Can the wiimote

  • Page

    of 2 First / Last

    Previous
  • Youthist 4 Sep 2007 13:19:59 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    Model 3d movement ?

    For example, could you develop a piece of demoware that shows somones virtual hand on screen holding a virtual Wiimote. Then you as the player simply move your hand around in 3d space (rotate, up , down etc you get the picture) and it fully show your "virtual wiimote" on the screen behaving in the same way ?

    Can this be done, opinions plz.

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Ignatius_Cheese Moderator 4 Sep 2007 13:23:13 10,861 posts
    Seen 7 hours ago
    Registered 11 years ago
    I don't think it can, unfortunately... I was hoping that a new Mario Paint title would be created allowing the sculpting or lathing of 3D blocks into shapes.

    AFAIK, it can sense rotation and tilt together with the distance of the remote from the sensor bar, which is not quite what I expected when I imagined the "hand inside a box and the movements being sensed in 3D space".
  • Youthist 4 Sep 2007 13:24:30 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    Ignatius_Cheese wrote:
    I don't think it can, unfortunately... I was hoping that a new Mario Paint title would be created allowing the sculpting or lathing of 3D blocks into shapes.

    AFAIK, it can sense rotation and tilt together with the distance of the remote from the sensor bar, which is not quite what I expected when I imagined the "hand inside a box and the movements being sensed in 3D space".
    Yes, but what you say above, means that you could develop my demoware as above using those principles?

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Hunam 4 Sep 2007 13:25:08 20,674 posts
    Seen 7 hours ago
    Registered 8 years ago
    I'm not sure if its the same thing, but the baseball bat in wii sports seems to be abit like 3d movement... as is boxing.

    Not sure if its the same thing you are talking about.
  • Youthist 4 Sep 2007 13:26:00 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    Hunam wrote:
    I'm not sure if its the same thing, but the baseball bat in wii sports seems to be abit like 3d movement... as is boxing.

    Not sure if its the same thing you are talking about.
    I am just talking about my scenario above. Which part of it do you not understand ? :-/

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Pirotic Moderator 4 Sep 2007 13:26:32 20,647 posts
    Seen 11 hours ago
    Registered 13 years ago
    It can, but not very well.

    the accelerometer could be used to keep track of it's guestimated position, but that's about it. If it could do it accurately you wouldn't have needed that infrared sensor bar to keep it in check with.
  • haowan 4 Sep 2007 13:26:56 7,385 posts
    Seen 5 days ago
    Registered 9 years ago
    No, it can't. You might create some kind of approximation, or be able to fool people (Hunam ;)) into thinking that tilt is doing some kind of 3D positioning, but it'll never work like, say, a properly modelled tennis racquet moving in 3D.

    edit: your demo could be done, but you'd need some calibration from the user etc.
  • Ignatius_Cheese Moderator 4 Sep 2007 13:27:38 10,861 posts
    Seen 7 hours ago
    Registered 11 years ago
    Youthist wrote:
    Ignatius_Cheese wrote:
    I don't think it can, unfortunately... I was hoping that a new Mario Paint title would be created allowing the sculpting or lathing of 3D blocks into shapes.

    AFAIK, it can sense rotation and tilt together with the distance of the remote from the sensor bar, which is not quite what I expected when I imagined the "hand inside a box and the movements being sensed in 3D space".
    Yes, but what you say above, means that you could develop my demoware as above using those principles?
    In theory, yes. But it won't be pinpoint accurate, due to the IR field being conical from the sensor bar.
  • gizmo 4 Sep 2007 13:28:24 2,051 posts
    Seen 12 hours ago
    Registered 13 years ago
    I read the thread title and thought 'I agree'.
  • Youthist 4 Sep 2007 13:28:58 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    gizmo wrote:
    I read the thread title and thought 'I agree'.
    Thanks for dropping by...see you in the 360 threads.

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Pirotic Moderator 4 Sep 2007 13:29:27 20,647 posts
    Seen 11 hours ago
    Registered 13 years ago
    Ignatius_Cheese wrote:
    Youthist wrote:
    Ignatius_Cheese wrote:
    I don't think it can, unfortunately... I was hoping that a new Mario Paint title would be created allowing the sculpting or lathing of 3D blocks into shapes.

