It’sachallengetogetpeopletotakepartinnewandunusualbehaviours–like using the remote control to buy pizza, or taking part in a programme rather than sitting back couch-potato style. To succeed, just like normal television programmes, interactive television applications need to offer viewers something they want or need in a way that keeps their attention. This chapter focuses on two key factors that help shape viewers’ experience of interactive television services: the usability and the graphic design.

Usability is about whether it’s easy to use a device or interface for the intended purpose. An application with great usability works well, is understandable and leaves a sense of satisfaction. This requires an appreciation of the different types of viewers that are likely to use the service, what they want, how they live their lives and what will be easy for them to understand on-screen.

Graphic and interface design, for the purpose of this chapter, is about creating meanings beyond the basic functionality by manipulating the style of the graphics. A great graphic design helps make a service emotionally engaging and aesthetically pleasing, helps signal how an application should be used, and helps boosts viewers’ enjoyment. Think about the different reactions viewers would have to the same application designed in the style of a Japanese comic book, compared to the style of one of the early bank cash machines or the style of a kindergarten classroom. Each graphic design style would result in a completely different experience.

Usability and graphic design are symbiotic. An interactive television service designed purely for usability may allow viewers to perform tasks, but risks leaving them feeling disconnected, uninvolved and without a sense of allegiance. A service built only with graphic design in mind may look fantastic but be difficult and frustrating to use. The two must work together, not against each other.

How do you get the right mix? You wheedle your way inside viewers’ heads – and stare right back out at the television set. This involves the application of common sense, experience, observation, technical knowledge and research. This chapter cherry-picks a little bit from each of these: first by looking at fundamental usability and design considerations for all devices, then by examining the experience of watching and using a television, and finally by suggesting a practical checklist for producing graphics that will work well on television screens.

Fundamentals of Usability and Design

There are some fundamental principles of usability and graphic design that apply to all devices, whether television, mobile phone or personal computer – any medium where human beings interact with a machine. These are well worth considering during interactive television production.

Users

Most successful user-interface designs have a common thread: they are built, as much as possible, with reference to users’ needs and wants. For example, to help get the Palm Pilot personal organiser right, the inventor walked around for weeks with a small block of wood and a chopstick, pretending to enter information. Although the Palm Pilot wasn’t the first hand-held computer, it was one of the first really successful ones. It succeeded where others had failed because the inventor’s research helped focus the specification and design towards the activities people actually wanted to perform – like looking up addresses and noting down phone messages.

There are tools and techniques that can be used to help observe and learn from users or to put yourself in their shoes. Some are informal and subjective; some are more formal and scientific – some are expensive to employ, others simply involve grabbing someone who happens to walk past your desk. Some common usability testing tools and techniques are described in Table 5.1 and in the case studies at the end of this chapter from Serco Usability Services and BBC Research and Development.

A production approach called user-centred design actively engages the target users in the production and design process. It involves asking questions, observing, testing and developing iteratively, building a product based on people’s feedback, as well as producers’ insights. The user-centred design methodology recognises that even the most visionary and brilliant producer of interactive systems will benefit from understanding that little bit more about the users he or she is trying to target.

Speed of Response

There’s a classic piece of user research from the late 1960s, which is still relevant today. Robert Miller looked at the effects on users of different response times with computer systems (response time being the time it takes for the desired result to happen after the user has pressed a button or key). He found that, on average, there needs to be:

• Less than about a 0.1-second response time, for users to feel that the system is reacting instantaneously and to feel that they are directly manipulating events.

• Less than about a one-second response time, for users’ thought processes to remain uninterrupted. That is, the delay will be noticed but will not interfere with the feeling of working directly on the system.

• Less than about a ten-second response time, for the users to keep their attention on what is taking place. Anything longer and users are likely to do something else while they wait.

