I showed ‘Lifeline’ during Thursday’s (9/20/2007) class of Survey of Interactive Media (CTCS 505), and I said that it was the first commercial use of Voice Control I had seen and Steve Anderson, Professor of said class, asked what came after? I paused and said ‘Nothing.’ Nothing I know of used voice control after Lifeline (2004). Why was this?
Was it because of the shoddy story? Was it because the voice recognition wasn’t 100% accurate (it was pretty good though)? Was it because the interface was too ‘new’ to the consumer? Was it because you needed to have a USB headset (not included)? It was actually all of these combining into the pivotal reason: it didn’t sell enough (it sold competently in Japan, apparently [according to Wikipedia], but from what I remember, not in the US). This is the problem with industry. Well, one of them. First off, they are afraid to do innovative things because they don’t know if they will sell or not. The standard fare has been proven to sell and is therefore ‘safe’, while innovation is unknown territory and ‘scary’ to the publisher. Secondly, if something has been released that was innovative but did not sell well, the whole idea is canned and treated as a diseased foot fungus by everyone in the industry. These are all my opinions by the way, you can think differently if you desire. My point is, I think voice controlled games are not being made today (commercially) because Lifeline didn’t sell well enough. Its sad story and clunky interface is more to blame then the technology behind it. It clearly demonstrated that voice-controlled games are possible, and with some work they can become more accurate, can have more words it understands (Lifeline knew “thousands of words and phrases” [from the back of the box]). Voice-controlled games, I believe, can be very immersive as it makes you play as you, and you can kind of think of the characters you interact with as being real (after all, it’s a real conversation. You say something, they respond. Well, ideally… more on Lifeline in a bit about this).
My final point is that I want to see more voice-controlled games being made. If you are thinking about making a game, don’t disregard a voice interface. They’re interesting, they’re fun, and right now they are still really innovative!
Okay, now I am going to talk about Lifeline specifically so stop reading if you don’t care. Lifeline is a game where you play an unnamed man stuck in a control-room after attending a swanky party in space that horribly goes wrong (as all parties in space generally do). See, mutant creatures attacked, turning the people into horrible monsters as well and generally making a muck of things (the space station was a research station & space hotel, a combination ripe for unfortunate happenstances). See, I told you the story wasn’t anything to hold in admiration. There is a waitress who is free to roam about and you talk to her, controlling the doors from the control room and generally bossing her around. She wants to find out about something in the research station (looking for her lost dad [lost before the party tragedy] and some other stuff) and you want her to find your missing girlfriend and to have her save you (as you are stuck in the control room, plus you are defenseless from the monsters. Rio, the waitress, has a gun at least). Any-who, that is the main setup. You can command Rio to walk somewhere, look at things, run, play games (tongue twisters, trivia, etc), shoot monsters, etc. She bosses you around by making you open doors for her. The voice recognition software is not 100% accurate like I said before, but once you get used to it, it actually does a decent job of it. You have to speak clearly, make sure you pause between holding down the button (that tells the game to pay attention to what you are saying) and speaking as otherwise I think it doesn’t hear the complete command (likewise, no speaking about something else while pushing the button, it hears stuff it doesn’t recognize). But at the end of the game, I could mostly have her do what I wanted. And mind you, I am more forgiving about such things then a lot of people.
No, in my mind it is not the voice-recognition that is its main failing point (although it is a minor one for me). No, it is the game-mechanics. To have it voice controlled was to make Rio an absolute moron. ‘What should I do, a monster is attacking and I have a gun? Tell me operator!’ Yeah, yeah. Shoot it, Rio, shoot it. ‘I’m about to be hit, what should I do??’ Dodge Rio! She has to be told everything! And she has to be. To have it otherwise is to not have a voice controlled game. It is the inherent limitation of the interface. The creators need to craft a story where the voice-control is natural and the characters believable, otherwise the game just doesn’t work right. Play as a voice in someone’s head to make them do things, or control a robot that has no free will, something like that. But to set up Rio as a human being and then have to tell her what to do 24-7 is silly. Be the operator who can look at the cameras and tell her where the monsters are, or be able to analyze the monster as she fights it to tell her the weak points, that is acceptable. But it is unbelievable that without you telling her to shoot, she will stand there, gun in hand, and be pummeled to death without any resistance (oh, she will complain at you, but that’s it). A stealth game would work well as being voice-controlled, as you could ‘hack’ into the place’s security system (or use satellite imagery or whatever, something to let you see the area) and you could guide the operative around the area, avoiding traps and people. If they had to fight, they could do it themselves and you could offer suggestions (like you saw on a camera two guys talking about how the legs in the new battle bot are weak or something like that so you could tell the agent to aim at the leg joint). There is a wealth of situations where voice-control is natural, and someone needs to exploit it. Maybe me, at thesis time, who knows. But I am tossing it out there to everyone because I think it has promise and I would like to see more of these games.