    AFAIK, it can sense rotation and tilt together with the distance of the remote from the sensor bar, which is not quite what I expected when I imagined the "hand inside a box and the movements being sensed in 3D space".
    Yes, but what you say above, means that you could develop my demoware as above using those principles?
    In theory, yes. But it won't be pinpoint accurate, due to the IR field being conical from the sensor bar.

    the IR field thing is just so it knows where the TV is, rather than where the wii mote is in space, obviously when it can see the dots it has an extra input so help it come up with a more accurate calculation of it's movement, but it can still calculate it without the dots using the accelerometer. the rotation thing is trivial.
  • Youthist 4 Sep 2007 13:30:21 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago


    the IR field thing is just so it knows where the TV is, rather than where the wii mote is in space, obviously when it can see the dots it has an extra input so help it come up with a more accurate calculation of it's movement, but it can still calculate it without the dots using the accelerometer. the rotation thing is trivial.

    thats what I was thinking - surely its just about the gyroscope within the wiimote?

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Pirotic Moderator 4 Sep 2007 13:31:07 20,647 posts
    Seen 11 hours ago
    Registered 13 years ago
    yup, but it's not very accurate at all - thus why they put the sensor bar over the TV to help boost it's accuracy, in 3D space that wouldn't happen :(
  • Youthist 4 Sep 2007 13:33:04 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    Pirotic wrote:
    yup, but it's not very accurate at all - thus why they put the sensor bar over the TV to help boost it's accuracy, in 3D space that wouldn't happen :(

    it seems "close" in wii sports though (particularly the baseball when waiting to receive a pitch and when doing a practise swing in the golf)....so I was hoping the "Brains" could take this further in the future...

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • JetSetWilly 4 Sep 2007 13:35:29 5,721 posts
    Seen 2 weeks ago
    Registered 9 years ago
    But isn't Wii Sports just sensing a start and end movement of any type and the speed between the two? In golf if I jab the Wiimote at the floor rather than pull it back the golfer swings even though I'm not.
  • Youthist 4 Sep 2007 13:46:53 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    JetSetWilly wrote:
    But isn't Wii Sports just sensing a start and end movement of any type and the speed between the two? In golf if I jab the Wiimote at the floor rather than pull it back the golfer swings even though I'm not.

    Yeah, but if you also go in the "correct swing position" and just move arms very slowly up, and tiny amounts up and down, it detects the smallest movements of the gyroscope and reflects it on screen immediately, ie 1:1 movement. Could you take this further onto the "virtual wiimote" scenario in my initial post?

    Yes there is away to make it do the movement in other ways, but also can the "correct" way can be modelled perfectly?

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Der_tolle_Emil 4 Sep 2007 13:50:22 7,301 posts
    Seen 7 hours ago
    Registered 9 years ago
    haowan wrote:
    No, it can't. You might create some kind of approximation, or be able to fool people (Hunam ;)) into thinking that tilt is doing some kind of 3D positioning, but it'll never work like, say, a properly modelled tennis racquet moving in 3D.

    edit: your demo could be done, but you'd need some calibration from the user etc.

    You definetly need some calibration. The Wiimote knows when and how it's moved - all directions, every angle etc. But it's all useless if it doesn't know where it originally was. The only thing you can really calculate is the relative position to the original, calibrated one. That in fact should be enough to replicate a virtual Wiimote.

    The question is how accurate this will be in the long run. It will definetly need some recalibration every now and then.
  • AliRay 4 Sep 2007 16:55:59 181 posts
    Seen 2 years ago
    Registered 8 years ago
    But just because it's not been done yet, doesn't mean it isn't possible, surely?

    Won't devs be able to get to grips with programming for the WiiMote in the same way as squeezing better graphics out of older consoles, surely its just a case of experience and optimisation?
  • T.G. 4 Sep 2007 16:57:39 5,989 posts
    Seen 6 years ago
    Registered 9 years ago
    The strict answer is no, but broadly speaking, it kinda can yeah. Just not very well.
  • LeD 4 Sep 2007 16:59:04 6,314 posts
    Seen 6 hours ago
    Registered 8 years ago
    The strict answer is 'Yes, please!'.
  • AliRay 4 Sep 2007 17:03:04 181 posts
    Seen 2 years ago
    Registered 8 years ago
    Anyway, this is all largely irrelevant. Once everyone realises that the wiimote/nunchuck combo is as good as keyboard/mouse for FPSs, we won't care about movement in 3D space. Too much free online deathmatch, boo-ya!
  • FiveManArmy 4 Sep 2007 17:03:20 506 posts
    Seen 6 months ago
    Registered 9 years ago
    Youthist wrote:
    JetSetWilly wrote:
    But isn't Wii Sports just sensing a start and end movement of any type and the speed between the two? In golf if I jab the Wiimote at the floor rather than pull it back the golfer swings even though I'm not.

    Yeah, but if you also go in the "correct swing position" and just move arms very slowly up, and tiny amounts up and down, it detects the smallest movements of the gyroscope and reflects it on screen immediately, ie 1:1 movement. Could you take this further onto the "virtual wiimote" scenario in my initial post?