Response times are particularly relevant to interactive television, because screens often have to be downloaded or processed by the set-top box, and viewers have to wait for them to appear on-screen. Ideally, these response times should be kept to less than one second. Interactive television producers are in a position to reduce the response time of applications they are building by finding ways of reducing, among other things, the complexity of graphics and the amount of information that needs to be processed. Alternatively, content (an egg-timer graphic, for example) can be put on-screen to maintain or divert viewers’ interest during the wait (see the remote control section later in this chapter) or the service can include a picture-in-picture television channel, so people can watch this.

Response times over ten seconds are particularly dangerous in a television environment, because there are so many other attractions that could draw the viewer away – such as the pleasures of switching channel. Only the most compelling and valuable information will survive a response time greater than ten seconds.

Simplicity

Whatever the device, the simpler it is to use, the more people are likely to use it – as long as it still fulfils the core user needs. You don’t have to look far for evidence of this. Video Plus sold millions, because video recorders were far too complicated to understand. The early success of the Apple Macintosh was partly due to the graphical user-interface, which was much easier to understand than command-based interfaces. The internet search engine Google has become a multi-million pound operation, by offering a fast-download front page, simple interface and effective responses to user requests. With all these examples of simplicity, users can get what they want, quickly, easily and without having to make cognitive leaps to decipher the interface.

The need for simplicity is doubly the case with interactive television. The control of television is an eminently straightforward experience already – and viewers seem to like it that way. Most people use just a few buttons on the remote – channel up and down, volume up and down, and the number keys. These always work as expected and aren’t difficult to become familiar with. Similarly, television programmes tend to use on-screen graphics that are big, bold and to the point. And when voice-overs or on-screen presenters are used, they tend to speak clearly and keep the messages simple.

The challenge for interactive television is to work within this environment, where simplicity is the norm and where new ways of doing things may even be feared. A study from the United Kingdom Independent Television Commission in 2001 found that people see interactive television as more complex than computers and video recorders – even if, in reality, this is far from the case.

One good way to keep an interactive television application simple is to avoid extra features that the majority of people don’t necessarily want or need. Consider what the core functionality is and make sure it is consistently presented and available throughout the application. Also, spend time thinking about how to help make that functionality as easy as possible to use. If advanced functionality for a minority of users needs to be included, it’s worth finding ways of presenting it so that it doesn’t get in the way of the core purpose of the application.

The Medium is the Message

The information in this book is presented in a certain way, to make the best use of a paper-based medium. It is broken down into chapters, there are long blocks of text that readers can digest at their own pace and there are lots of nice colour pictures. By building on the strengths of the delivery medium like this, it’s possible to create an experience that is as engaging and as useful as possible.

To work well on a computer, Interactive Television Production would be delivered in a completely different way. The text would need to be broken into smaller chunks, to make it easier to read on a computer monitor. The illustrations would be a lower resolution, to make file transfer faster. And there could be on-demand interactivity, such as quizzes. To work well on a mobile phone, most of the content would probably need to be dropped. Reading lots of information on a small mobile-phone screen is tedious in the extreme. Interactive Television Production on a mobile phone might work best by concentrating on quick-access but useful interactive television production tips, perhaps sent to the phone each time the user arrives at the office in the morning.

Similarly, a service that has been designed for one medium then transferred directly to another is unlikely to be very successful. As discussed in Chapter 1, web sites that get plonked, unchanged, onto television screens are usually unmitigated disasters. They are strange and difficult to use on a television set. Most interactive television applications should be designed with the television experience in mind, not just moved across, unchanged, from other media.

Usability and Design Across Media and Platforms

The Television Experience

The television experience is very different from reading a book, sending a message on a mobile phone, or using a personal computer. Television itself necessarily affects the way interactive television services are designed. Unfortunately (or perhaps fortunately), there has not yet been a great deal of research carried out to determine the best way to design for interactive television. Current practices are often a mixture of opinion plus trial and error. The following are factors that are worth bearing in mind during design – not hard and fast rules.

