to the context, and the ordering from
representation to connotation evolves from verisimilarity, indexation, iconicity and symbolism.
To ensure the complete understanding of the message by several different receptors, the sign
system has to be grounded on a relationship that avoids, as much possible, arbitrary alternative
interpretations. The research aim is to devise a ‘semio-system’ for use with rural landscape
imagery or real-time models, which is consistent with BERTIN’S (1983) semiology, to convey
the same message, when read at the same level, by different observers.
2.2 Visual Variables
To produce a cartographic representation, the cartographer uses elementary signs, and the
combination of the signs is configured through seven visual variables: shape (point, line,
polygon), size (quantitative information, e.g. quantity), colour (qualitative information, e.g.
species), value (order expression, e.g. density), texture-structure (quantitative variation within
one of the above), grain (quantitative and qualitative variation within one of the above) and
orientation ( direction of the variable used).
In addition, the transparency of a flat tint can also be considered as a visual variable, linked
with geographical location. In the context of landscape visualisation, in a vertical or horizontal
view, the user is often being offered thematic representations that require modifications to the
underlying cartographic representation. The user may require to correct one or several visual
variables to produce a clear and understandable document.
2.3 Visual Signs
The visual communication through a semiotic system can be described with a sign classification
going from abstract meaning to physical reality (SHERMAN, 2002). In the context of this
study, the focus is on four representations, which are verisimilitude, indexation, iconicity and
symbolism. Fig. 1 presents the gradient of information content in one of the representations
used in the study.
Verisimilar Indexed Iconic Symbolic
Fig. 1: A ‘realism axis’ of the Scots Pine (adapted from SHERMAN, 2002)
False colours + species denomination (dark green) Tree biologically realistic Handmade model
Pictorial Quality and Communication Information in a Rural Landscape Context
3
The verisimilitude, named “photo-realism” or “biological realism” in the landscape modelling
field, represents the highest level of similarly to visual reality. In this example, the Scots pine is
one of the ‘Xfrog’ library (GREENWORKS, 2005). Each branch and needle is represented by
a geometric primitive, and the model can be oriented, its age set, and the result is a modification
of the tree to fit different stages in its growth.
Indexed objects are more generic than the verisimilar ones, with the generalization of shape
supporting the level of feature recognition. The model presented is extracted from a photograph
and neither the age nor the orientation can be changed. The size can be adapted but does not
reflect biological realism. The iconic representation of the Scots pine provides two forms of
information (in the context of virtual landscapes). These are, for instance, the genre coniferous
is represented by the triangular shape and the species by the colour. In the model, every species
has one colour, the broadleaves are distinguishable by the lollipop shape. A key gives
information about the species. The Symbolic representation contains one single meaning: the
type of tree. As for the previous images, the size variable can be changed.
In this realism axis, the closer the object is from the verisimilitude, the more specific the
instance.
2.4 Relating Semiologies
A correspondence can be identified between the cartographic visual variables and the visual
features at three levels.
1. A correlation between vertical objects and the visual variables of shape, size and colour.
2. The use of planimetric inputs (orthophotography, maps, images of textures or procedural
textures) in relationship with texture-structure and grain visual variables.
3. Value and orientation, which may be considered as intuitive perception: the value, in this
context, the density of high vegetation, is perceptible directly when there is a big variation
within the neighbourhood, but asks for more attention when the difference is minor.
2.5 Semiologies and pictorial quality
We could define a pictorial quality as a mode of representation. In a landscape representation
context, the term pictorial quality includes the GIS database and the landscape modeller. In
terms of data, the level of graphic generalisation or the amount of detailed information within
the database attributes defines the informative content of the pictorial quality, when the
landscape modeller, through the choice of the visualisation tools and the filter of his own
knowledge and skills, defines the visual quality of the picture (APPLETON 2003, SHEPPARD
2001, DEBARDIEUX, 2003)
A distinction between qualitative and quantitative visual information has been defined,
assuming two working definitions. The qualitative visual information refers to aesthetics codes,
the realism of the features representation and the closeness to the verisimilitude. The
P. Messager Belveze and D. Miller
4
quantitative information refers to more abstract representations, to which verbal information
(tree species) and geographical information may be added.
The survey described considers a high level of information (verisimilitude of representation and
nominative attributes of Geographic Information System (GIS) data) and a low level of
information (generalized or schematic representation and aggregated GIS attributes).
3 Methods
3.1 Survey Structure
To test the hypothesis, visualization tools are being used to produce different combinations of
diagrammatic and realistic representations of a landscape, which are then used in the
assessment of the capability of different levels of pictorial quality to convey information about
a rural landscape to public and expert audiences. The visualisation materials involve,
principally, the representation of vegetation, traditional hand-made features, physical landmarks
and patterns of agricultural and forestry land uses, and use the software packages Visual Nature
Studio (3DNature) and ERDAS VGIS (the 3D viewer of ERDAS IMAGINE v8.6).
Two steps in the survey development have to be distinguished. The first survey was a prototype
the results of which are presented here, included only three pictorial qualities (Diagrammatic,
realistic and verisimilar). The actual survey’s results, based on four pictorial qualities
(Diagrammatic, iconic, realistic and verisimilar) will be presented at the conference.
The design of the survey uses a framework of four combinations of qualitative and quantitative
information, combining cartographic and visual communication semiology (Table 1). The
results presented below come from the first prototype where only three of these different visual
qualities are modeled using the same landscape from three different view points, each
viewpoint containing different information and repartition (Table 2).
Table 1: Framework of qualitative and quantitative information being tested
Name Qualitative information Quantitative information
Diagrammatic Low Low
Iconic Low High
Realistic High Low
Photo-realistic High Medium to High
Pictorial Quality and Communication Information in a Rural Landscape Context
5
Table 2: Summary of view points, view content and information level
View point View type Information View length Composition in perspective
VP1 Ground level 3 < 500 m 53 % Open area
7 % Forest
40 % Sky
VP2 Ground level 5 < 1 km 47 % Open area
14 % Forest
38 % Sky
VP3 Overview >5 > 1 km 52 % Open area
33 % Forest
15 % Sky
The first page of the survey provides an introduction to the objectives and content, followed by
images in the four visual qualities, displayed from the same view point. The view point is
chosen randomly for each participant, from a set of three possibilities. The responses to the
questions on ease of understanding and the assessment of rural landscape representation,
determine which set of three view points of only one pictorial quality is then displayed. The
observer is asked to choose their favorite view point, from which a third set of paired questions
are posed, the choice of image being based upon the information and models used to represent
the landscape in the favorite image. Fig. 2 summarizes the steps of the survey.