I was very glad you brought this in to show. I can’t believe I had never even heard of this game until you demo-ed it.
I think my impression of voice control is not very good… I think it is a little more awkward and takes longer than just directing a character with a gamepad. But you’re absolutely right in that it’s a completely different experience, and I really liked how Rio was a character you “interacted” with… it’s very different from BEING the character.
You are 100% on the mark with the fact that you have to tell Rio to do everything, though. But I suspect this was because the game developers had no idea what else they could do and due to voice recognition software still being relatively poor. (Karaoke Revolution is all about freaking singing, and yet it doesn’t bother recognizing your lyrics at all.)
It’ll be incredibly amazing when people are making games based on voice recognition that are more than just what Lifeline did. But I still think Lifeline was the necessary first step to see what it might be like.
Yeah, I hum along to Karaoke Revolution so I can actually be graded well (when I sing it, for some reason it doesn’t like me :^( ).
And I think you make some great points. Making you control another character and not be that character does make it a completely different game and experience. And I too look forward to when a truly decent voice-controlled game comes out (which I hope someone will do eventually).
As for the awkwardness of voice-control, I think that has to do with how Lifeline went about it. If they had chosen a different story, it wouldn’t have been bad. I believe it was because they were making you not the lead character (the one doing everything) but were still having you do everything. It was contradictory and thereby frustrating (if you want me to make her go everywhere and do everything, why didn’t you just make me play as her? Eh, Konomi? Why?). I think you need to have a control structure in which you do not control everything the character you are guiding does, but that you are an integral part in helping the character make the goal, or to develop, or what have you. Does that make sense? You need a role separate from the role of the character you are guiding. They are joined, to be sure, but they should not be the same. Once developers realize this, and make games that reflect this, I don’t think you will have the frustration anymore and the voice-control will be perfectly normal and natural.
Here’s a blurb on some new tech developments in voice control and conversation that might make for a useful game interface:
http://www.technologyreview.com/Biztech/19403/?a=f
That is really cool, thanks for sharing the link!
I think the best use of voice control in games so far is World of Warcraft, where players use VOIP to coordinate massive raids.
In terms of prior art for consoles, Socom got a lot of positive press for its use of a headset, which was included in the box. You could give commands to friendly AI in the single player game as well as talk to other humans in multiplayer. Apparently Rainbow Six had the same feature, which is discussed on Slashdot.
Also, I think one of the big reasons we haven’t seen a lot of voice recognition in games is because the technology tends to suck. You can train it for simple tasks if you speak slowly enough and condense your vocabulary, but speech is still a very hard thing for computers to deal with.
I heard a great story about recognition failure in a class this semester (maybe 541, so apologies if you’ve heard it before…). When Apple was releasing the Newton the CEO went on TV to demo the handwriting recognition and the Newton failed miserably, even though the CEO had practiced with the device (with coaching from engineers) for hours and hours. The problem was, when he was on TV he got nervous. And that affected his handwriting. The same thing happens with voice recognition when people get nervous and their voice changes.
Of course, someday we’ll be able to talk to computers in a natural way. But I think it’ll stay tantalizingly out of reach for awhile.
I was very glad you brought this in, Brandi. Based on what you showed, it seemed like the game was trying too hard to be totally voice-controlled, as opposed to, say, partially voice controlled in ways that make sense. If you were able to move your character manually, but interact with figures on screen, or shout commands, for example, the whole thing might seem less forced and arbitrary. Ian’s example of the voice communication in WoW, seems like a good model — you talk to other team players when you need to coordinate, but use your hands to move around or kill things (and stuff).
There are a number of DS games that use the microphone, like electroplankton, mario kart ds, and several others.
And let’s not forget the dreamcast classic Seaman, which (according to rumor) will be getting some sort of sequel in the future…
One of the first voice controllers i can remember for games was a terrible headset for either the SNES or the Genesis, that was a terrible hybrid of a lightgun and a microphone that would shoot whenever you yelled “fire!”
I was very much interested in the idea of introducing voice as a ‘player’ to the game-player interaction.
My first impression was that Brandi was going to conduct more of a dialog with the character than what was being presented. That’s how far I’m detached from what can vs. cannot be achieved with technology!
Voice invites a whole lot of dynamic to the interaction with digital characters as well as potential for bugs and underdeveloped features.
Most importantly, the richness of voice in a game can allow to forgo a lot of the narrative context-building and feedback from the characters about their environment and inner states, and turn player and character into partners in pushing the game forward.
Hopefully more experiments with that as interface will see the light.
Thanks for this thread. I found this link byGoogle. I’m bookmarking this page of yours and I’ll be back when you have more updates like this.