Lean Forward Versus Lean Back

One distinction that is often referred to in relation to interactive television is between lean forward and lean back. Personal computers are thought to be lean-forward devices, because users sit near the screen and lean forward to actively engage in what they are doing. Television is thought to be lean back, because viewers sit back on the sofa to watch programmes – and expect everything to be delivered to them.

The premise is that people are in very different modes when leaning forward and leaning back, and that it is very difficult to persuade television viewers to sit up and lean forward to actively ‘use’ television. If any interactivity works at all, it will be interactivity that fits in with a very passive lean-back style.

It is the case that television is often less cognitively and physically demanding than using a personal computer or reading a newspaper. Televisions are frequently left on in the background, as just one of the many activities people are engaged in (vacuuming, reading, vigorous love-making and so on), and television is often used while viewers are in a very relaxed mode. Consequently, interactive television producers do need to recognise that it may well be more difficult to persuade television viewers to interact than it would be to persuade personal computer users. Nevertheless, the lean-forward, lean-back concept is, perhaps, an oversimplification. People already lean back and forward when watching regular television. Sports events and quiz shows, for example, are often far from lean-back experiences – with viewers shouting out comments or arguing with the rest of the family. And console games on television are played actively rather than merely watched. Furthermore, United Kingdom and European viewers are familiar with switching instantly into a lean-forward information-retrieval mode when they use analogue teletext.

What this means for producers of interactive television applications is that the level of cognitive and physical activity that can be expected from viewers is likely to be dependent on the type of content on offer and the mode the viewer is in. ‘Generalisation of mood experience when describing television consumption can be misleading’, says Andrew Lowe, managing director of interactive television strategy company EnhanceTV. ‘It is better to describe the lean-back and lean-forward postures around types of programming or activities, as opposed to a way that describes our posture for a particular content delivery mechanism.’ For example, a viewer absorbed in a romantic movie is only likely to be prepared to engage in enhanced television interactivity that has a very low level of cognitive and physical effort, at least until the end of the show, and the reward on offer for interacting would probably need to be very high. Conversely, a viewer watching a quiz show may well be prepared to pick up the remote and get stuck in to the enhanced television, without needing too much persuasion at all.

Within interactive applications themselves, different types of content are also likely to be used in different ways. Viewers can reasonably be expected to proactively engage with a game, for example, but may only have a fleeting interest in a shopping or information application. Time should be spent working out how the usability and graphic design will tie in with the content and objectives of a given application.

Distance from Screen

The distance between television viewers’ faces and the television screen is typically 2–3.5 metres (children tend to sit the closest). This means that television screens generally take up much less area on the retina than personal computer monitors, which are viewed from less than a metre away. Activities in the rest of the room will easily distract television viewers and content on-screen will have to work harder to get attention, compared to content on a personal computer monitor. There are several tricks that can be used to sustain viewers’ attention. These include:

• Using colour, sound, high-impact images and movement to attract and keep the viewers’ attention – without getting in the way of the key functionality. Sound, in particular, can add tremendous impact to the interactive television experience.

• Making sure the screen is not cluttered. There should be one area of focus and/or a clear order for the viewers’ eyes to follow. The number of elements on-screen should be kept to a minimum. Some producers try to have less than seven different elements on-screen at any time.

• Fonts and graphics should be clear and large. See the design checklist, later in this chapter.

• Written material should be broken up into small and easily digestible chunks, with plenty of spacing between lines of text.

Remote Control

Remotes control television; hence, interactive television applications need to be designed with the remote control in mind, not a mouse or joystick.

Navigation

Many television viewers will not have any understanding of the remote control beyond what the channel up and down and the number buttons generally do. It is illuminating (and to be recommended) to watch the subjects of user tests struggling with even the simplest interactive applications – unable, for example, to make the connection between arrow keys on the remote control and the movement of a selector box on-screen.