Fig. 2: An overview of the steps in the survey and the expected outcomes
OutcomesSurvey steps
Information on visual quality preferences
for easy understanding
Information on view point’s perspective and
level of information
Random set of view point
Selection of one visual quality
Three viewpoints with the same
visual quality
Three pictures representing three visual
qualities, from the same viewpoint
Information on identification of features and
relevancy of chosen models
P. Messager Belveze and D. Miller
6
3.2 Images Used in the Final Survey of Pictorial Quality
The visual qualities being tested (diagrammatic, realistic and photo-realistic) were presented in
the first prototype of the survey. The Iconic pictorial quality has been added as the combination
between low qualitative information and high quantitative information (Fig. 3).
Fig. 3: Images used in the survey of visual qualities, showing four levels of information
content and three view points
4 Results
The survey prototype has been displayed at three major public events: The Countryside
Awareness Day (May 2004), the Royal Highland Show (June 2004), the Aberdeen Technology
Festival (September 2004), which attracted a large number of agricultural related people (22%
of the respondents) and students and pupils (27%), and several smaller events. There are 110
relevant questionnaires used for the prototype survey analyses.
Table 3: Sample composition: Age, gender and place of living
Age
<12 16 Gender Dwellers
12 - 17 16
18 - 30 25
31 - 40 19
Men 64 Urban area 47
41 - 50 21
51 - 60 8
>60 5
Women 46 Rural area 63
Diagrammatic quality
Iconic quality
Realistic quality
Photo-realistic quality
Pictorial Quality and Communication Information in a Rural Landscape Context
7
0
5
10
15
20
25
Aca
demic
and
teac
hing
Agr
icul
ture o
r fishi
ng
Cen
tral o
r loc
al gov
ernm
ent s
ervi
ces
Energ
y indu
strie
s
Environ
men
t related
pro
fess
ion
Hou
sewife / H
ouse
hus
band
Land
man
ager
Lands
cape
relate
d pr
ofes
sion
Leisu
re o
r tou
rism
indu
strie
s
Scien
tific
and
tech
nica
l
Schoo
l pup
ils
Stude
nts
Fig. 4: Sample composition: Activities
4.1 Ease of Information Interpretation
Survey Question: How easily can you recognize the different features in the landscape?
A comparison between pictorial quality (opposition easiest - hardest) shows that the
diagrammatic pictorial quality is the hardest to understand in comparison with the easiest,
which was the realistic pictorial quality. The photo-realistic images appear to be rated as
neither difficult nor easy to understand (Fig. 5).
The urban public has a clear choice about the easiest representation: the realistic pictorial
quality (almost 50%). The choice of the hardest representation choice is quite close between
Photo-realistic (38%) and Diagrammatic (32%). The rural publics easiest choice is between two
pictorial qualities: the realist representation (42%) and the photo-realistic (37%) whereas the
Diagrammatic representation is harder to understand (almost 50%) (Fig. 6).
Fig. 5: Ease of instant recognition of pictorial quality (all view points)
percentage of respondents
0 .0
5 .0
1 0 .0
1 5 .0
2 0 .0 2 5 .0
3 0 .0
3 5 .0 4 0 .0
4 5 .0
D ia g r a m m a tic R e a l is t ic P h o to - r e a l is t ic
E a s ie s t
H a r d e s t
P. Messager Belveze and D. Miller
8
Fig. 6: Comparison of results of ease of interpretation of landscapes with respect to
urban and rural dwellers
Fig. 7: Comparison of results of ease of interpretation of landscapes with respect to
view point
Fig. 7 shows the distribution of interpretation of ease or difficulty in understanding the content
of the four visual qualities for each of the three view points. At VP1, there is a clear difference
between the easiest-hardest for the realistic (B) and the photo-realist pictorial quality (C). At
VP2, the differences in opinions are minor between the realistic (B) and the diagrammatic (A),
and there is no great opposition to the use of this level of quality. For VP3 the distinction
between realistic-diagrammatic is strong, as is that between photo-realistic (C) and
diagrammatic (A).
4.2 Rural Landscape Representation
Survey Question: How well do these pictures represent a real landscape?
The analysis uses a ranking of 5 classes, from ‘very badly’ to ‘very well’. Generally, the
diagrammatic representation is not a good way of representing scenery. Although people
consider this representation at ground level (VP1) as being acceptable (class 3 to 5: 60% for
VP1, 40% for VP2 and VP3). Whereas, the realistic representation is accepted by more than
0
10
20
30
40
50
60
Iconic Realist Photo-realist
Easiest
Hardest
0
10
20
30
40
50
60
Iconic Realist Photo-realist
Easiest
Hardest
Diagrammatic Realistic Photo-realistic
Pictorial quality Diagrammatic Realistic Photo-realistic
Pictorial quality
Ease of recognition: urban dwellers Ease of recognition: rural dwellers
0
5 10 15
20 25 30
35 40
View point 1 View point 2 View point 3
A /B B/A A/C C/A B/C C/B
Pictorial Quality and Communication Information in a Rural Landscape Context
9
80% (classes 3, 4 and 5) as a good way to represent a rural landscape. The photo-realistic
representation is accepted at more than 90% (Fig. 8).
Fig. 8: Responses to the images representing a real landscape
4.3 Viewing Perspective
Survey Question: Which view point do you prefer?
VP3 is selected as the favourite view point by 52% of the respondents; VP2 is identified by
40%, and the VP1 by 9%. People living in an urban place prefer the VP2 (22% of answers,
compared with 18% of rural habitants), whereas people living in the rural place prefer the VP3
(35% of the answers, and 15% of urban respondents) (Fig. 9).
Fig. 9: Preferences expressed for the view from each view point
4.4 Identification of Landscape Features
This theme considers two questions.
1. Are the chosen models relevant or too complex to each pictorial quality and for the basic
signification they transmit?
2. Do people have some common base to their understanding of rural landscapes?
9%
40%
51%
0
10
20
30
40
50
60
VP1 VP2 VP3
Urban
Countryside
0 %
2 0 %
4 0 %
6 0 %
8 0 %
1 0 0 %
V P 1 V P 2 V P 3 V P 1 V P 2 V P 3 V P 1 V P 2 V P 3
V e r y w e l l
4
3
2
V e r y b a d l y
Photo-realisticDiagrammatic Realistic
5
4
3
2
1
5
4
3
5
4
3
2
5
4
3
2
1
5
4
3
2
5
4
3
2
5
4
3
2
5
4
3
2
1
5
4
3
2
1
P. Messager Belveze and D. Miller
10
Survey Question: How many types of ground vegetation and woodland can you identify?