    Yes there is away to make it do the movement in other ways, but also can the "correct" way can be modelled perfectly?

    It's not 1:1 movement though. It's just sensing the wii remote is being tilted and hoping that you are in a golf position. You could just hold the remote out and tilt it and it would complete the golf swing exactly like you said.
  • Youthist 4 Sep 2007 17:06:37 10,083 posts
    Seen 14 hours ago
    Registered 9 years ago
    FiveManArmy wrote:
    Youthist wrote:
    JetSetWilly wrote:
    But isn't Wii Sports just sensing a start and end movement of any type and the speed between the two? In golf if I jab the Wiimote at the floor rather than pull it back the golfer swings even though I'm not.

    Yeah, but if you also go in the "correct swing position" and just move arms very slowly up, and tiny amounts up and down, it detects the smallest movements of the gyroscope and reflects it on screen immediately, ie 1:1 movement. Could you take this further onto the "virtual wiimote" scenario in my initial post?

    Yes there is away to make it do the movement in other ways, but also can the "correct" way can be modelled perfectly?

    It's not 1:1 movement though. It's just sensing the wii remote is being tilted and hoping that you are in a golf position. You could just hold the remote out and tilt it and it would complete the golf swing exactly like you said.

    Agreed. But I am fine with that. I just want it to be "possible" if i hold it in the correct way and tilt and move it. Yes I could "cheat the system" , but I am fine as long as it can be done "correctly".

    For example, I never stand "side on" when playing golf games on the wii. I always stand square on to the screen, and swing "in front" (ie from left to right) of the screen as opposed to "away from and towards" the screen.

    It's much bettter, as you can see the screen as you play through your shot as opposed to having your face not directed at the screen when you at rigth angles to the screen.
    Works much better in Pangya golf, for example, as it means you can see the power bars and contact points easily as you are always facing the screen.

    "Good chess players think one move ahead. Great chess players think two or three moves ahead."

  • Der_tolle_Emil 4 Sep 2007 17:29:26 7,301 posts
    Seen 7 hours ago
    Registered 9 years ago
    FiveManArmy wrote:
    Youthist wrote:
    JetSetWilly wrote:
    But isn't Wii Sports just sensing a start and end movement of any type and the speed between the two? In golf if I jab the Wiimote at the floor rather than pull it back the golfer swings even though I'm not.

    Yeah, but if you also go in the "correct swing position" and just move arms very slowly up, and tiny amounts up and down, it detects the smallest movements of the gyroscope and reflects it on screen immediately, ie 1:1 movement. Could you take this further onto the "virtual wiimote" scenario in my initial post?

    Yes there is away to make it do the movement in other ways, but also can the "correct" way can be modelled perfectly?

    It's not 1:1 movement though. It's just sensing the wii remote is being tilted and hoping that you are in a golf position. You could just hold the remote out and tilt it and it would complete the golf swing exactly like you said.

    But you realize that this is still 1:1 movement? Just not relative to the golf ball. If it knew where the ball is lying (thus calibrating) it wouldn't work. But it still detects the same movement. If you calibrated the Wiimote once, like having to point at a certain point on screen using the sensor bar, the game knows that you are in front of the TV now. All you have to do now is track every movement the Wiimote detects (and it can detect every kind of movement there is. All 3 axes and in which angle it's held) and the game knows exactly where you are relative to the point you were when you were standing exactly in front of the TV. It would know if you moved 2 meters to the left and are now holding the Wiimote a few centimeters higher than before.

    You can take it to the extreme as well. Stand in front of the TV, point at the screen. Then move as far as possible to the wall left of you, point at the screen again to tell the game that you are at the far left. Next move to the wall right to you, point at the screen and let the game know that you have hit the wall. To the same with the wall behind you and then get as close as possible to the TV with the Wiimote still being recognized. You now have 5 readings - you don't have the exact coordinates of your walls, or the space you can move in, but the game knows - thanks to the accelerometers - how far you can move to the left/right/back from right in front of the TV until you cannot go any further. This info alone is enough to calculate where the Wiimote is in the room - and thanks to the tilting sensors it knows how it's being held.

    The problem with this method is that it's probably not very precise. The Wiimote doesn't automatically know where it is. You cannot turn it on and the game knows exactly where you are. You can however set a specific point like the middle of the room and calculate the relative position to this point. This is more than enough to let the game know the 'exact' position from now on. Depending on the quality of the accelerometers and the tracking of the movement (quite cpu hungry I suppose) you can easily create a 'virtual movement space' for simpler tasks. The biggest problem is the preciseness - but technically it's definetly possible.
  • Deleted user 4 September 2007 21:27:57
    Infra red depends on nothing being in the way so triangulaiton wouldnt work very well. Even if you had 2 other sensors and a camera with a wider field of view on the wiimote.

