[PATV] Wednesday, April 11, 2012 - Extra Credits Season 4, Ep. 9: Kinect Disconnect

Geth

Extra Credits Season 4, Ep. 9: Kinect Disconnect
http://penny-arcade.com/patv/episode/kinect-disconnect

This week, we look back on the last year and a half of Kinect and offer some new observations about the device.

Come discuss this topic in the forums!

jackal

This is something I've been thinking about because I got my kids Kinect Rush: A Disney Pixar Adventure (why is the name so long?). The Incredibles game mode is actually fun because it's basically a rhythm game that isn't set to a beat. It's very linear and performing actions in rapid succession at the right times is important. 'Pump you arms' 'jump' 'arms up' 'jump' that sort of thing. The Toy Story mode uses the same controls, but it is terrible. It's basically pump your arms and lean your shoulders to navigate the environment looking for batteries. The Cars mode is actually physically painful. I don't know how the game designers didn't realize that making the player hold their arms straight out for four minutes would be a bad idea.

So I basically came to the same conclusion that what you are doing with your body needs to be fun for it to work at all.

RMS Oceanic

Hmm, using the Uncanny Valley to describe the issues of Kinect's immersion is a pretty novel way of approaching it. I approve.

2 Marcus 2 Ravens

This is a pretty great episode.

My experience with Kinect is minimal, but I can honestly say the most fun I had with it was dicking around in the menus. That's a bit less harsh than it sounds, because I actually had quite a bit of fun with those menus.

But I find it interesting to think about this compared to the Wii or Move. I know when I'm playing the Wii, most of the gestures I have to do are trying to merely hint at some greater gesture. In Skyward Sword, I swing that Wii-mote around, hacking and slashing my way to good times. But my hand is going through a fairly limited range of motion. Using your uncanny valley robot metaphor, it's the cute, Wall-E type motion controls, only hinting at actual human motion, while Kinect is the horrifying rubber-faced robots from hell.

I had never considered there would be such a line to be drawn with motion controls, as they're all quite gimmicky by nature, but it's absolutely true.

You learn somethin' every day.

Aurich

Considering the point was that the motion controls need to be inherently fun movements (which is a fairly straightforward concept) I'm not sure I followed the uncanny valley comparison. I doesn't seem to me that any motion control is like enough to its real counterpart to turn someone off for that reason. I mean.. wouldn't something like Dance Central be guilty in that case?

Tubular

Aurich wrote: »

Considering the point was that the motion controls need to be inherently fun movements (which is a fairly straightforward concept) I'm not sure I followed the uncanny valley comparison. I doesn't seem to me that any motion control is like enough to its real counterpart to turn someone off for that reason. I mean.. wouldn't something like Dance Central be guilty in that case?

No, because with something like Dance Central, you're actually doing the thing that's happening on screen.
The uncanny valley type issue sets in when you're doing something in Kinect that is removed from what you're physically doing. With a controller of some kind, your brain is able to make the distinction that you're using a tool, and make the proper association. With Kinect, your brain, subconsciously, is saying 'hey, wait a sec, this isn't matching up'.

Framling

Good episode, and the uncanny valley explanation was pretty solid. It really highlights one of the biggest problems I've had, historically, with 'natural' interfaces. They don't feel natural. They feel like I'm trying to figure out what the interface wants from me, instead of letting my kinesthetic brain parts pick up the slack.

I think the lesson to learn is that it's better to move the in-game activity closer to the actual physical activity than the other way around. Dancing in a video game by actually dancing can work. Running in a video game by running in place in the real world feels weird. I guess I just don't know how many activities that includes. Large movements of the body, while remaining in place, and probably not holding anything?

Personally, I've gotten more value from being able to yell at my TV. I've mentioned elsewhere, how one of the most compelling moments I experienced in ME3 was

on Palaven's moon, coming around a corner to a huge vista of Palaven burning in the sky above us, and involuntarily letting out a low, 'holy shit' whistle, and having James react.

I don't know, I may have imagined it, but even if I did, that kind of interaction has a lot of potential.

FULL DISCLOSURE: I AM A FULL-TIME EMPLOYEE OF MICROSOFT CORPORATION. ANY OPINIONS EXPRESSED HEREIN ARE MY OWN, AND NOT THOSE OF MICROSOFT CORPORATION OR ITS SUBSIDIARIES OR AFFLIATES. ANY LIKENESS TO ANY PERSONS OR EVENTS, LIVING OR DEAD, ARE COINCIDENTAL. THIS DISCLAIMER NOT TO BE REMOVED UNDER PENALTY OF LAW EXCEPT BY THE END CONSUMER. THIS TELECAST IS COPYRIGHTED BY THE NFL FOR THE PRIVATE USE OF OUR AUDIENCE. ANY OTHER USE OF THIS TELECAST, OR OF ANY PICTURES, DESCRIPTIONS, OR ACCOUNTS OF THE GAME WITHOUT THE NFL'S CONSENT IS PROHIBITED.