To work well, interactive television navigation should pass the ‘intra-ocular trauma test’ – that is, it needs to be so obvious that it hits viewers right between the eyes. There are a number of ways of ensuring this happens:

• If there is an on-screen selector that viewers need to move around, it should be very prominent – perhaps it changes the colour and size of the selected element compared with the rest of the screen.

• Short text instructions can be put on the screen, explaining what the viewers need to do. (Don’t expect viewers to navigate to a help section to find out this kind of information. They may not know how to select help unless that too is explained on-screen. And most people never make the effort to read help sections anyway – even if they need to.)

• Important parts of the on-screen navigation design should be logically positioned (in terms of the likely movement of the viewers’ eyes across the page). For example, many applications use on-screen arrows to help viewers understand that they need to press an arrow key on the remote to get more information on-screen. If the arrows are not positioned near the part of the screen the viewer will be looking at, just before they need to press the arrow key, they are less likely to make the cognitive link between the on-screen arrow and the fact that it will help them get more information.

  

  5.6  A purple on-screen selector. Courtesy of Carlton Active.

  

  5.7  A yellow box selector. © NTL.

• Abstract icons should be avoided. For example, viewers may not be able to interpret that a hand with a pointing finger means they should press an arrow on the remote for more information. It will help viewers make the connection between the screen and the remote control if on-screen icons mimic the designs and buttons used on the remote.

• Navigation using number keys can work well. Viewers press a number to go to a particular option. The model is familiar to people from television channel changing and analogue teletext. Ideally, the numbers on-screen should look very similar to the number buttons on the remote control, so that viewers can immediately make the connection between the two.

• Minimise the number of key presses required. Using a remote control for selection can be a reasonably strenuous activity (relative to using a mouse for pointing and clicking). ‘Instead of being a single action, pointing turns into a sequence of actions that have to be planned and monitored with a much larger degree of cognitive load than when using a mouse,’ said Jakob Nielsen in one of the first studies of WebTV usability (in 1997).

• Anticipate the viewer’s next selection when they are navigating between pages and pre-set the selector box or highlighter in that position for them.

• The range of finger movement required between different keys on the remote control should be limited, as should the number of keys in use. The more remote control keys that viewers are required to use, the more they will have to look away from the screen to check the remote and the more effort it will take. ‘We try to minimise the number of different keys used in navigation and stick to arrows and the select button,’ says Matthew Stratfold, manager of enhanced television at digital cable operator NTL. ‘In an ideal world, I’d even like to stop using the coloured keys as well, as the viewers have to keep looking back at the remote to use these, which can be time-consuming and annoying.’

• Remote controls can make it difficult for viewers to operate scrolling pages. Some viewers may also find this computer-based concept confusing. Pages that appear in series, one after the other, are likely to be more familiar to viewers, as this is the style of graphic presentation used in television programmes.

  

  5.8  The first paragraph of the news story on one screen. © Teletext Ltd.

  

  5.9  The second paragraph of the news story on the second screen. The whole page is updated for each section and there is no scrolling. © Teletext Ltd.

• Avoid personal computer or internet navigation devices, like drop-down menus, which viewers may not have encountered before and therefore will take more time to understand.

• Avoid having too many planes of navigation (for example, just up/down navigation options are likely to be easier to understand than on-screen options that require movement of the selector up/down, left/right and the pressing of various coloured fast-text keys on the remote).

Text Input

Inputting text is tedious with a remote control. Although some platforms provide keyboards, these usually get to just a small percentage of customers – and, even if every viewer has one, there’s no reason to assume they will want to use them. Would you enjoy juggling with a full size keyboard and your TV dinner?

Text input using a remote control, rather than a keyboard, is possible using letters etched onto the remote next to the number keys, in the style of a mobile phone, or using mini-keyboards that appear on-screen, in the style of arcade machines. Both are time-consuming and annoying. An interactive television application that requires lots of text input (a full name and address, for example) may find that the only people to use it are teenagers, who, from constant phone texting, seem to have the muscular thumbs suitable for marathon text sessions. If you absolutely have to get someone’s address, it’s best to ask only for the postcode, then look up the full address on a database. Then, even better, get the system to remember the full address for the next time the viewer uses the service.