Table 4 shows the amount of land cover information by view point. The results suggest that
people are responding to the information in the aerial photograph, which includes more colours
and variation in ground texture.
Table 4: Number of types of land cover information by view point and
pictorial quality
Diagrammatic Realistic Photo-realistic
VP1 3 > 5 3
VP2 5 > 6 5
VP3 5 > 8 7
From the questionnaire the main conclusions regarding the identification of different types of
land cover is the importance of the aerial pictures. Despite a low quality of picture (e.g. an
error of colour correction during the mosaic process), the realism of the colour and grain gives a
mental association with the natural colour and grain of the existing ground texture. However,
the wording of the question might not have been clear enough as people could have
misunderstood it: was it “counting the number of patch” or “counting the different class of
visual similarity”? Table 5 shows the results for each view point.
Table 5: Percentage of respondents for each VP
Diagrammatic Realistic Photo-realistic
VP1 33 56 22
VP2 48 68 45
VP3 39 77 23
Survey Question: How recognizable are the following features?
In this question, features such as the broadleaves, the coniferous woodland and the ground
vegetation are three features which the respondents are asked to assess. This assessment is
required to define the relevance of the object used to represent a natural feature and the
interpretation of this object by the observer.
Respondents appear to have reacted positively to the representations of the trees in the
diagrammatic type of image. However, the use of monochromic ground cover, in a uniform
texture was the least acceptable of the three visual qualities. The ground cover for the realistic
imagery was rated as the most acceptable of the three visual qualities, closely followed by that
in the photo-realistic images. The results are summarized in Fig. 10.
Pictorial Quality and Communication Information in a Rural Landscape Context
11
Fig. 10: Preferences expressed for the interpretation of land cover
features for each view point
5 Discussion
It appears that respondents concentrated on the information in the images, and considered
aesthetics secondarily. This first conclusion follows Debardieux statement about the risk of
using an aesthetical representation aiming to an immediate response from the observer rather
than using a less aesthetical representation aiming for easier understanding and thinking.
Further statistical analysis will be required to enable conclusions to be drawn about the
relevance of the different visual qualities with respect to the perspectives they show. Future
analysis will also seek to explore differences in understanding with respect to the structures of
the land cover content, in terms of view point and public perception.
The representation of the ground surface may have a negative visual impact on the perception
of the view, at least in terms of responses to the visual impression of the foreground, due to the
lack of clarity in the image details in the graphics used. Kati Appleton, in a survey focused on
realism of representation, crossed similar results for the foreground representation: the drapped
ortho-photograph has a lower rating than the procedural texture, used here for the verisimilar
pictorial quality.
Such effects will require to be tested and analysed for to aid in the design of visualisations, and
to better understand the factors which should be considered when selecting view points (e.g. to
avoid a point in which text would be presented close to the viewer in the image foreground).
The inclusion of the fourth pictorial quality, the iconic representation uses visual keys, such as
colour by species and a topographic map for representation of the ground surface will be tested
in a follow-up survey. The inclusion of this image is required to enable an analysis of the
understanding of a representation that couples the cartographic approach (i.e. the use of
distinctive visual keys) and perspective views of the landscape.
2129 27
1120
26 2615 17
35 2421
18
25
25 23
3020
44 48 51
71
5549 51 55
63
0%
20%
40%
60%
80%
100%
VQ1 VQ2 VQ3 VQ1 VQ2 VQ3 VQ1 VQ2 VQ3
Acceptable
Neutre
Unacceptable
Diagrammatic Realistic Photo-realistic
P. Messager Belveze and D. Miller
12
6 Acknowledgements
The authors wish to acknowledge the funding of the Scottish Executive Environment and Rural
Affairs Department and the EC VisuLands project for this research. Thanks are also due to
Andrew Thorburn for the implementation of the WWW survey. The internet address of the
survey is: http://ginkgo.macaulay.ac.uk/landscapevisualquality/
7 References
Appleton, K. & A. Lovett (2003): GIS-based visualisation of rural landscapes : defining
‘sufficient’ realism for environmental decision-making. Landscape and Urban Planning
65
Barthes, R. (1964): Elements of Semiology, Hill and Wang
Béguin, M. & D. Pumain. (1994): La représentation des données géographiques. Paris, A.
Colin
Bertin, J. (1983): Semiology of Graphics. University of Wisconsin Press
Debardieux, B. & S. Lardon (2003) : Les figures du projet territorial. Ed. de l’Aube - DATAR
Eco, U. (1976): A Theory of Semiotics. Indiana University Press, Macmillan
Eco, U., & T.A. Sebeok (1984): The Sign of Three: Dupin, Holmes, Pierce (Advances in
Semiotics), Indiana University Presses
Greenworks Organic Software, (2005): Xfrog 4.2. www.xfrogdownloads.com
De Saussure, F. (1916): Cours de linguistique générale, Payot
Sheppard, S. (2001): Guidance for crystal ball gazers: developing a code of ethics for
landscape visualization. Landscape and Urban Planning 45
Sherman, W.R. & Craig, A.B. (2003): Understanding Virtual Reality: Interface, Application,
and Design, Morgan Kaufmann
Tufte, E.R. (2001): The Visual Display of Quantitative Information, Graphics Press USA
Optimizing the Visualization of 3D-Information for
Participative Planning of Landscape Development Concepts
Ulrike WISSEN, Eckart LANGE and Willy A. SCHMID
1 Introduction
The integrative planning of landscape development concepts with the balancing of ecological, economic and social aspects requires the cooperation of various stakeholders. Thereby, the exchange of knowledge is crucial to enable all participants to an objective assessment of planning alternatives (SELLE, 1997). As maps and texts turn out to be hard to understand for lay people, there is an increasing use of 3D-visualizations to show concepts in a clearer way (LANGE ET AL., 2003a). To ensure a useful design of these visualizations as information and communication tools in participatory settings, it is strongly recommended to include scientific results on human perception (LANGE, 2001) and cognition (KLIMSA, 2002) as well as didactic principles (DEMUTH & FÜNKNER, 2000). This paper describes research undertaken in a Ph.D. thesis concentrating on editing information relevant for the participative planning of landscape development concepts by means of 3D visualizations. A better comprehension and estimation of consequences of activities and their effect on landscape change over a long period of time by lay people is aimed for. Results from psychological and pedagogical research are used to identify requirements for a user friendly design of 3D visualizations as regards information communication. The focus is laid on the image functions and the needed degree of realism and complexity in the visualizations, with the aim of minimising the effort of content understanding. The stronger inclusion of the human factor in the development of the 3D visualizations as a tool for ‘shared learning’ should provide for principles with regard to design aspects and lead to a component for a standardized 3D landscape visualization method.