Aurich

TubularLuggage wrote: »

Aurich wrote: »

Considering the point was that the motion controls need to be inherently fun movements (which is a fairly straightforward concept) I'm not sure I followed the uncanny valley comparison. I doesn't seem to me that any motion control is like enough to its real counterpart to turn someone off for that reason. I mean.. wouldn't something like Dance Central be guilty in that case?

No, because with something like Dance Central, you're actually doing the thing that's happening on screen.
The uncanny valley type issue sets in when you're doing something in Kinect that is removed from what you're physically doing. With a controller of some kind, your brain is able to make the distinction that you're using a tool, and make the proper association. With Kinect, your brain, subconsciously, is saying 'hey, wait a sec, this isn't matching up'.

In that case I just don't see how it's uncanny. That would be like a robot that is supposed to be charming but has zero human characteristic: It's just ineffective.

Framling

Aurich wrote: »

TubularLuggage wrote: »

Aurich wrote: »

Considering the point was that the motion controls need to be inherently fun movements (which is a fairly straightforward concept) I'm not sure I followed the uncanny valley comparison. I doesn't seem to me that any motion control is like enough to its real counterpart to turn someone off for that reason. I mean.. wouldn't something like Dance Central be guilty in that case?

No, because with something like Dance Central, you're actually doing the thing that's happening on screen.
The uncanny valley type issue sets in when you're doing something in Kinect that is removed from what you're physically doing. With a controller of some kind, your brain is able to make the distinction that you're using a tool, and make the proper association. With Kinect, your brain, subconsciously, is saying 'hey, wait a sec, this isn't matching up'.

In that case I just don't see how it's uncanny. That would be like a robot that is supposed to be charming but has zero human characteristic: It's just ineffective.

Did you watch the video? The point of the uncanny valley is that a robot that looks like a robot can be cute because of human-like qualities, but a robot that looks like a human will tend to be creepy because of its robot-like qualities. Your brain sees one as obviously a robot, but the other as a human with some unsettling defects. On a less-than-conscious level, you're not thinking "Look how real that robot looks," you're thinking "What is wrong with that person?"

The analogy here is that with a controller, your brain is already very good at associating actions with unrelated results, e.g., moving a piece of plastic on your desk to make a little picture of a hand or an arrow move on a screen in front of you. When you perform that kind of action frequently enough, you don't even consciously make a distinction anymore. You get used to pushing A to jump, so that when you want to jump, you don't think "Push A," you think "jump." But with something like the Kinect, the input you provide is much closer to the desired output in the game. Instead of pushing a thumbstick forward to walk or run, you actually move your legs. It's too close for that amazing part of your brain that associates things like pushing a piece of plastic with moving a little arrow on a screen. But it's too far away to just leave it to the part of your brain that associates running with moving your legs. Just like the too-real robot is seen as a human, but messed up, running in place is seen as running, but messed up.

Aurich

Framling wrote: »

Aurich wrote: »

TubularLuggage wrote: »

Aurich wrote: »

Considering the point was that the motion controls need to be inherently fun movements (which is a fairly straightforward concept) I'm not sure I followed the uncanny valley comparison. I doesn't seem to me that any motion control is like enough to its real counterpart to turn someone off for that reason. I mean.. wouldn't something like Dance Central be guilty in that case?

No, because with something like Dance Central, you're actually doing the thing that's happening on screen.
The uncanny valley type issue sets in when you're doing something in Kinect that is removed from what you're physically doing. With a controller of some kind, your brain is able to make the distinction that you're using a tool, and make the proper association. With Kinect, your brain, subconsciously, is saying 'hey, wait a sec, this isn't matching up'.

In that case I just don't see how it's uncanny. That would be like a robot that is supposed to be charming but has zero human characteristic: It's just ineffective.

Did you watch the video? The point of the uncanny valley is that a robot that looks like a robot can be cute because of human-like qualities, but a robot that looks like a human will tend to be creepy because of its robot-like qualities. Your brain sees one as obviously a robot, but the other as a human with some unsettling defects. On a less-than-conscious level, you're not thinking "Look how real that robot looks," you're thinking "What is wrong with that person?"

The analogy here is that with a controller, your brain is already very good at associating actions with unrelated results, e.g., moving a piece of plastic on your desk to make a little picture of a hand or an arrow move on a screen in front of you. When you perform that kind of action frequently enough, you don't even consciously make a distinction anymore. You get used to pushing A to jump, so that when you want to jump, you don't think "Push A," you think "jump." But with something like the Kinect, the input you provide is much closer to the desired output in the game. Instead of pushing a thumbstick forward to walk or run, you actually move your legs. It's too close for that amazing part of your brain that associates things like pushing a piece of plastic with moving a little arrow on a screen. But it's too far away to just leave it to the part of your brain that associates running with moving your legs. Just like the too-real robot is seen as a human, but messed up, running in place is seen as running, but messed up.

