Skin in the game: the use of sensing smart fabrics in tennis costume as a means of analyzing performance

Tennis costume in performance

In the late Nineteenth century, tennis decorum required men to wear a suit and tie (Fig. 2) and women a cumbersome dress with a corset and long sleeves. Figure 3 shows the extent of fabric carried by female tennis players at that time, including a high neck, long sleeves, draped apron and bustle that quickly caused the body to overheat. Contemporary fashions made it difficult to find a practical alternative to the restrictive clothes of the period; elegance took precedence over performance, much to the frustration of players unable to move as well as they might or showcase their skills effectively (Delanoë 2014). As a data space, these costumes function largely, via written and photographic record, as a statement on fashion trends and uses of apparel in different periods. The actual bodily performance within such a costume is largely surmised. In 1919, American Lawn Tennis advised women that a corset was “a requisite component of the correct togs for tennis” (Cassel 1919, p. 30). Cassel provides early evidence of a costume that supported and controlled the body but which hampered performance; contrast this, nearly one hundred years later, with twenty-first century compression fabrics that still act to support and control the body but this time in order to enhance physical performance. Fabric technologies, and in particular breathable fabrics, have slowly revolutionized sportswear (So 2015).

One of the dominant forces in the design of female tennis players’ costumes was Teddy Tinling. A tennis player himself at the age of 21, Tinling decided on dressmaking as a profession and set out to bring a touch of glamour and individuality to women’s tennis clothes. His drive to bring a sense of drama through costume took little account of the effects of fabric choice on sweat or physical performance and included a three-piece ensemble in black velvet for Rosie Casals and a shimmering opalescent cellophane dress for Billie Jean King’s “battle of the sexes” match against Bobby Riggs in 1973. King, however, never wore the now lost dress which she found too “scratchy and abrasive” next to her skin (Leibowitz 2003) and opted instead for a crystal studded back-up dress (Fig. 4).

Tinling’s designs were controversial and, in 1948, his dress for champion Betty Hilton’s Wightman Cup match “so outraged Hazel Wightman she threatened to ban colour—if not Tinling—from future Wimbledon games” (Blausen 2016, p. 4). Tinling correctly predicted that an all-white rule would prevail at Wimbledon and, in 1949, he designed a dress for Gussie Moran in white satin-trimmed rayon, which shimmered and included the now infamous lace trimmed shorts:⁴ “Tinling was accused of introducing sin and vulgarity to a gentleman’s game and was banned from Wimbledon for the next 33 years” (Blausen 2016, p. 5). Despite this his clothes were not banned, and they continued to bring a sense of flair and glamour to Centre Court performances. Blausen (2016, p. 6) notes that “Tinling had an easy rapport with the stars of the game. He designed to suit the playing style and personality of the players he came to know so well, matching fabric, trim, and cut to the individual”. By doing so Tinling found ways to subvert the all-white restrictions, which he felt led to costumes that lacked spectator appeal and did nothing to showcase a player’s individuality. From 1971 to 1978 Tinling was the official designer for the Virginia Slims circuit, where he would design some 100 dresses per season, each unique to the player (Blausen 2016).

The period following Tinling largely saw costumes become more uniform in cut and aesthetic and within the narrow design scope of commercial brands. Until very recently, costumes worn by leading players at Grand Slam events were dominated by the symbols of an exclusive group of brands, for instance the Nike swoosh, Adidas’ three stripes and Lacoste’s crocodile. In the men’s game, led by Djokovic (Uniqlo) and Andy Murray (Under Armour), there has been a shift towards brands that place an emphasis on fabric technologies and data. The founding principle of Under Armour was to “make athletes better,” and to achieve that the company uses its clothing to “approach human performance as a big data problem”. In so doing, Pierce (2016, p. 5) contends “Under Armour isn’t a clothing company, but a tech company. That’s where this entire industry is heading. The days of wristbands and straps are numbered”.⁵ As an artistic expression, costume enacts a dialogue between the body’s proportions and movement and the shape, volume and materiality of the fabric. However, with the current focus on a fabric’s computational and data collection capacity, the potential exists to disregard the importance of a costume’s theatrical effect in the quest to acquire big data.