2 Characteristics of Information
Information is knowledge shaped in a definite way to convey or adopt messages, as well as the result of the conveyance and adoption. Crucial characteristic of the knowledge to be transferred is that it serves for a specified use, it contains significant aspects of an issue, and it is composed and bound to a communication medium ruled by conventions (BOLLMANN, 2001). This definition leads to different dimensions with influence on the production of 3D visualizations as a communication medium for participative landscape planning:
• The use / function of the 3D visualization in a specified planning situation.
U. Wissen, E. Lange and W. A. Schmid 2
• The aspects of an issue relevant for the planning to be shown with the 3D visualization.
• The conventions for the production of the 3D visualization. Also, indications for the chronology of the visualization process can be derived from this definition. First of all, the planning process and the implementation of the visualizations define the functional character of 3D visualizations. The information needed in the specific planning situation is decisive for the content to be visualized. Both require for a certain design of the visualization so that this can be decided on when these factors are known. In the following chapter, the functions of images regarding communication and learning of knowledge with results from psychology and pedagogy are specified. In addition, design aspects assisting these functions are mentioned.
3 Functions of 3D Visualizations for Knowledge
Communication
Images can have various functions. Studies assessing the efficiency of learning with images underline the following functions as important for human information processing (LEWALTER, 1997, WEIDENMANN, 2002):
• Motivating the beholder by activating his interest and curiosity.
• Situating the content in a larger context and providing for relations to the real world.
• Demonstrating a significant aspect.
• Supporting the construction of complex mental models or providing for an addition to an existing mental model.
These aspects are also important in phases when information has to be communicated in a participative planning context, e.g. in moderated workshops (SEIFERT, 2002). In the beginning of a workshop the audience has to be interested for the topic. Images are very suitable for this task as they can provoke something because of their direct emotional effect. Motivation can be reached with images that have an appellative function calling for an action. So called push images or images acting as a stimulant that are giving an impulse fulfil this function. “Push” is a term used in advertising and techniques used in this branch of business are suitable to produce images with appellative function (DOELKER, 1999), e.g. provocative, attractive or shocking images and headlines as well as animated images that cause a subconscious automatic reaction (DRANSCH, 1997). Certainly, it has to be deliberated carefully about the choice of the technique regarding the aim of the workshop, especially if ‘shocking images’ should be used. Images providing for a scenario, in the sense of a certain setting or scenery, help to integrate detailed information in a larger context. Scenarios can activate the imagination of the situation from everyday life experiences due to effecting the emotions (WEIDENMANN,
Optimizing 3D-Information Visualization for Participative Landscape Planning 3
2002). These situating images can be spatial or textual. They give a hint to the real situation or a real event in space and time (DOELKER, 1999). A (photo)realistic design of visualizations is most effective but can cause a search for errors which affects the process of getting involved. Therefore, it is suggested to reduce the content to characteristic elements illustrating the situation (DRANSCH, 1997). In contrast, scientific visualizations are produced with the intention rather to uncover relationships of visual or non-visual functions than to represent reality. The visualization of spatial-functional landscape relationships, e.g. the impact of agricultural management over time on the landscape functions as economic, ecological, and social area, can be a very complex issue (LANGE & WISSEN, 2004). With 3D visualizations, the data can be reduced to a compact format that is simultaneously presented to the human eye. Using the abilities of human perception, the time to take up the information can be optimised (HEINTZ & HUBER, 2001). In this context, so called demonstrational images are useful for showing and explaining landscape functions. Research on visual perception has proved that solely presenting the required information can maximise the speed of perception (EASTERBY & ZWAGER, 1984) Abstract contents can be shown more comprehensible by stressing significant characteristics and omitting irrelevant features. These are the main techniques used to draw the attention of the beholder on the important and characteristic elements (DOELKER, 1999; WEIDENMANN, 2002). One problem is that the accuracy of data may suffer due to the reduction of complexity and false interpretation is possible (EASTERBY & ZWAGER, 1984). Simulating images e.g., future scenarios can be seen as coaching for a lack in imagination. The benefit of these images with regard to landscape development is that experience can be anticipated. This means that it could be learned from errors without conducting them. This is a rather crucial function as decisions that can cause harmful and not reversible environmental impact must not be made (DOELKER, 1999). The advantage of 3D visualizations compared to maps or texts is that the function of these images is not solely to explain certain issues, but to include and address the emotional aspect of the experience, too. Beyond this, the better imagination of different dimensions of a landscape development can be the basis for individual ideas of alternatives (LANGE ET AL., 2004). Like this the visualizations can contribute to the construction of complex knowledge structures (mental models) and activate their further enhancement. Images assisting in the comprehension of complex sections of reality by providing visual information on the elements and their interrelationships have to deal with the problem of complexity as well (WEIDENMANN, 2002). In summary, it can be said that the design of 3D visualizations with the use of didactical principles requires for the arrangement of the information content based on the function the visualization should have in the communication process. STROTHOTTE (1998:19) states that “information visualization deals with how to depict an extract of an information space, how to visually communicate important features of data”. The recommendations found regarding the design of images supporting the itemised functions, are rather general and refer to various aspects. Nevertheless, in pedagogic studies it was determined that a medium degree of reality and complexity of images had a positive effect on the learning
U. Wissen, E. Lange and W. A. Schmid 4
performance (LEWALTER, 1997), and thus the comprehension of the issues. In the following chapter, the terms “realism” and “complexity” are defined as to their meaning for information visualization in the context of landscape planning.
4 Realism and Complexity in 3D Landscape Visualizations
Landscape visualizations are a more or less abstracted imitation of the complex reality (LANGE, 2001). Reality can be understood as the sum of the existing and the actually given, and thus it is concrete. The opposite of “concrete” is “abstract”, in the general sense, by separating the essential from the randomly existing (BROCKHAUS, 1997). Regarding the prior described image functions, information visualization requires for different degrees of abstraction. These have to match the content to be transferred by containing the most important information needed for a specific purpose. In geological, biological and environmental sciences the expression “complex” stands for spatial or systematic structures that appear as functional entities or holisms. These can be differentiated by partial complexes standing for a part of the whole system. For example vegetation, soil form and water balance are key characteristics in landscape ecology expressing complicated relationships, and representing a part of the geo-ecological system respectively (LESER, 1997). In 3D landscape visualizations the differentiation of the complex spatio-functional relationships of the real landscape are an important task to make these interactions more comprehensible as they are almost only partly considered in landscape planning processes (LANGE ET AL., 2004) Visualizing smaller entities of the whole system indicating the extent of specific functional relations or their spatial dimension for a particular issue are challenging but have proven to be doable (HEHL-LANGE, 2001). A comprehensive evaluation of landscape development concepts requires the combination of these indicators which is again a task of complexity for their visualization. How these principles found can be used, is analysed amongst other items in the frame of the EU-project ‘VisuLands’. 3D landscape visualizations are applied to current planning tasks of participative landscape planning processes in the Entlebuch UNESCO Biosphere Reserve (UBE), Switzerland. With social science techniques the responses of the users are evaluated (LANGE ET AL., 2003b). The 3D visualizations given in the next chapter show examples for relevant issues and demonstrate possible applications of the principles found regarding image functions and degree of abstraction and complexity.
