Not too sure about angle. That can be seriously wicked.
Sounds like a serious tech problem. Maybe OpenNI is the easiest way of placing the hands in a 3D space (IMO).
3D? is this a similar problem to the theremin problem? no tactile feedback. There are probably other game types that are more intuitive with Kinect. Actual XBOX kinect games do something similar right?
Invent something new, rather than set people up for something they expect that you can’t recreate. You get the asshole who’s like, “why don’t you just gimme a f’in basketball?”
Real ball? If not (And I assume not), it seems like you definitely want OpenNI and basically Gesture Recognition stuff. If you think about how wii bowling and stuff handles this type of thing, you don’t need to totally realistically recreate the physics etc. of real basketball – you just need to make an analogous gameplay challenge, with analogous levels and types of skill.
I think a major issue will be how to judge the release of the ball. I could imagine the ball ‘sticking’ to someone’s hand and creating issues.
Its an interesting idea, but I have no idea how to get the proper release data, it seems catching the ball would be easier, for rebounding, or setting up. I think abstracting it would be helpful some, instead of pure horse.
Would the ball come back to you after you throw it? Or would you just get a new ball? Not sure about the hand release. Maybe you can explore other interactions. One thought is maybe you can do something more abstract with the ball. Perhaps throwing the ball around leaves some imprint or does something to the virtual environment, and if multiple people participate you create something as a group.
I think you should explore the body pose matching game that john suggested. I’d like to play it :P
One of the biggest drawbacks of the kinect is that it’s bad at “firing” actions. Because everything is determined with some percent certainty, it’s difficult to get a Kinect game to jump exactly when you want to jump or shoot exactly when you want to shoot. I think you’ll run into that problem pretty fast. At what point does the person’s hand actually let go of the ball? You can never be 100% sure when they mean to let go.
I think the trick is to NOT put the kinect in front of the user. Put it to the side and let it capture the user’s profile so you can see the shot better.
Requiring that the user jump when they shoot might make it easier to detect the shot. Just a thought.
Doing this game without OpenNI could be… challenging. Maybe you could architect it in a way that plays into a simpler hand detection algorithm, like the method from Ben’s Kinect project.
This sounds really fun! will there be a point system or something like that?
What if people had to wear gloves that had IR lights on them. You could maybe do this with a Wii. < -- ++1
What about creating one of those games where you do a horse race with water guns. Maybe you do a horse race based on how quickly people can jump up and down.
You will need high-speed cameras, e.g. PS3 Eye is 120fps, that's a minimum.
John's suggestion to match someone's pose is REALLY INTERESTING. that could be an extremely fun app for 2 people!!!
This will definitely be harder then you think. I agree, simulation of a existing interface/interaction will be difficult since people have such a high expectation based on what they are already used to ... easier to look for new interactions in this case.
Check this out Ward. A pose matching game based on video stills: http://johnhenryshammer.com/projects/WatchWhat/watchWhat.html It’s a single player thing and I think a 2 perosn faceoff adds a much nicer social dynamic. More info here: http://interactivos.eyebeam.org/dokuwiki/doku.php/watch_what_you_are
Here are todays comments:
Not too sure about angle. That can be seriously wicked.
Sounds like a serious tech problem. Maybe OpenNI is the easiest way of placing the hands in a 3D space (IMO).
3D? is this a similar problem to the theremin problem? no tactile feedback. There are probably other game types that are more intuitive with Kinect. Actual XBOX kinect games do something similar right?
Invent something new, rather than set people up for something they expect that you can’t recreate. You get the asshole who’s like, “why don’t you just gimme a f’in basketball?”
Real ball? If not (And I assume not), it seems like you definitely want OpenNI and basically Gesture Recognition stuff. If you think about how wii bowling and stuff handles this type of thing, you don’t need to totally realistically recreate the physics etc. of real basketball – you just need to make an analogous gameplay challenge, with analogous levels and types of skill.
I think a major issue will be how to judge the release of the ball. I could imagine the ball ‘sticking’ to someone’s hand and creating issues.
Its an interesting idea, but I have no idea how to get the proper release data, it seems catching the ball would be easier, for rebounding, or setting up. I think abstracting it would be helpful some, instead of pure horse.
Would the ball come back to you after you throw it? Or would you just get a new ball? Not sure about the hand release. Maybe you can explore other interactions. One thought is maybe you can do something more abstract with the ball. Perhaps throwing the ball around leaves some imprint or does something to the virtual environment, and if multiple people participate you create something as a group.
I think you should explore the body pose matching game that john suggested. I’d like to play it :P
One of the biggest drawbacks of the kinect is that it’s bad at “firing” actions. Because everything is determined with some percent certainty, it’s difficult to get a Kinect game to jump exactly when you want to jump or shoot exactly when you want to shoot. I think you’ll run into that problem pretty fast. At what point does the person’s hand actually let go of the ball? You can never be 100% sure when they mean to let go.
I think the trick is to NOT put the kinect in front of the user. Put it to the side and let it capture the user’s profile so you can see the shot better.
Requiring that the user jump when they shoot might make it easier to detect the shot. Just a thought.
Doing this game without OpenNI could be… challenging. Maybe you could architect it in a way that plays into a simpler hand detection algorithm, like the method from Ben’s Kinect project.
This sounds really fun! will there be a point system or something like that?
What if people had to wear gloves that had IR lights on them. You could maybe do this with a Wii. < -- ++1 What about creating one of those games where you do a horse race with water guns. Maybe you do a horse race based on how quickly people can jump up and down. You will need high-speed cameras, e.g. PS3 Eye is 120fps, that's a minimum. John's suggestion to match someone's pose is REALLY INTERESTING. that could be an extremely fun app for 2 people!!! This will definitely be harder then you think. I agree, simulation of a existing interface/interaction will be difficult since people have such a high expectation based on what they are already used to ... easier to look for new interactions in this case. Check this out Ward. A pose matching game based on video stills: http://johnhenryshammer.com/projects/WatchWhat/watchWhat.html It’s a single player thing and I think a 2 perosn faceoff adds a much nicer social dynamic. More info here: http://interactivos.eyebeam.org/dokuwiki/doku.php/watch_what_you_are