14/07/2014 5 Applications of Text to Speech in Video Games

Gaming has never been bigger business. Ever-more elaborate and sophisticated titles are being released all the time. Cutting-edge technologies are not only being deployed in game development, but are being researched and produced directly as a result of that development, as we've seen with innovations like Oculus Rift. CereProc has worked with large games studios around the world to bring our technology to a number of exciting projects, and we wanted to share with you a few of the ways in which text to speech can be used both in development, and to enhance the gaming experience.

Placeholding

Video games get through vast quantities of recorded dialogue to go along with their expansive, immersive storylines and worlds. But developers can't just sit around and wait for all of this spoken material to be created, which creates a real problem when trying to judge timings in particular scenes. Game producers often use TTS before they get their hands on the actors' rendition of the script for an idea of how long phrases and sentences take to say, thus allowing them to block scenes and plan durations. CereProc's SDK also allows for the retrieval of visemes: these provide information about how a character's mouth should be animated as they move through various speech sounds.

Accessibility

TTS and accessibility go hand-in-hand. From something as simple as adding a voiceover option to menus, to providing an audio description of what is currently on the screen as you go through the game, there are tons of ways to improve video game accessibility using speech synthesis technology.

The costs of using TTS to make video games accessible to visually impaired gamers are small, especially when you consider the number of people who stand to benefit. In the UK alone, it is estimated that two million people are visually impaired at a level that cannot be fully corrected with glasses or contact lenses. Blind Adrenaline owner Che Martin has previously commented on this issue, saying: “There’s not enough money in it for the mainstream developers to make their games blind-accessible, so they don’t even worry about it.”* TTS provides a cost-effective way for developers to begin addressing this issue.

Localisation

After pouring all that time, effort and money into recording hours upon hours of dialogue for their game, having to translate it all for releases in other languages is a bitter pill for developers and publishers to swallow. Subtitling and dubbing - subs and dubs, if you're an industry kid – are the two main methods of localisation in video games, but both have their drawbacks. Subtitles compound the problems faced by visually impaired users. Dubbing gets around this, but can be prohibitively expensive. Text to speech is the best of both worlds, allowing publishers to localise their games efficiently and economically.

Dynamism

Voice talents for lead roles on big-budget video games do an inimitable job. We're not shy about blowing our own trumpet when it comes to the naturalness and characterfulness of our voices, but even we are prepared to admit that it isn't quite ready to put trained actors out of work. An approach that does work well is bringing together real speech and TTS to expand the horizons of game dialogue. The game studio already has a large body of recordings for at least one or two actors in the game; CereProc can make the most of that by turning those recordings into TTS versions of the actors' voices. In this way, dynamically produced content can be incorporated – for example the player's chosen name and attributes. NPCs in expansive open-world titles can be given unique things to say for greater realism, ridding the player of that spooky deja-vu feeling as they move from place to place. And modders can get even more creative with their expansions!

Chat

TTS used to 'enhance' in-game chat is kind of a love-hate thing. The sheer number of users online and using group chat features at any given time makes it pretty much impossible for synthesis to actually be useful, as so many people discovered recently on Twitch Plays Pokemon. But it can still be great fun! Check out this classic example of synthesized chat madness from Moonbase Alpha:

Which titles would you like to see add TTS? Which studios nail it on the accessibility front? Tweet us your views @cereproc

*Che Martin quote taken from a 2009 Gamestop interview.