The importance of a costume’s aesthetic performance is as key for many leading players as its technical sports capabilities; how the fabric and cut of the sportswear perform under conditions of movement and light, body-shape distortion of patterns and cut, and especially colour when chosen to contrast with the vibrant orange of clay courts or bright blue of hard courts. A costume’s aesthetic image is arguably more prominent for female players, though one should not forget Agassi’s denim shorts⁶ or Federer’s study in monogrammed elegance.⁷ Costume helps to shape the image projected by the player and to transform them in their performance; to stand out from the crowd of players wearing similarly-branded outfits. In this sense, as Tinling recognized, a theatrical costume becomes part of the player’s personality and helps him or her to manage and shape a performance persona.

In 2017, the shapes and expressions made by tennis players mean that any costume is put through demanding paces by the performing body as shown by Novak Djokovic in Figs. 5 and 6. Costume can take the brunt of the player’s frustration with their performance and this is illustrated by Fig. 6, with Djokovic provoking comparison with the Incredible Hulk (Parry 2015) yet being unable to rip his Uniqlo shirt.

Manzini et al. (1989) noted that the identity of materials can generally be defined both in terms of their physical and cultural performances. As smart fabrics facilitate the integration of electronics with textiles, there should be little reason why flamboyant and spectacular designs cannot also be of physiological calibre or feature a variety of thermochromic or LED effects that change in accordance with the body’s performance.⁸ Whilst smart fabrics allow us greater and more precise control over how our bodies perform sports skills, they similarly alter our understanding of materiality and bodily presence in terms of fashion more widely.

Fabric science and smart sportswear

The introduction of responsive materials and computing technology in textile structures offers an opportunity to develop smart sportswear with a new type of behaviour and functionality. Opportunities for smart clothing to drive research around use and utility arise as these textiles collect different kinds of data and use it in new and combined ways. For Berzowska (2007), smart fabrics are not only about combining textiles and electronics but about the conception of an entirely new medium which has to be tested, transformed and re-invented. The previous section of this article considers costume in performance at the surface of the body, where players respond to the feel and form of their costume and viewers to its visual properties and appeal. The remainder of the article posits costume as performance by exploring the interaction of fabric and technology in the ultimate generation of “living” fabrics and data spaces.

Smart textiles are possible through three developments: “(i) new types of textile fibres and structures such as conductive material; (ii) miniaturisation of electronics; and (iii) wireless tech that enables technology to be wearable and communicating at the same time” (Berglin 2013, p. 7). The five main functions of smart fabrics are: sensors, data processing, actuators, storage and communication. The extent of a fabric’s intelligence can be divided into three subgroups: “passive smart,” “active smart” and “very smart”. Passive smart fabrics can only sense the environment, active smart fabrics sense stimuli from the environment and react to them, and finally very smart fabrics adapt their behaviour to external circumstances (Das and Chowdhury 2014, p. 2).

The current wearable technology market consists mostly of sports and activity trackers (Rogers 2013). Nike and Apple introduced fitness tracking wearables in a 2006 collaboration to allow users to track their movements with iPods. In 2009 the first FitBit was released as a clip-on device that counted steps by means of an accelerometer (Paulson 2014). The first wave of “Fitness Trackers” including FitBit, Jawbone and Nike + FuelBand were essentially glorified pedometers. A second wave of devices appeared as fashion brands including Swarovski and Gucci teamed up with Intel, Google, and Samsung to create aesthetic and fashion-driven ways to wear data devices. Aside from aesthetics, devices also demonstrated technological advances, including “biosensing,” the collection of personal data related to health such as heart rate, hydration, and muscle fibre activation (Yang 2014).