Feedback

Remote controls are used at some distance from the television. Therefore, it’s helpful if there is some confirmation that instructions from the viewer are being received and understood. Ideally, there should be some kind of on-screen or other feedback (such as a click sound) within 0.1 seconds of buttons being pressed on the remote control, even if the main response to the key press is going to take a little longer. For example, if the viewer presses OK to select something on-screen and the next screen is going to take a second or two to appear, it’s worth considering making the selector box flash once in response to the OK press. This should help the viewer understand that the selection has been registered – rather than risk them pressing the button again during the delay.

If there is likely to be a significant delay in the response following a viewer action (anything much more than one second), this should certainly be signalled back to the viewer, for example, by putting a message on-screen that says, ‘Retrieving information now … please wait.’ If there is absolutely no way of getting the delay below ten seconds, then it’s worth considering putting something on-screen that will both distract and give an indication of the time left – perhaps an animated clock, a percentage figure counting up or designing the service so a television channel’s video and audio are constantly available to provide some level of entertainment during waiting periods.

It is sometimes tempting to implement a repeat function on remote control keys. With this, an action will be performed many times if a key is held down. This can help viewers perform some tasks faster. However, repeat keys can be risky with interactive television, since some viewers will not be familiar with the idea that holding keys down means the function will be repeated. If it is used, the repeat should take a little while to kick in and should repeat slowly. A fast repeat function combined with lack of on-screen signals for delayed responses can be a deadly combination – since viewers will hold keys down if it doesn’t look as if they are working. If key presses are still being registered while screens are updating, this could cause all sorts of mayhem.

Finally, when creating a navigation design, it’s worth considering providing an escape route if viewers get lost, bored or press the wrong button. This should be a button or option that will quickly take them back to somewhere they know – such as the main menu. A back button on the remote control can work well for this – but it needs to be implemented consistently across the platform, so that viewers learn that they can always press it repeatedly to back-track through what they’ve been doing.

Navigational Consistency Across Applications

Viewer Expectations

Television viewers have certain expectations about what a television is used for and how things should work on it. The Henley Centre, Planning for Consumer Change study (2000), looked at how consumers view different devices. The research showed that television is viewed primarily as a source of entertainment and information, while the internet is viewed more as a tool for communication. This is perhaps partly due to television’s history of being used in the living room, for relaxation and family activities, and the internet’s history of being used at the office and for email. It’s probably easier then for interactive television services to be designed to build on these existing expectations than fight against them. And there’s some evidence of this being the case: games are one of the most popular applications on United Kingdom interactive television platforms, while email hasn’t taken off. Similarly, some standalone interactive shopping services are struggling – perhaps because the weekly shopping task is not associated with television at all.

Furthermore, as discussed in Chapter 3, new interactive services may be easier for viewers to understand and use, and therefore be more popular, if there is a frame of reference from the viewers’ previous experience with television. Digital text services in the United Kingdom and Europe have the advantage of building upon existing viewer expectations of analogue teletext. Video-on-demand and personal video recorders have the advantage of building on video cassette expectations. Interactive shopping services linked to television channels, like QVC (as opposed to standalone shopping services that are not linked to television channels), have the advantage of tying in with an existing activity – that of watching a channel for information then buying products that look interesting.

This is not to say that out and out revolution isn’t possible with television, but when thinking about the usability and graphic design of an application, it is worth considering whether evolution is the easier route to take – by thinking about what viewers already know, then building on that.

Viewer Demographics

Television is used by a huge percentage of the population – many more people than use computers or mobile phones. This means the potential market for interactive television includes a wide variety of different people. To avoid excluding potential viewers, the design of most interactive television applications should take into account the variety of needs and abilities of an audience that could include senior citizens, children, the visually impaired, and so on.

