Kim Halskov, Aron Fischel
Media architecture, a subfield of interactive surfaces and architecture, is the term for digital technologies at large scale integrated into the built environment. This field emerged from the common practice of mounting large displays on the facade of a building, also known as urban screens. Iconic media architecture installations include Blinkenlights, which transformed Haus des Lehrers at Alexanderplatz in Berlin into an interactive installation by turning each of the 18x8 windows of the high-rise building into an interactive display, where people could play the game of pong using their mobile phone or upload low-resolution images and animations. A remarkable case is Jean Nouvel’s Institut du Monde Arabe in Paris, where each window is a mechanical device similar to a camera shutter that can open and close, allowing more or less light and fresh air to enter the building while at the same time acting as pixels of the facade. Another example is the Danish Pavilion at the 2010 EXPO in Shanghai, a building with a 300-meter-long curvilinear facade, where 3,500 ventilation holes are turned into pixels forming a three-dimensional and extremely low-resolution display with an aspect ratio of 25:1. As summarized in Table 1, media architecture differs from conventional interfaces in several aspects .
|Table 1. Comparison of media architecture with conventional displays.|
Clearly, media architecture is a unique subfield of urban computing that has gained popularity since the first Media Architecture Biennale was held in 2010. Numerous catalogs and books (e.g., ) have been compiled, providing an overview of this highly innovative field of interactive systems. Though systematic, such overviews tend to focus on the individual media architecture installation or categories of these (e.g., ), and do not provide detailed insight into how an individual case positions itself in relation to the field. Building on the idea that a design space may be conceived as the collection of all potential design ideas considered at a particular point in time during the design process , we suggest that a collection of final design objects may be organized according to which design space they populate.
Among the several conceptualizations of what constitutes a design space, a productive one is to understand a design space as a conceptual space defined by a set of aspects, each with a number of options . Based on this definition, a design space may be represented as a table visualizing the aspect-options sets. Table 2 illustrates the basic idea in the case of a sub-area of a media architecture design space. For each aspect (location, interaction style, content, and format) in the top row, the alternatives considered are listed as options below the aspect. For instance, the options gesture, touch, movements, and input device are considered as non-exclusive alternatives for the choice of the interaction style aspect for the installation.
|Table 2. A design-space schema showing an example design space for a media architecture installation, based on .|
The design space analysis method (DSAM) is a systematic approach to studying the design space populated by a collection of design objects. The method consists of five steps (Table 3).
|Table 3. The five steps of the design space analysis method.|
The first step primarily consists of delineating the collection of design objects to be analyzed and ensuring that data about the design objects (for instance, images, video, or Web documentation) or the design objects themselves are available. For this case study, the delineation is straightforward: We have selected all media architecture cases nominated for awards at the Media Architecture Biennale (MAB) since 2012, all presented at the conference websites. The fact that our design objects are well documented at the MAB sites makes step 2 straightforward too: recording data from the empirical material. Step 3 is much more demanding. In this step, the aspect-option sets of the design space are decided upon in order to develop the coding schema, which is applied in the coding of each of the design objects—that is, for each aspect, which options are relevant for the particular design object? Step 3 is an iterative process, moving back and forth developing the coding schema based on (data about) the objects and applying the coding schema to each of the individual design objects. Step 4 is where all the hard work in step 3 starts to pay off. This is where patterns in the design space may be identified, for instance, that the choice of an option with respect to one aspect often correlates with the choice of a specific option with respect to another aspect. Or you may observe that a certain area of the design space is only sparsely populated. The final step, step 5, is not only about communicating your finding to others, but also about refining your insight from the process as a whole. To facilitate the process, we have developed a design space analysis tool, which supports steps 3 and 4 of the method.
To illustrate the approach, we will now look into a study of the media architecture installations nominated for awards at the Media Architecture Biennales in 2012, 2014, and 2016, and focus on the visual interface subspace of the design space, including pixel-related aspects. The corpus of media architecture consists of a total of 31 installations. Data about the installations was collected from the MAB12, MAB14, and MAB16 websites, as well as websites documenting the individual installations. The final coding schema consists of 19 aspects organized into five overall categories: architectural, interaction, visual interface, auditory interface, and location. The visual interface category consists of five aspects: content, together with four pixel-related aspects: shape, pattern, spacing, and color range. Associated with each aspect are between three and six options. At the end of the article, we will look deeper into the visual aspects of the design space.
The 31 installations were coded based on the coding schema. The aspects and associated options in the coding schema themselves present interesting insights into the design space. One distinctive aspect is the physical relationship to the architecture (the relation of the digital media to the architecture). The distinction here is basically between three options: inseparable, integrated, and attached. Inseparable is the sense that the media (typically in the sense of some kind of display) is the architecture. Integrated refers to media that have been integrated physically into existing architecture, but not in such a way as to be inseparable. An example would be covering an existing structure with a media architecture “skin” that could conceivably be removed. An example of attached is Le Circuit de Bachelard, where translucent tubes containing LEDs are mounted on the wall of a tunnel connecting two buildings.
