One of the biggest complaints I hear about audio in video games is the amount of homophobic and racist language used in voice chat during online game play.  As a Game Audio Designer, this is an issue that I want to fix.  We Audio Designers put our hearts and souls into developing rich audio experiences and so, we need to avoid creating situations where gamers feel compelled to turn down the volume.

A recent post in the blog of our local weekly newspaper The Stranger, suggested that we [game audio professionals] should “…get around to magically filtering the system’s voice chat,” so as to inhibit the sort of trash talking that goes on.

How can this be addressed?

A few options are available, muting offensive players, hosting “no-foul-language” or “foul language” rooms, compartmentalizing voice communication, enabling community moderation, using speech recognition to censor and having verbal abuse as a feature.  Each of these solutions comes with it’s own challenges.

1. Manually muting offensive players:

Giving each player the ability to selectively ignore other players makes it so that the annoying player only communicates with people that actually want to listen.  Selecting players to ignore could become tedious if there are too many annoying players.  Also, muting inherently reduces the potential amount of useful gamestate information that can be relayed.  For example, if you mute a teammate, you won’t be able to hear him when he tries to warn you about someone creeping up behind you.

2. Foul Language and No Foul Language Rooms:

Hosting a “no-foul-language” room creates a type of exclusivity that separates gamers from playing with one another.  Creating a “foul language” specific room might be undesirable by game companies wishing to avoid the perception of being complicit in facilitating morally offensive behavior.  Nonetheless, these are useful ways to manage expectations.

3. Compartmentalize players:

Reducing the number of players that can communicate with each other at any given point in time makes it less likely that an offending player will end up in your chat space although it doesn’t guarantee that you won’t hear any foul language.  Many games compartmentalize by making it so that players may only communicate with their own team.

4. Community moderation:

Community moderation puts the policing of behavior in the hands of the gaming community rather than the publisher.  Ebay has buyer and seller reviews, Linkedin has recommendations, Craigslist allows people to flag postings, anyone may edit Wikipedia and Xbox live has gamer reviews.

Community moderation can either be active or passive.  Active community moderation has tangible consequences.  These could be temporary removal of voice chat or banishment from rooms that require a certain percentage of positive reviews.  This could result in a lot of complaints from players who feel unfairly restricted, or could lead to abuse of rating systems by groups of people targeting a single individual for the purpose of hindering their gaming experience.

Passive community moderation is merely a review or ranking in a gamer card to establish reputation.   This would be useful to people who wanted to research their teammates and decide whether to play with them or not based off of the opinion of others.

5. Speech Rejection: “Magical filtering” using Speech Recognition

Speech recognition and voice communication have been in games for a while. Back in the day, SOCOM on PS2 had speech recognition enabling AI characters to respond to spoken commands. It would be a relatively simple process to add a delay and analyze phrases before broadcast and a keyword based volume control to that system. The problem is that this system would make using voice communication useless because of the lag.

Speech recognition software capable of running on a game console generally relies on compliance of the player to pronounce keywords the same way each time.  People would find ways around these roadblocks by changing the pitch or tempo of their voice, publishing the banned words and using different ones.

Additionally, all homophones to potentially foul language could get banned. The censors would need to ask themselves if words like cockeyed, titmouse and uranus should be included.  If this happened, the value of voice communication systems would be worsened and we might see Lenny Bruce styled protests and complaints about censorship, freedom of speech, etc…

6. Embracing verbal abuse as a regulated feature

According to Wikipedia:

In Monkey Island, Insult Swordfighting consists of a series of Call and Response exchanges, in which an insult must be countered with a witty retort. Should the responder counter with an appropriate retort, they win the right to call the next insult; fail to respond, and the caller gains an advantage. Win enough of these exchanges, and the duel is won.

A well known insult from The Secret of Monkey Island, in which the insults were written by author Orson Scott Card, is “You fight like a dairy farmer!” to which the correct response would be “How appropriate. You fight like a cow!”

While this is a very clever way to encourage more creative and family friendly banter, selective playback of pre-recorded dialog isn’t a full substitute for real time voice communication.

In conclusion:

All gamers have a common interest in participating in virtual spaces. While none of these solutions will single-handedly give the best experience to all gamers, combining them in the right ways can make the majority of gamers quite happy.

Processing power and memory budgets are best spent on creating fun rather than limitations.  To achieve success, game developers should focus on building environments that encourage participation by as many people as possible.

For those who are completely intolerant of the language on XBox live or any other gaming chat rooms, we can always take the headphones off and play with our friends.

– Adam Smith-Kipnis