One way of doing this is to offer viewers a choice at the beginning – for example, ‘If you want the easy-access service, press 1, if you want advanced functions, press 2.’ It also helps to choose interface styles that most people can relate to and use writing styles that everyone can understand. There’s not much point putting a message on-screen that says ‘Error #4309303 HTTP’, if you want to appeal to a broad market.

Even so, it is always worth bearing in mind that however useful interactive television becomes, some viewer groups may never be persuaded to use it at all. Also, applications aimed at particular niche audiences may not need to care too much about using language, designs or navigation models that exclude those outside their target group. They can be designed with that audience in mind. A pop music service, for example, is likely to want to use design styles that appeal to teenagers and will not worry too much about alienating those over the age of 40.

There’s one view on the different ways interactive television viewers can be categorised in the case study from Netpoll at the end of this chapter.

The Social Context

According to Statistical Research Inc (2000), 60 per cent of living room television viewing happens in a group. And it’s the living room where most televisions with digital interactive services are located (the older analogue sets, without interactivity, tend to get relegated to bedrooms and studies). Therefore, interactive television applications intended for use during group-viewing times should take this into account.

One way of building interactive applications that work in a social context is to make sure viewers can get what they need quickly and easily – during advertising breaks, between programmes and during boring bits. The Henley Centre (in 2001) found that 25 per cent of the users of Sky’s interactive service went there during television advertising. Enabling quick access means that the interactivity is less likely to annoy the rest of the people watching.

Another route is to make the interactivity work in such a way that it doesn’t interfere with the video programming. This is the main thrust of enhanced television. The less the interactivity gets in the way of the enjoyment of others in the room the better. This can be done by:

• Overlaying information on the video channel using semi-transparent graphics, rather than solid graphics.

• Using inserts, rather than taking up the whole screen. For example, if an enhanced television application has a function to allow viewers to select a camera that follows a particular sportsperson, this may be better done as a video insert into the main picture rather than taking up the whole screen, so other viewers don’t risk missing the main action.

• Always keeping a television channel on-screen, even when the viewer is using unrelated interactive content (this is usually achieved by running the TV channel in a picture-in-picture window).

• Providing multiple remote controls and designing the application so lots of people can use it at once. Two Way TV and Sky Digital sell multiple remote controls, so the whole family can play along with enhanced television games.

In this same vein, there are some interesting possibilities around providing interactive functionality to accompany the radio stations and audio streams that are available on digital television platforms (these are designed to be played via the television or using a stereo connected to the set-top box). Viewers tend to use these as background noise while they do something else, and the television interactivity can be designed to complement this fact. Music Choice in the UK already provides an interactive application that allows easy switching between its audio streams. Several UK radio stations are also looking at how interactive content can be used to support their output on digital television.

Time of Day

Television tends to be watched by more people in the evening. Therefore, unless specifically pitched at the kind of people that use television during the day, interactive applications will tend to get more usage in the evenings too. This means that it’s worth thinking about providing the best standard of interactive television service (for example, more updates and the biggest customer service team) during the evenings, at peak television viewing times – not during office hours.

Furthermore, in the morning and afternoon, television is more likely to be on while viewers are doing other things. To get round this, television programme makers will often try to structure programmes in such a way that dipping in and out is easy and enjoyable. Niche television channels use the same tricks – they are happy if they get a few minutes of people’s viewing time per week. Next time you see a specialist news channel, watch how often graphics come up explaining who is speaking – even if the person had been identified two minutes before. This is done because it’s difficult to know when viewers have switched on or started paying attention. You may also notice that morning magazine-style programmes often have much snappier items than the evening magazine-style programmes, and morning and daytime soaps have stories that are easy to pick up without viewers having to have watched the rest of the programme or series.