Browsing the different aspect-option sets and combinations thereof provides insight into the design space populated by the 31 media architecture installations. Among the most striking findings is the fact that three installations used sound (music and soundscapes), whereas media architecture installations are predominantly only visual, and that all three of these media architectures are temporary as opposed to permanent.
The analysis of the media architecture installations has disclosed four aspects related to pixels: shape and pattern, how the pixels are arranged to form a display (Figures 1 and 2); spacing, the distance between the individual pixels; and color range. Table 4 shows those four aspects together with the associated options. The four pixel aspects have decisive implications for which kinds of content may be displayed; hence, we also have included the content aspect of the design space.
|Figure 1. Five pixel shape options. Left to right: square, rectangular, circular, 3D polygonal, 3D spherical.|
|Figure 2. Five options for pixel patterns. Left to right: rectangular matrix; rectangular matrix, 45-degree angle; tessellated matrix; segmented vertical lines; segmented horizontal lines.|
|Table 4. The design space of pixels.|
Three media architecture installations illustrate the uniqueness of pixel subspace of the media architecture design space:
The Danish Pavilion at Expo 2010 in Shanghai by BIG Architects has a double-loop structure with an almost 300-meter-long facade, appearing from some angles as two bands, one above the other  (Figure 3). The facade of the pavilion is perforated with approximately 3,600 holes of various sizes, which serve the purpose of allowing sunlight and air into the interior of the building. Inside each hole, a translucent PVC tube is mounted together with an LED fixture, turning each hole into a pixel. When unfolded into a two-dimensional surface, the facade of the pavilion yielded a 300-meter-long, 12-meter-high structure, giving an extreme aspect ratio of 25:1. Basically the pixel shape is round, but due to the tubes the shape of pixels actually depends on viewing angle (Figure 4). The pixels are organized in a rectangular matrix configuration with a 45-degree angle, and with a varied pixel spacing larger than the pixel size. The color range is limited to RGB colors. The unconventional shape, pattern, and spacing of pixels have a decisive implication for the content, which is constrained to abstract graphics and video . The media architecture was developed by Martin Professional and CAVI in collaboration with Bjarke Ingels Group (BIG).
|Figure 3. Danish Pavilion at Expo 2010 in Shanghai.|
|Figure 4. Shape of pixels depending on viewing angle.|
Crystal Mesh (Figure 5) is an integrated part of the Ilumina building in Singapore, developed by realities:united studio for art and architecture, Berlin. The 5,000-square-meter facade is covered by 3,000 polycarbonate modules, of which 1,900 contain a light source. The shape of the pixels is 3D polygonal, organized into a tessellated matrix pattern with a varied pixel spacing less than the pixel size. The color range is limited to monochrome (Figure 6). The unique visual interface allows for display of abstract graphics.
|Figure 5. Crystal Mesh.|
|Figure 6. Tessellated matrix pattern.|
Spine, designed by the Danish design agency Kollision, is an interactive installation consisting of 20 glowing cubes (1 meter on each side) mounted in the ceiling of an old warehouse building and forming a 50-meter-long spine (Figure 7). Each cube is connected to a computer-controlled motor, which together with motion sensors enables the light and fluid movement of the cubes to respond to the movement of people in the space. The shape of the pixels is also in this case 3D polygonal, but organized as a dynamic matrix pattern with pixel spacing larger than the pixel size. The unique visual interface with only a monochrome color range allows for the display of extremely abstract graphics.
|Figure 7. Spine.|
As illustrated by the three cases and the design space of pixels (Table 4), media architecture is a highly specialized subfield of interactive surfaces and architecture. Media architecture displays are unique with respect to the five aspects of the pixel design space: shape, pattern, spacing, color range, and content. For instance, pixel shape may be square, rectangular, circular, 3D polygonal, and 3D spherical. Whereas the pixel shape and configuration of conventional displays are hardly noticeable, shape and configuration are part of the visual quality of media architecture. The mapping of the design space of a collection of media architectures may potentially inspire designers and researchers. What other shapes may pixels have, and how may they be organized in space?
This research has been supported by Innovation Fund Denmark (1311-00001B, CIBIS).
Kim Halskov is professor of interaction design at Aarhus University. He is head of CAVI (www.cavi.au.dk) and co-director of the Centre for Digital Creativity (www.digitalcreativity.au.dk). His research areas include innovation processes, design processes, media architecture, and experience design. In 2007, Halskov established the media architecture research group at Aarhus University. firstname.lastname@example.org
Aron Fischel is a research assistant at the Centre for Digital Creativity, Aarhus University, where his research focus is design space analysis, methods, and tools. email@example.com
Copyright held by authors. Publication rights licensed to ACM.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2019 ACM, Inc.