In 2008, Adidas purchased Textronics^®, a team of experts in the fields of textiles and electronics to design sportswear using the latest fabric technologies. Their designs since have utilized soft textile sensors to capture the electrical activity or the mechanical movements of the body. Elastic yarns are used as “building block fibres” to weave or knit conductive or optical fabric structures and elastomeric polymers exhibit changes in electrical conductivity as the material is stretched. These polymers, with variable resistance properties, can behave as strain gauges, switches and sensors to track and capture a materiality of movement. Wallace et al. at the University of Wollongong developed a sports bra that changes its properties in response to breast movement; the fabric alters its elasticity in response to information about how much strain it is under. This smart bra is capable of instantly tightening and loosening its straps or stiffening cups when it detects excessive movement (Pitt 2014). The tennis player’s body can increasingly be considered a functional platform in sportswear design since it causes property change in the textile surface—referring not just to the live presence of a player within a match, but also to their role as an agent in activating the costume.

At present, smart shirts often still need some data storage capacity as the fabric in itself does not have computing power. Pieces of hardware are still necessary, but they are now available in miniaturized and flexible forms. One of the most important issues for sportswear is that the garment should be washable and the electrical components insulated to prevent water/detergent damage; thus, many designs opt for a removable black box that is only attached during the period of data capture. In one example, Canadian-based OMsignal worked with Ralph Lauren to create the Polo Tech shirt.

Ralph Lauren, designer of the uniforms for the Wimbledon umpires, line judges and ball boys/girls, has embraced the big data possibilities of fabric technology through the Polo Tech shirt (Fig. 7). This product is “a compression shirt with biometric sensors that act like an extremely sophisticated fitness band without having anything around your wrist” (Kooser 2014, p. 1). The nylon shirt is infused with conductive silver-coated fibres which act as sensors to track distance, calories burned, heart rate, stress rate and intensity of movement. The information is collected by a black-box data module and fed into an iOS app that then streams this data in real time to a smartphone. An athlete can adjust training by breathing more deeply, increasing exertion to hit a target heart rate, or focusing on reducing stress in competitive situations (Kooser 2014). The shirts were trialled by some of the ball boys at the 2014 US Open. The data from the shirt can be analysed using algorithms that pick out the key biometrics and psychometrics that the athlete and their coach choose to monitor (Collins 2014).

Although wearable technology has quickly infiltrated the fitness industry, the literature shows that there is a large gap where higher-level performance is concerned. The notion that wearable technology has the capacity to extend human capabilities, shape behaviour, and amplify social and behavioural responses (Joinson and Piwek 2013) invites investigation of the complex nature of micro-level motion capture in tennis and attracts brands such as Under Armour to view the big data potential in their textiles.

A number of companies, including Heddoko,⁹ Mico,¹⁰ and Cityzen Sciences,¹¹ are exploring ways to utilize a fabric’s intelligence to improve players’ abilities on the court. Others are pursuing the integration of technological components and capacities to the point of disappearance, where the technology becomes an undetectable part of the clothing save for a data exhaust streaming off the textile. Vigano observes, “we are in a convergence zone… Imagine if a plasticky glorified pedometer could be injected into what people decide to wear. We think wearable computing will become transparent and the devices will disappear to the eye” (Stables 2015, p. 5). The convergence between the miniaturisation of electronic components, intelligent textile production, advances in biotechnology and the growth of wireless and cloud computing locates wearable technology at the intersection of ubiquitous computing and functional clothing design.

Despite an extensive research effort over the last ten years only a few smart textiles are on the market. Whilst smart garments have clear use cases in sport, medicine and the military—and the origin of many smart fabrics came from military innovations—critical issues concerning the real need for smart textiles and the ethical issues of being monitored remain. Big data brings together large amounts of data and allows for the intersection of multiple types of data that previously would not have been considered together, the results of which may be utilized by insurance companies or advertisers. Health institutions are required to protect the privacy of health data but individuals are free to share their own data and post it publicly, and increasingly individuals are choosing so to do. Boundaries between public and private space are blurring and one of the major challenges of smart fabrics that generate this type of data is preserving individuals’ privacy, including how far data is captured beyond the fabric surface, the security of data use and storage.¹² Despite this, wearable technology is gaining traction alongside advances including the Internet of Things (IoT), 3D printing and Augmented Reality (AR) glass tech (Hololens, Magic Leap).