Interactive television applications can benefit from the same approach. To get viewers to use the service in the mornings and daytime, it can be worth thinking about the kind of interactive content that will provide satisfaction quickly. For example, an interactive television games service could consider running shorter games or ones that can be dipped in and out of during the day. An interactive television news service could promote quick updates and weather in the mornings – and promote more in-depth material at night. It may even be worth changing designs to reflect the different states of mind of people in the morning and the evening – in the same way that late-morning television chat shows are often very different in design to evening ones.

Design Checklist

As well as considering the television experience, the graphics themselves for interactive television services need to be adjusted to work on the television set – otherwise they can look really bad.

Set-top boxes and middleware systems make some adjustments to graphics and text. Unfortunately, these automatic processes can’t deal with every kind of problem, so it’s better to get screens looking right before they are sent to the set-top box for processing (doing this will help the application run faster too). Also, some viewers connect their televisions to the set-top box using poor-quality RF connections, rather than the better-quality connections like SCART and S-Video. This can exacerbate problems with interactive television graphics.

The following is a checklist for producing graphics that will work well on television screens.

Design for the Television Screen, not the Computer Monitor

Canvas

So, you’re sitting down in front of Photoshop ready to design your interactive television application. What size should your design be? It depends. A regular television screen has a 4:3 aspect ratio. The picture is built up with horizontal lines scanning across the screen – 625 lines with the United Kingdom and European PAL television system and 525 lines with the American NTSC-M system. Some lines are used for carrying signal synchronisation and other information, rather than picture information – see Chapter 2. Interactive television services are left with an available screen size of around 720 × 576 pixels for PAL and 640 × 480 pixels for NTSC-M televisions.

This is approximate, because different set-top boxes sometimes use different screen sizes.

Safe Area

Unfortunately, every television model is built slightly differently and some show more of the available picture area than others. Therefore, any important information needs to be shown away from the edges of the screen, to avoid disappearing completely on some televisions – in part of the picture known as the safe area. A background image can take up the whole screen, but any text or important graphics should be within the safe area. There’s some debate about what this safe area is, but most broadcasters advise keeping important information within an area around 20 per cent smaller than the full screen – that’s 10 per cent off each edge – although it might be possible to get away with a bit less if you don’t mind cutting off the edges of your graphics and text on some televisions.

Some middleware systems and platforms will set a specific screen size and force information into this area. Designs that are too big will be squeezed into this. They do this to make sure information is inside their view of what the safe area should be. Some platform operators also include their own navigation or other elements within the screen area, which will restrict the available space still further. Check with the platform operator for the exact available screen dimensions and the recommended safe area.

Pixel Shape

If the interactive television design is being carried out on a personal computer monitor, bear in mind that personal computer monitors have square pixels, while televisions have pixels that are slightly rectangular. This means that shapes designed on a personal computer monitor will look a little wider when finally displayed on television – a circle will become an oval. To get round this, graphic work in progress should be viewed on a television screen not a personal computer monitor. Or, failing that, the graphics should be squeezed a little on the horizontal axis before being sent to a set-top box (UK cable operator NTL recommends, for their platform, re-sizing to 94 per cent of the original width, after designs are finished on a personal computer monitor).

Additionally, it’s worth taking into account the fact that interactive television applications designed for 4:3 may be stretched further if they are viewed on widescreen (16:9) televisions – making the fonts and graphics appear fatter than they would on a 4:3 television. To get round this, some interactive television producers design their graphics and fonts in such a way that they look okay both on 4:3 and 16:9 – a compromise position.

Avoid Detail

The available television screen size in pixels doesn’t give a wholly accurate portrayal of the screen resolution. This is because television pictures are built using two interlaced fields (see Chapter 2). The effect this has is that images on television appear softer than they would on a personal computer monitor, because only half of the image is being updated with each pass of the electron beam. There’s no point using images with fine details then, as these will blur and television viewers won’t be able to see them properly.