5 Examples for 3D Visualizations and Their Functions in a
Participative Planning Context
In Fig. 1, a scenario on forest growth on the alpine farms in the UBE is shown. It was used in the initial workshop on the participative elaboration of the Forest Development Plan in the UBE and had principally the function to motivate the participants to give their own
Optimizing 3D-Information Visualization for Participative Landscape Planning 5
opinion on the shown topic. The 3D visualizations depict the provocative hypothesis by the facilitators of the workshop that the forest growth affects valuable habitats and leads to a monotonous view of the landscape. The visualizations were produced with a medium level of realism and a rather low level of complexity. For the representation of the agricultural land use classes, textures are used that allow for a very coarse classification in different pastures and thus habitats. All forest is made up of billboards of firs (Picea abies). The whole process of forest growth on alpine pastures due to management changes and the affect on landscape ecosystems is rather complex and the functional relationships are manifold. Making it less complex, the functional relationships were differentiated and only the aspect of the changes in the land use and the correlating changes in the visual landscape were demonstrated.
Fig. 1: Examples of 3D visualizations presenting a scenario of forest growth on alpine farms in the Entlebuch UNESCO Biosphere Reserve used in a communicative planning context as a stimulant. (Visualizations VisuLands 2004; Geodata © Cantone Luzern, GIS-Coordination office).
In the actual workshop, the participants reacted actively on the hypotheses presented on the posters. However, the reaction could not be assigned clearly to the visualizations as the text had stimulating effect as well. For a better location of the presented situation the images were linked to a 3D visualization providing for the overview on the area. This 3D model is characterised by an abstract representation of the landscape with a topographic map draped on the relief. Few labels
U. Wissen, E. Lange and W. A. Schmid 6
name the landmarks important for the orientation. The degree of complexity is rather low, too. The polygons show the position of the pastures visualized in the more realistic visualization. A line from the abstract model to the more realistic one indicates the view point of the latter. Feedback on the abstract overview model give first hints that it allows for a better orientation than 2D maps. However, one has to be careful with the change of the perspective in this model during the presentation as beholders get lost very quickly.
Fig. 2: Situating the single more realistic view in a larger spatial context by the link to a more abstract 3D visualization providing for the overview on the area. (Visualizations VisuLands 2004; Geodata © Cantone Luzern, GIS-Coordination office).
The 3D visualizations of the development of a small area demonstrate the impact of changes in the grazing system on the alpine pastures in the study area over time. Here the focus is laid on the development of the visual landscape quality and on the shift in the presence of species at vegetation level. The degree of realism is the highest compared to the visualizations shown in the other figures. But even though single species on the pastures are represented, there is an abstraction of the actual vegetation to indicator species of the mapped vegetation types. Participants of a workshop in the UBE regarded these visualizations as useful to explain crucial landscape processes to people not familiar with the area.
Optimizing 3D-Information Visualization for Participative Landscape Planning 7
Fig. 3: 3D visualizations demonstrating the impact of a shift in the grazing system over time (from left to right: today, within 5 years, within 30 years). (Visualizations VisuLands 2004; Geodata © Cantone Luzern, GIS-Coordination office).
These are only a few examples of the design and application of 3D visualizations with different functions assisting in the comprehension of landscape planning issues. Results from empirical testing in planning practice reveal their potentials for effective information communication. Nonetheless, for the definition of explicit design guidelines the findings are too weak as yet.
6 Conclusions & Outlook
Visual communication is of interest for various disciplines but while the practical side of visual communication design is very well established, the theoretical part - analysing the production, the product and its impact - lacks literature (MÜLLER, 2003). In addition, only general advices for a user friendly design of images can be found. As the degree of realism and complexity is of importance for both, the human process of managing the received information, and the technical process of structuring and designing the content presented in the visualizations, these aspects will be further analysed. In the next step, the 3D visualizations will be tested in an experiment to get an answer on the degree of realism and complexity appropriate to support the functions of images that are assisting lay people in the comprehension of information relevant for the planning process. The development of design principles is not only of interest for an effective information transport. A lower degree of realism and complexity of production comes with lower time consuming design effort and rendering time. This aspect is crucial with regard to the economic efficiency of the images in the landscape planning practice. It should be kept in mind that with the increased use of 3D visualizations in landscape planning, expectations for their added value for participative planning processes are high. Therefore, it is more important than ever to work on a guideline for the production of these images to avoid weak or ineffective design that restricts the communication of information or lead to false interpretations. Optimizing the representation of content in 3D visualizations regarding the functions of the images in the communication process can help to avoid mistakes before they are put into practice.
U. Wissen, E. Lange and W. A. Schmid 8
7 Acknowledgements
This work is part of the project «VisuLands– Visualization Tools for Public Participation in the Management of Landscape Change», funded by the European Union 5th Framework Programme, Quality of Life and Management of Living Resources, and the Swiss State Secretary for Education and Research (SER). Special thanks apply to the Canton Lucerne for providing the digital data of their geographical information system and to the public of the Entlebuch UNESCO Biosphere Reserve for testing the 3D visualisation tools in their local planning processes.