We are arguably at a pivotal moment for what is possible in terms of integrating data collection and communication power into our clothing and surrounding paraphernalia; with this comes an opportunity to reinvent how we think about the relationship between technology and our clothing/objects and their related data streams. The Quantified Self (QS) movement is a response to self-tracking behaviours, described as “an advanced user community of people who have begun to explore and experiment with novel uses for personal data” (Watson 2013, p. 1). Consumers have access to a growing number of gadgets designed to gather real-time signals from their bodies, convert this information into digital data and expose it to algorithms programmed to reveal insights and inform future behaviour. Kelly (2012), a QS community founder, contends that the current moment of self-quantification is merely an intermediary step toward “the future self”—one that Swan (2013, p. 95) describes as “spatially expanded with a broad suite of exosenses”. The body becomes a “knowable, calculable and administrable object” (Viseu and Suchman 2010, p. 162). Whilst this is an emerging trend for wearable technology in society, in sports smart fabrics are already materialising the data body for scrutiny. Wolf (2010) proposes that such data serves as a new kind of “digital mirror” in which to see and learn new things about ourselves.

Designed by Sabine Seymour, SoftSpot is a “plug + play” sensor system for clothing that monitors biometric and environmental data and automatically connects with the Internet of Things: “its proprietary technology leads the way from wearables to ‘disappearables’. It is invisible, soft, washable, flexible, and wireless” (Seymour 2015, p. 3). Seymour (2010, p. 13) describes “fashionable wearables” as the intermediary “between the human body and the spaces we navigate… our clothing, accessories and jewellery are the epidermal interfaces with which we can experience the world”. As fabrics assume “epidermal properties”—either as a material that mimics the functions of human skin or as the intimate layer between a body and its environment—there is potential for intelligent fabrics to act as an epidermal portal between flesh and data performance spaces. In this case a tennis player’s costume would act simultaneously as the layer between his/her body and the world and as the intermediary between his/her body and a remote data world, as a portal to a parallel performance dimension. Smart costume captures the time-matter which constitutes a performance, and yet because the costume is also the process of generating the data space from biometric and biomechanical impulses it frames the body in terms of its “sensorial architecture” (Tomas 1989).

The performativity of human and fabric skin has led to a keen interest in developing energy-related fabrics. This new class of materials focuses on circulation, muscle recovery and blood flow to enhance energy and wellness. Labelling itself “a different kind of performance technology”, Hologenix’s Celliant consists of “minerals embedded in a synthetic polymer that interact with the body’s electromagnetic emissions to induce increased oxygenation and blood flow… the technology modifies visible and infrared light, recycling them into energy that the body can use more effectively” (Schwarz 2011, p. 3). Powered by the body’s own metabolism, Celliant claims to be “more hybrid engine than textile. It recycles and converts radiant body heat into something that gives the body a measurable boost—infrared energy” (“What is Celliant?” 2016, p. 1). As fabric science explores how costume can measure heart rates and respond to temperature, capture the nuanced movement of muscle and also utilize the power of the wearer’s own metabolism to generate energy, one has to consider the extent to which our costumes are becoming a second skin.

Skin in the game?

Skin-tight fabrics such as elastic and Lycra-spandex have long been integral to the costumes of athletes engaged in activities ranging from swimming and running to speed skating and skiing. For the most effective data capture the fabric needs to fit as flexibly and comfortably close to the skin as possible, mirroring the body’s own contours. Stretchable sensors combine electronic components, energy supply and actuators on a stretchable substrate with stretchable conductors. To assimilate flexible printed circuit boards with fabric, the components need to be arranged to avoid body “flex zones” where excessive bending might damage sensitive components, yet in tennis it is these areas with the greatest variation of bend or movement that generate the most valuable data. A method for direct screen-printing of biological sensors onto clothing has been demonstrated by the University of California San Diego’s Laboratory for Nanobioelectronics. By printing the sensors onto an elastic fabric they are able to maintain tight contact with the skin (Yang et al. 2010); this also means that the shape of the costume evolves as the player gains and loses muscle.