Did you read what I was responding to?

Really, I'm saying I think the analogy was unnecessary for the point they were trying to make: That motion controls need to be fun of themselves rather than an imitation of whatever's going on onscreen. We all know what uncanny valley is and why it is bad. I don't think motion controls are anywhere near the level of fidelity where a person can have that kind of reaction. They will just have fun or be bored or frustrated. Also probably tired, but they will not be revulse'd or creep'd out.

Ivar

Aurich wrote: »

Framling wrote: »

Aurich wrote: »

TubularLuggage wrote: »

Aurich wrote: »

Considering the point was that the motion controls need to be inherently fun movements (which is a fairly straightforward concept) I'm not sure I followed the uncanny valley comparison. I doesn't seem to me that any motion control is like enough to its real counterpart to turn someone off for that reason. I mean.. wouldn't something like Dance Central be guilty in that case?

No, because with something like Dance Central, you're actually doing the thing that's happening on screen.
The uncanny valley type issue sets in when you're doing something in Kinect that is removed from what you're physically doing. With a controller of some kind, your brain is able to make the distinction that you're using a tool, and make the proper association. With Kinect, your brain, subconsciously, is saying 'hey, wait a sec, this isn't matching up'.

In that case I just don't see how it's uncanny. That would be like a robot that is supposed to be charming but has zero human characteristic: It's just ineffective.

Did you watch the video? The point of the uncanny valley is that a robot that looks like a robot can be cute because of human-like qualities, but a robot that looks like a human will tend to be creepy because of its robot-like qualities. Your brain sees one as obviously a robot, but the other as a human with some unsettling defects. On a less-than-conscious level, you're not thinking "Look how real that robot looks," you're thinking "What is wrong with that person?"

The analogy here is that with a controller, your brain is already very good at associating actions with unrelated results, e.g., moving a piece of plastic on your desk to make a little picture of a hand or an arrow move on a screen in front of you. When you perform that kind of action frequently enough, you don't even consciously make a distinction anymore. You get used to pushing A to jump, so that when you want to jump, you don't think "Push A," you think "jump." But with something like the Kinect, the input you provide is much closer to the desired output in the game. Instead of pushing a thumbstick forward to walk or run, you actually move your legs. It's too close for that amazing part of your brain that associates things like pushing a piece of plastic with moving a little arrow on a screen. But it's too far away to just leave it to the part of your brain that associates running with moving your legs. Just like the too-real robot is seen as a human, but messed up, running in place is seen as running, but messed up.

Did you read what I was responding to?

Really, I'm saying I think the analogy was unnecessary for the point they were trying to make: That motion controls need to be fun of themselves rather than an imitation of whatever's going on onscreen. We all know what uncanny valley is and why it is bad. I don't think motion controls are anywhere near the level of fidelity where a person can have that kind of reaction. They will just have fun or be bored or frustrated. Also probably tired, but they will not be revulse'd or creep'd out.

I think they meant that it's near enough to the uncanny valley that something feels a bit weird, if not so close that it's creepy

Zachary Amaranth

I don't disagree with the video per se, but I do think that perhaps more significantly is the issue of trying to apply Kinect-style controls to games which may be fundamentally at odds with it.

Trying to mesh the console game experience and Kinect controls may be the wrong way to go, and I think another commonality amongst Kinect titles that are good is that they are "different" experiences. A lot of them, in fact, are arcade-style. Not all, but there's a pretty common trend here.

jackal

Things that work well with Kinect:
Dancing
Jumping
Throwing
Kara-te chopping

Things that don't work well with Kinect
Every-fucking thing else.

RoyceSraphim

Did anyone like the Gunstringer?

mispeled

I made an account to join this discussion. It's awesome to see Extra Credits discuss stuff like this. I love this show. This is something I've been thinking about for awhile, ever since Heavy Rain came out. That game added a weird layer of abstraction, too, because, especially in the beginning parts of the game, it abstracted actions into nonsense, actions that we could perform MORE effectively in real life than with a controller. Stuff like setting the table is easier to do in real life than it was in the game, and yet it still hit this uncanny valley feeling. If anyone is interested, I wrote more about this back in Feb 2011. Rather than repost it here, here's a link:

http://mispeled.net/2011/02/08/video-games-abstraction-and-motion-controllers/

It talks about the interesting aspects of learning curve using motion controllers - I think eventually games will progress to the point where we don't need tutorials any more, because actions are mapped 1 to 1, rather than abstracted with a tool or a limited set of possibilities.