8 References
Bollmann, J. (2001): Information. In: Bollmann, J., Koch, W.G. (eds): Lexikon der Kartographie und Geomatik in zwei Bänden. – Erster Band, A bis Karti. Spektrum Akademischer Verlag Heidelberg, Berlin, 396-398
Brockhaus, F.A. GmbH (1997): Brockhaus – Die Enzyklopädie: in 24 Bänden. – 20. überarb. und aktual. Aufl., F.A. Brockhaus GmbH, Leipzig – Mannheim
Demuth, B., Fünkner, R. (2000): Einsatz computergestützter Visualisierungstechniken in der Landschaftsplanung – Chancen, Risiken und Perspektiven. In: Gruehn, D., Herberg, A., Roesrath, C. (eds): Naturschutz und Landschaftsplanung – Moderne Technologien, Methoden und Verfahrensweisen, Festschrift zum 60. Geburtstag von Prof. Dr. H. Kenneweg, Mensch und Buch Verlag, Berlin, 97-111
Doelker, C. (1999): Ein Bild ist mehr als ein Bild. Visuelle Kompetenz in der Multimedia-Gesellschaft. Klett-Cotta, Stuttgart
Dransch, D. (1997): Funktionen der Medien bei der Visualisierung georäumlicher Daten. Geoinformatik_online 3/97, http://gio.uni-muenster.de
Easterby, R., Zwager, H. (eds) 1984: Information Design. The design and evaluation of signs and printed material. Written by contributors to the NATO Conference on Visual Presentation of Information, Het Vennenbos, Netherlands, 1987, John Wiley and Sons Ltd., Chichester, New York, Brisbane, Toronto, Singapore
Hehl-Lange, S. (2001): Structural elements of the visual landscape and their ecological functions. Landscape and Urban Planning, 54: 105-113
Heintz, B., Huber, J. (2001): Der verführerische Blick. Formen und Folgen wissenschaftlicher Visualisierungsstrategien. In: Heintz, B., Huber, J. (eds): Mit dem Auge denken. Strategien der Sichtbarmachung in wissenschaftlichen und virtuellen Welten. Institut für Theorie der Gestaltung und Kunst. Edition Voldemeer, Zürich, Springer, Wien, New York, 9-26
Klimsa, P. (2002): Multimedianutzung aus psychologischer und didaktischer Sicht. In: Issing, L.J., Klimsa, P. (eds): Information und Lernen mit Multimedia und Internet. Lehrbuch für Studium und Praxis. Verlagsgruppe Beltz, Psychologische Verlags Union, Weinheim, 7-17
Lange, E. (2001): The limits of realism: perceptions of virtual landscapes. Landscape and Urban Planning, 54: 163-182
Lange, E., Schroth, O., Wissen, U. (2003b): Interaktive Landschaftsentwicklung. Potenziale von Computervisualisierungen in partizipativen Planungsprozessen zur
Optimizing 3D-Information Visualization for Participative Landscape Planning 9
Lenkung der Landschaftsentwicklung am Beispiel der UNESCO Biosphäre Entlebuch (Schweiz). DISP 155.4: 29-37
Lange, E., Schroth, O., Wissen, U. (2004): Fallbeispiele von 3D-Visualisierungen zur partizipativen Landschaftsentwicklung. Geomatik Schweiz, Geoinformation und Landmanagement, 4, 246-249
Lange, E., Schroth, O., Wissen, U., Schmid, W.A. (2003a): Anforderungen an Visualisierungstools zur Partizipation der Öffentlichkeit bei der Bewertung der Landschaftsentwicklung. Proceedings CORP 2003, Wien, 553-556.
Lange, E., Wissen, U. (2004): Ein Bild sagt mehr als 1 000 Worte. Kooperative Entscheidungsfindung mit Hilfe der 3D-Visualisierung. Garten+Landschaft, 11, 51-52
Leser, H. (1997): Diercke – Wörterbuch Allgemeine Geographie. Deutscher Taschenbuch Verlag GmbH & Co. KG, München und Westermann Schulbuchverlag GmbH, Braunschweig
Lewalter, D. (1997): Lernen mit Bildern und Animationen - Studie zum Einfluss von Lernermerkmalen auf die Effektivität von Illustrationen. In: Rost, D.H. (Hrsg.): Pädagogische Psychologie und Entwicklungspsychologie, Bd. 2, Waxmann Verlag GmbH, Münster; New York; München; Berlin
Müller, M. G. (2003): Grundlagen der visuellen Kommunikation. Theorieansätze und Analysemethoden. UVK Verlagsgesellschaft mbH, Konstanz
Seifert, J. W. (2002): Visualisieren, Präsentieren, Moderieren. Der Bestseller – überarbeitet und erweitert. 18. Auflage, Gabal Verlag GmbH, Offenbach
Selle, K. (1997): Planung und Kommunikation: Anmerkungen zur Renaissance eines alten Themas. DISP 33.129: 40-47
Strothotte, T. (1998): Computational Visualization. Graphics, Abstraction, and Interactivity. Springer, Berlin, Heidelberg
Weidenmann, B. (2002): Abbilder in Multimediaanwendungen. In: Issing, L.J., Klimsa, P. (eds): Information und Lernen mit Multimedia und Internet. Lehrbuch für Studium und Praxis. Verlagsgruppe Beltz, Psychologische Verlags Union, Weinheim, 83-93
Display Methods for Real-Time Landscape Models: An
Initial Comparison
Katy APPLETON and Andrew LOVETT
1 Introduction
Due to continuing improvements in available hardware and software, interactive, real-time
landscape models are becoming a more viable proposition both for landscape professionals
involved in communication and decision-making, and researchers investigating
perceptions, opinions and responses to landscape issues (LOVETT et al., 2002; STOCK &
BISHOP, 2002). The prime forces behind the development of many of these techniques, in
whole or in part, are military users and industries such as aerospace and automotive
engineering, oil and gas exploration, and of course entertainment. Their demands have led
to sophisticated viewing methods being developed alongside software and hardware,
including head-mounted displays, multi-sided “caves”, and screen shapes including curved
planes and spherical segments, any of which may be further enhanced by the use of
stereography (KASIK et al., 2002).
Such non-traditional viewing equipment is often marketed as providing a more immersive
experience, thereby implying that the viewers’ experience and understanding will be better.
When undertaking environmental decision-making exercises based on computerised
landscape models, it is of considerable importance to be aware of any potential differences
introduced by the way in which the information was displayed to the audience, yet little
such information exists at present. Research on the effects of using different kinds of
graphics displays has focused to date on the ability of viewers to complete a particular task
such as wayfinding or component recognition (e.g. SWINDELLS et al., 2004, KASIK et al.,
2002), but landscape research and environmental decision-making usually involve audience
perceptions and other more subjective assessments of the information on display. RICE
(2003) examined audience perceptions of an urban campus design comparing a
hemispherical screen display to a physical scale model and a real-life walkthrough, but
much of the information gathered was more relevant to aspects of the built environment
than the large natural or semi-natural areas usually considered in environmental decision-
making (e.g. the balance between ceiling and walls, or plants and paving). It was therefore
decided to assess the effects of using one such piece of non-traditional display equipment, a
1.5m diameter spherical section screen (Elumens Vision Station:
http://www.elumens.com), to show a real-time interactive landscape model of a relatively
large area.