Although the advantages of tight compression fabrics for physical performance have been well established of late (Marqués-Jiméneza et al. 2016), for women in particular there are still aesthetic concerns about revealing such a fit. Despite players such as Serena Williams wearing body-hugging catsuits,¹³ women are more often criticised for body shape and weight in ways that leave them potentially more vulnerable in public performance situations to hecklers, social media and Press comments.¹⁴ Increasingly we are seeing players wearing compression and skin-tight garments underneath looser items, such as men wearing shorts under shorts. This presents a situation in which the data capture element is separated from the aesthetics of a public presence.

The Tesla VR Suit uses tiny electric pulses which stimulate the skin surface, muscles and nerve endings to create tactile feedback and to simulate diverse sensory experiences. So, whilst smart fabric in sportswear captures data from the skin and transmits it to a data space, the Tesla Suit also operates in reverse to feed information from a data space to the skin. There is potential for sportswear to adapt this approach and assist training by feeding remote pressure data back to the skin and, in particular, to support resistance training of subtle muscle combinations. Integration of wearable electronics introduces new challenges for thermal management by adding the heat produced by certain electronic components to that produced by the body. In addition, plastic or resin components forming an impermeable barrier too close to the body can hold in heat and moisture hereby reducing comfort.

As an expansion of the concept of a second-skin, MIT’s Tangible Media Group is experimenting with a more literal “bio-skin,” a “living material” that responds to perspiration and body heat (Stinson 2015). As the person wearing the bio-skin becomes warm and begins to sweat, the material peels away to reveal breathable holes in the clothing. One of the researchers, Chin-Yi Cheng explained: “we are trying to create an interactive feedback loop between the human body, biofilm and the environment” (McGoogan 2015). The bio-skin fabric is made from natto cells, which expand and contract in response to atmospheric moisture. This material is traditionally used in a soybean-based breakfast dish in Japan and more recently has been tested in melt-resistant ice cream. The MIT team, led by Hiroshi Ishii and Lining Yao, grew natto cells and bio-printed them into scale-like shapes (see Fig. 8a, b).

Aligning the cells in a specific pattern, the team could programme them to behave in a certain way—in this instance to curl open when heated (McGoogan 2015); in addition to a cooling system, this fabric also presents the player with the opportunity to design for deliberate patterns or even sponsor logos, and, of course, has possibilities for fashion away from sportswear—especially theatre where stage lighting has the potential to generate a similar body heat effect. As a source of data, biosensors and GPS technology could capture motion data via the position and micro-condition of these cells/fabric shapes at any one moment.

Barad argues that “matter is not produced and productive, generated and generative. Matter is agentive, not a fixed essence or product of things” (Barad 2007, p. 137). Costumes are “not mere static arrangements in the world, but rather dynamic reconfigurings of the world” (Barad 2003, p. 816). Although my focus is on tennis, there are clear possibilities for advances in smart sports fabrics to inform studies of stage and theatrical as well as couture fashion designs, and vice versa, because wearable technology interfaces increasingly encourage us to use costume to manage and interrogate information and our actions in the world in new ways.

Sportswear as a data space

Although the performance of players wearing smart shirts is predominantly watched on court, it can also be experienced in a virtual space and abstract form by downloading data from the costume. According to Whitlock (2006, p. 85), “in an age of mediated bodies and avatars, the potential of motion capture in performance suggests a new method of character creation as well as new possibilities for recording and re-using human motion”. In recent decades, film studios have identified the potential of costume as a tool to translate human movement into data that can then shape the movement of a CGI character. Motion capture costumes for films are typically full body suits with reflective markers at strategic points on the body to plot movement through space. Despite the prevalence of this technology, Pingali et al. (2001, p. 76) observe that, “although there is some effort towards placing active and passive sensors on the players and objects to facilitate real-time motion capture, most sports do not allow such interference”. Sheets et al. (2011) note that quantitative kinematic measurements in realistic environments are limited by current motion capture technologies. Increasingly, designers are exploring the potential of integrating these capture techniques into fabrics worn by athletes. A smart sports shirt enables a more holistic capture of not just external points moving through space but also the body’s internal movement through muscle, respiration and temperature cycles. The capabilities of such a costume allow the body’s performance to be transformed into a cloud of numerical data.