Interlacing also means that any very thin horizontal lines are likely to flicker, as they will only be updated with every other pass. Middleware systems and televisions will often have some kind of filter to minimise this effect, although it’s still worth avoiding anyway.

Finally, some detailed designs can cause unwanted movement and shifts on-screen known as moiré patterns (one version of this is the effect seen on-screen when someone is wearing a checked shirt). To avoid this, don’t use patterns made up of lots of thin lines of light and dark next to each other.

Avoid Highly Saturated and Very Bright Colours

Saturation

Colours that are highly saturated (very rich and intense) don’t work well next to each other on television. The television set can’t always keep up with the transition between the two, and blurring or crawling occurs at the edges. There isn’t usually a problem in the middle of areas of colour, just at the edges.

The saturation level of graphics and text should ideally be dropped to less than 80 per cent. Red and yellow tend to be the worst offenders. Pastel colours work well on television, as do blues.

Brightness

Very bright (high in luminance) colours don’t look great on television – particularly if used as backgrounds. Pure white is the worst offender. This is because pure white is made by turning up the red, green and blue cathode rays to full blast. The intensity of electrons upsets the electrical balance of the screen and makes the beams go wide of the intended targets – creating wavy effects or curving, known as bloom, around the sides of the screen or around the edges of bright text or graphics. To deal with this cut down brightness levels below 90 per cent.

Television Colour Terminology

There can also be problems when there are different colours next to each other with very similar brightness levels. As with the bleeding caused by transitions between highly saturated colours, televisions have trouble keeping up with rapid changes between different colours with similar brightness levels. The end result is blurring between the two. The best way around this is to make sure colours next to each other have different brightness levels.

Note that the NTSC-M system displays colours brighter than PAL, as well as having a different screen size. Interactive graphics will need to be designed especially for each system.

Colour-Check Tools

Most middleware companies supply PC-based tools for checking combinations of colours for problems with saturation or brightness. These can be really useful, although they should ideally be used in combination with checking the work on a television set (preferably a mid-range television that is reasonably representative of what the viewing public have at home). Checking work on lots of different television sets is an even better idea.

Choose Fonts for Television

Most broadcasters use font sizes on graphics that are no less than one fifteenth of the screen size. The relative sizes of fonts depend on the middleware and font processing in the set-top box, but as a rule of thumb it’s rarely worth using font sizes less than 22 points – and bigger is better. Get a small portable television and see if you can comfortably read the text from three metres away. If you can’t – make the font bigger. Moreover:

• Sanserif fonts tend to look better on television than serif fonts. Serif and delicate fonts can break up around the edges.

• Anti-alias fonts as much as possible. Anti-aliasing adds transitional colours between high-contrast colours (like text and its background). This has the effect of smoothing out the edges of fonts, reducing boundary blur and making the font easier to read. Regular television caption generators traditionally have an incredible number of levels of anti-aliasing.

• Drop shadow (a slightly offset shadow of each character just underneath it) helps pull the text away from the background, reducing blur at the edges and making text easier to read.

• Upper and lower case words are better than all upper case, as the shapes of the letters and words are easier to identify.

• One of the easiest to read colour combinations for text on television is white on blue – as opposed to the black on white or grey that works well on personal computer monitors.

Unfortunately, there may not always be much flexibility with font choices. Some settop boxes support just one or two different fonts and these are available in a limited range of sizes. Furthermore, most set-top boxes don’t have the processing power to do much anti-aliasing on fonts, so they can look a little ragged compared to graphics used in television programmes.

A common font used by United Kingdom middleware systems and set-top boxes is Tiresias. The Royal National Institute for the Blind designed the font to be as easy as possible to read by partially sighted people – and perhaps because of that, it works very well for people with normal vision too.

Next: the future.

Case Study: The Usability of Electronic Programme Guides

Case Study: User Research for Interactive Television

Case study: Hands on TV – Interactive Television Consumer Research