K. Appleton and A. Lovett 2
2 Data & Methods
2.1 Input Data
The landscape model employed in this study is a 2 x 3 km area in the Humberside region of
England (Figure 1). It consists of a flat, floodplain area, a wide river (the Trent)
constrained by embankments, a small village (East Butterwick), and scattered farms among
arable fields. This area was chosen primarily for the ready availability of geographic data
and landscape models from ongoing work on climate change scenarios.
GIS data provided the base for the visualisations. A 10m resolution elevation model
(Ordnance Survey Profile® data) was used with land use information derived from
Ordnance Survey MasterMap® and fieldwork. Land use data showed the layout of fields,
water, buildings and built-up areas as well as the location of a set of electricity pylons.
Crop information was not contained within the OS data, but was derived to a basic level
from aerial photography online at Multimap (http://www.multimap.com). While crops
could not be reliably identified, the distribution of different vegetation (e.g. cereals, root
crops or bare earth) could at least be discerned and a “best-guess” label given to each field.
Fig. 1: Location and detail of the area for the landscape model
2.2 Landscape Model
The model was constructed using Visual Nature Studio version 2 from 3DNature
(http://www.3dnature.com), with the additional Scene Express real-time export extension.
GIS files were imported, and visual properties assigned to all landscape features; as only a
low-detail model, suitable for real-time display, was to be produced, less effort was
required than for a full-detail still rendering or animation. Buildings were shown using
simple extrusions with flat roofs instead of importing full 3D geometry; the river was
represented by a bluish-grey flat colour on the land surface, rather than simulated water
with reflectivity, transparency and rippled surface; and instead of many individual items of
Display Methods for Real-Time Landscape Models: An Initial Comparison 3
crop vegetation, procedural textures were used to represent land cover types. Labels (tall
“billboards” anchored to a specific surface point) were used to inform viewers of the field
contents as well as to identify points of interest (the river, village, specific farms). The
landscape has few trees, but a set of three distinctive poplars were included, and trees were
also included in the “village” area As the 10m resolution of the elevation data was
insufficient to show the river embankments and field ditches, these were added in at the
visualisation stage using the “Terraffector” feature in VNS. This allows for local terrain
modification according to a user-defined elevation profile, applied symmetrically along a
vector.
In preparation for export to real-time format, a number of pre-set fixed and animated
cameras were set up. These comprised an overhead view, a circling camera at an altitude of
approximately 500m, a stationary ground level view, and a panning ground level view.
Such cameras give an alternative overview of the model for viewers who do not wish to
navigate interactively.
Fig. 2 Screenshots giving an overview of the landscape model used.
3DNature’s own NatureViewExpress format was used to export and view the landscape
model. This was chosen for ease of export and the speed of interaction; it was run on both a
K. Appleton and A. Lovett 4
laptop (2.8GHz, 1Gb RAM, 64Mb video card) and desktop (2.8GHz, 1Gb RAM, 256Mb
video card) with no problems. Screenshots of the landscape model used can be seen in
Figure 2.
The landscape model was exported with a terrain resolution of approximately 2.5m and a
drape image resolution of around 1.5m. These settings are the primary determinant of the
resulting file sizes; the total archive, including elevation files, drape images, foliage
images, and 3D object files as well as the main model file, was approximately 56Mb. As an
example, halving the resolution of both the terrain and the drape image leads to a total size
of 30Mb.
2.3 Display Methods
Three display methods were compared:
• A set of six images, screenshots of the real-time model, laser printed in colour on
three sheets of A4 paper. The set comprised two high-level, two medium-level,
and two ground-level views.
• The NatureViewExpress model running on a 17” flat (TFT) monitor (via a
desktop PC) at a resolution of 1280 x 1024.
• The NatureViewExpress model running on a 1.5m Elumens Vision Station (via a
laptop PC) at a resolution of 1024 x 768, the maximum for the projector.
In the case of the NatureViewExpress models, the setup was identical for both display
methods. Control is via the mouse and keyboard, and participants were given oral and
written instructions on using both pre-set cameras and interactive navigation.
2.4 Surveys
Experimental sessions were run over four days; two of these were in the foyer area of the
School of Environmental Sciences at UEA, and two were in the Virtual Reality Suite
within the department. Advertising took place throughout the school via e-mail and posters,
and participants were able to pre-book or simply turn up. Seventy-one people participated
in the survey.
In order to keep participation time short, only eight responses were requested for each
display method, with a further four demographic questions (age, sex, familiarity with
computers, familiarity with computer games) and space for comments if desired. The main
questions were, for each display method:
• “To what extent does this model/viewer combination allow you to imagine the
real landscape?” (using a rating scale of 1-9, where 1 = not at all, and 9 = very
well).
• Six adjective pairs, also on a 1 to 9 scale (the adjective at the lower end of the
scale is named first in each pair): beautiful/ugly; cramped/spacious;
simple/complex; interesting/boring; static/dynamic; attractive/unattractive. The
adjectives were chosen from KASMAR’s (1970) lexicon.
Display Methods for Real-Time Landscape Models: An Initial Comparison 5
• The order in which the method was seen (first/second/third).
3 Results
Participants found the landscape easier to imagine using either of the two computer
methods than with the prints (p < 0.01, Kruskal-Wallis H test), but there was no significant
difference between the computer displays, each having a median rating of 7 compared to 4
for paper. Figure 3 shows the distribution of responses for each method. It can be seen that
the results for the visualisation dome are slightly more skewed to the high end of the scale
than those for the flat screen, but not significantly so. Both computer methods also have a
small tail of results at the low end, indicating a small number of viewers were dissatisfied
with the model shown.
Fig. 3 Participants’ ability to imagine the real landscape, split by display method
This result implies that the larger curved screen gave little or no benefit over a conventional
flat screen in terms of the viewer’s ability to understand the information presented and
picture the landscape in a holistic sense. In fact, participants’ comments indicated a variety
of opinions on the three methods, not all of which can be reflected in the simple rating
scale used. Some positively enjoyed the novelty of the visualisation dome, while others
appreciated the familiarity of the flat screen and paper images. A number reported a greater
sense of immersion or “being there” when using the dome, while for several others the size
of the screen and the closeness of the viewing position made the experience uncomfortable
and even led to feelings of motion sickness in some cases. The relatively low resolution of
the projector compared to the monitor was noted by some as a disadvantage, as was the
K. Appleton and A. Lovett 6
distortion of the projected image, and the fact that the central desk obscures part of the
display, including in this software the viewer orientation/location map. Some of these
problems could be addressed by technical changes such as a higher resolution projector,
and software changes including a different screen layout and incorporating spherical
geometry correction, but others are limitations of the equipment, with a great deal being
down to individual preferences. There is no doubt that an additional benefit of the dome is
generating interest in whatever you wish to present to people; while conducting the survey
in the foyer, many survey respondents were students and staff who had been intrigued by
the equipment while passing by.