Raw data from the costume builds the performance as a virtual interactive space, open to analysis by player and coach: “once a sporting event is stored in a database in the form of motion trajectories, scores and other domain specific information, a viewer can explore and interact with the virtual version of the real event” (Pingali et al. 2001, p. 77). Hence, the conception of an exo-self and a time-series self as a new entity complicating the unified human body where, for Schüll (2016, p. 26), the “body” is “a data-generating device that must be coupled to data-monitoring systems; together they inform a new episteme that devotees find empowering”. It becomes possible to compare a player’s style and strategy in one match with his/her performance in other matches; to create situations in which bodies are connected and extended in space as well as in relation to one another. The collected data highlights an indexicality of the costume’s relationship to that which it records, but also a shift from this data being primarily the residue of the flow of the live performance to instead being conceived as some kind of ontological force in its own right. In this way, the garment becomes its own event, carrying its own sense of mobility and temporality.

Data captured via a player’s costume manifests in a variety of visualizations, from animated 3D characters to line graphs, plot points and charts. Sensors track more than 100 metrics such as distance, speed, acceleration, deceleration, and heart rate. They also assess spatial changes in direction using 3D accelerometers, 3D magnetometers and 3D gyroscopes. For Hannah (2014, p. 20), “within the dilating layers of the body-object-event as objectile, the costume can be interpreted as matter-in-action and action that matters”. Figure 9 shows some of the outputs from the Ralph Lauren Polo Tech shirt (see Fig. 7), while Fig. 10 demonstrates dashboards from Sony’s Smart Tennis Sensor, a device which fits onto the end of the racket handle. Whilst data capture from intelligent fabrics is still in test mode, this example of performance information captured via the player’s racket highlights the variety of data visualizations which could be built into costume capture data space and which focus specifically on tennis performance rather than generic training platforms. According to Harris (2013, p. 242), “as design practitioners challenge perceived notions of textile, substrate, and skin through relatively newfound digital contexts, resulting propositions include high-tech interfaces for communication, surface embellishment, and guise”. Furthermore, Birringer and Danjoux describe the telepresence possibilities for costume that converts the dynamics of moving design “through algorithmic mappings of the kinetic data into a virtual format where data can be varied infinitely and transmitted between geographically distant wearers” (Birringer and Danjoux 2009, p. 102).

Data capture allows the player to embrace his/her skills in trend and plotted progression. Once captured, the data can be integrated and re-fleshed to fit with bodily performance and training offline. Manning (2012, p. 129) suggests that “it is not movement become form, but movement unforming… the creation of a strange interval through which image and body begin to intertwine” and where costumes are not things, but phenomena—“dynamic topological reconfiguring/entanglements/relationalities/(re)articulations” (Barad 2003, p. 818). Within the data space we might find various single tennis actions de-fleshed—translated into lines, graphs, percentages and ratios by a fabric skin that contributes to the capture of a de-fleshed performance. Via their costumes, players are able to monitor their heart rate during play, measure the resting and peak rates and recovery time, measure calories burnt and track data from multiple performances in comparison. Individuals can map the total number and placement of steps taken during a point or the whole match and this detailed analysis has a tangible influence on how players think about their performance, mentally mapping the data back into performed actions. This example also raises questions about the ways in which motions are read differently when seen on a human form as opposed to data charts on screen. Further, with the rapid advances in the representation of the human body in CGI and the games industry, more sophisticated models for displaying sports data may soon be available.

Developers argue that those invested in competitive sports, including engineers, athletes, and stakeholders, will see a revolution in the industry where players’ performances, coaches’ strategy, and the resulting fans’ viewing experience will be enhanced through ubiquitous computing (Springer 2014). The costume and its associated functions can help players to see their bodies through a scientific lens, to see the body as a system, and to think of it in terms of function. It also provides players and coaches with a language for thinking about bodily structures and systems. Thus, as much as fabric science is about possibilities for researching the costume, the costume itself becomes a tool for researching the body beyond as smart fabrics and wearable technology initiate what Balsamo (1995, p. 215) calls “new technologies of corporeality”.