Figure 4 shows the effects of display method on the six adjective pair ratings scales. As
with the participants’ ability to imagine the real landscape, it is the paper information
which gives a significantly different result in each case (Kruskal-Wallis H test, p < 0.01 for
all except attractive/unattractive, p < 0.05), with the two computer methods not differing
from each other sufficiently to be significant. In general, from viewing the printed
information, participants rated the landscape as uglier, more cramped, simpler, more
boring, more static, and less attractive. Perhaps not unexpectedly, the largest difference
between paper and electronic methods occurred on the static-dynamic scale, with the
computer models leading to the landscape being rated as far more dynamic, even though
there are no moving elements such as trees blowing in the wind, flowing water, or people
and vehicles travelling. Only the viewer’s location changes, although technology
increasingly offers the opportunity for animated elements within a scene, and this would be
an aspect worth revisiting.
Display Methods for Real-Time Landscape Models: An Initial Comparison 7
Fig. 4 Ratings on the six adjective scales, split by display method. Boxes show the mean
rating, and the bars indicate the 95% confidence interval above and below the mean.
Examination of participants’ responses with regard to the order in which they viewed the
three methods revealed no consistent effect, although the split between all six possible
viewing orders was not particularly even due to “drop-in” participants preferring to view
the methods as others finished rather than waiting to proceed in a specific order. Several
respondents made comparisons between the methods, but any noted good or bad points
were largely consistent across all possible viewing orders. The demographic and
background information collected was simple, and again no consistent pattern was
observed in terms of the impact on participants’ ratings of the three viewing methods.
4 Conclusions & Outlook
It is interesting that this survey, admittedly a relatively small one, indicates little difference
between a conventional flat-screen monitor and a moderately sized visualisation dome in
terms of communicating information to viewers about the nature and certain qualities of a
landscape. The equipment is expensive, with a visualisation dome of this size costing well
over ten times as much as a relatively high-specification PC, and it is logical to assume that
there must be certain expectations on the part of those who buy it about the effects of using
it. However, the research literature reveals no real qualitative information on the benefits of
using such equipment. While some people clearly do find this sort of display beneficial,
and there are undoubted advantages in terms of engaging people’s interest, sufficient
drawbacks were noted that care is advised in using a dome, at least in considering the setup
and the configuration of the landscape display. However, the results also suggest that a flat
screen could be set up as an alternative/additional display point with little impact on the
responses of viewers who saw one or the other.
The differences noted in Figure 4 suggest that caution is necessary if different groups
within a consultation exercise are presented with different information, for example if
meeting attendees can view a real-time landscape model while non-attendees are sent
printed images. Questions about the acceptability of a particular landscape future could
clearly be affected by the differences in perception, and the results in this paper indicate
that equivalence of information cannot be assumed across paper prints and real-time
information. In particular, the matter of a landscape’s perceived attractiveness is of great
importance in determining audience approval, although it is acknowledged that aesthetics is
just one of several aspects of any future landscape scenario which might be presented,
alongside biodiversity, natural resource and socio-economic considerations.
In this study, the objective was to compare display methods using identical information,
which is why the paper-based images were simply screenshots of the interactive model.
However, one previously-noted strength of non-interactive information is that greater detail
can be incorporated (APPLETON et al., 2002). An extension to this research would
therefore be to include more detailed visualisations of the area, either on paper or as still
K. Appleton and A. Lovett 8
images on a computer screen, and obtain ratings to determine any difference in preferences
and perceptions between more realistic images from a limited number of viewpoints, and
the ability to explore a more abstract model at will.
It must be said that the landscape model used here represents an area which is by nature
quite open and not very complex, nor would it be considered particularly attractive
according to commonly accepted theories of landscape aesthetics. An interesting further
step would be to assess a range of landscape types and see whether the effects described
above are consistent.
Finally, a very desirable addition would be to include further methods of communicating
landscape information. The department’s Virtual Reality laboratory contains a large (2 m ×
5 m) screen, and to use this would be an obvious extension, giving yet another sense of
scale. Another practical progression would be to actually visit the landscape in question, as
was done by RICE (2003). The distance to the study area in this case prevented such a visit,
but to compare visualisations of any kind to real life experiences is perhaps the most
important extension to this sort of work, and the one which will allow the effects of
different presentation methods to be most fully quantified.
5 References
Appleton, K. Lovett, A., Sünnenberg, G., Dockerty, T. (2002), Visualising Rural
Landscapes From GIS Databases: a comparison of approaches, options and problems.
Computers, Environment and Urban Systems, Vol. 26, No. 2-3, pp 141-162.
Kasik, D. J., Troy, J. J., Amorosi, S. R., Murray, M. O., Swamy, S. N. (2002), Evaluating
Graphics Displays for Complex 3D Models, Computer Graphics and Applications Vol.
22, No. 3 (May/June 2002), pp. 56-64
Kasmar, J. V. (1970), The development of a usable lexicon of environmental descriptors,
Environment and Behaviour, September 1970, pp. 153-169
Lovett, A., Sunnenberg, G., Appleton, K., Dockerty, T., Dolman, P.M., Cobb, D. &
O’Riordan, T. (2002) The use of VRML in landscape visualisation, in Buhmann, E.,
Nothhelfer, U., and Pietsch, M. (eds.) Trends in GIS and Virtualization in
Environmental Planning and Design. Heidelberg: Wichmann
Nelson, L., Cook, D., and Cruz-Neira, C. (1999), XGobi vs. the C2: Results of an
Experiment Comparing Data Visualization in a 3-D Immersive Virtual Reality
Environment with a 2-D Workstation Display, Computational Statistics Vol. 14, No. 1
(January 1999), pp. 39-51
Rice, A. (2003), Exploring the Impact of Emerging Landscape Visualisation Tools on
Spatial Perception and Design Education, in Buhmann, E., Ervin, S. (eds.), Trends in
Landscape Modeling, Heidelberg: Wichmann
Stock, C., Bishop, I. D. (2002), Immersive, Interactive Exploration of Changing
Landscapes, Proceedings, iEMSs 2002, pp30-35;
http://www.iemss.org/iemss2002/proceedings/pdf/volume%20uno/40_stock.pdf