  • tiddles 4 Sep 2007 21:31:15 3,555 posts
    Seen 7 days ago
    Registered 13 years ago
    gizmo wrote:
    I read the thread title and thought 'I agree'.

    My thoughts exactly... so I was extremely disappointed to find out that the thread was in fact a reasoned and intelligent exploration of the technical potential of Nintendo's motion sensing peripheral.
  • [+..] 4 Sep 2007 21:58:26 341 posts
    Seen 2 years ago
    Registered 8 years ago
    There would be no way to triangulate the vertical axis. You could work out the angle but not the height.
  • FiveManArmy 4 Sep 2007 22:18:43 506 posts
    Seen 6 months ago
    Registered 9 years ago
    Der_tolle_Emil wrote:
    FiveManArmy wrote:
    Youthist wrote:
    JetSetWilly wrote:
    But isn't Wii Sports just sensing a start and end movement of any type and the speed between the two? In golf if I jab the Wiimote at the floor rather than pull it back the golfer swings even though I'm not.

    Yeah, but if you also go in the "correct swing position" and just move arms very slowly up, and tiny amounts up and down, it detects the smallest movements of the gyroscope and reflects it on screen immediately, ie 1:1 movement. Could you take this further onto the "virtual wiimote" scenario in my initial post?

    Yes there is away to make it do the movement in other ways, but also can the "correct" way can be modelled perfectly?

    It's not 1:1 movement though. It's just sensing the wii remote is being tilted and hoping that you are in a golf position. You could just hold the remote out and tilt it and it would complete the golf swing exactly like you said.

    But you realize that this is still 1:1 movement? Just not relative to the golf ball. If it knew where the ball is lying (thus calibrating) it wouldn't work. But it still detects the same movement. If you calibrated the Wiimote once, like having to point at a certain point on screen using the sensor bar, the game knows that you are in front of the TV now. All you have to do now is track every movement the Wiimote detects (and it can detect every kind of movement there is. All 3 axes and in which angle it's held) and the game knows exactly where you are relative to the point you were when you were standing exactly in front of the TV. It would know if you moved 2 meters to the left and are now holding the Wiimote a few centimeters higher than before.

    You can take it to the extreme as well. Stand in front of the TV, point at the screen. Then move as far as possible to the wall left of you, point at the screen again to tell the game that you are at the far left. Next move to the wall right to you, point at the screen and let the game know that you have hit the wall. To the same with the wall behind you and then get as close as possible to the TV with the Wiimote still being recognized. You now have 5 readings - you don't have the exact coordinates of your walls, or the space you can move in, but the game knows - thanks to the accelerometers - how far you can move to the left/right/back from right in front of the TV until you cannot go any further. This info alone is enough to calculate where the Wiimote is in the room - and thanks to the tilting sensors it knows how it's being held.

    The problem with this method is that it's probably not very precise. The Wiimote doesn't automatically know where it is. You cannot turn it on and the game knows exactly where you are. You can however set a specific point like the middle of the room and calculate the relative position to this point. This is more than enough to let the game know the 'exact' position from now on. Depending on the quality of the accelerometers and the tracking of the movement (quite cpu hungry I suppose) you can easily create a 'virtual movement space' for simpler tasks. The biggest problem is the preciseness - but technically it's definetly possible.

    I could be wrong, but I don't think this is correct. The remote can't detect where it is in 3d space, it may be able to detect roughly how it's being moved or tilted but not accurately enough to calculate it's exact position. I don't think the remote has enough information to work this out.
  • afray 4 Sep 2007 22:23:50 2,251 posts
    Seen 2 weeks ago
    Registered 10 years ago
    The info's there - enough to get distance from the TV screen - but errors would creep in very rapidly when you start moving it about, and you'd have a tough time telling gravity from the acceleration forces. I don't think it would be practical to do the app in the original post.
  • Der_tolle_Emil 5 Sep 2007 08:28:23 7,301 posts
    Seen 7 hours ago
    Registered 9 years ago
    [+..] wrote:
    There would be no way to triangulate the vertical axis. You could work out the angle but not the height.

    True, I forgot about that aspect. But you can still calculate up and down movement.

    FiveManArmy: It has all the information to track every kind of movement possible. You are probably right about the accuracy though - as afray said, it's unlikely that it's accurate enough (without using full CPU load all the time) and thus probably not really an option in a real game or any useful application rather than a tech demo.
  • Page

    of 2 First / Last

    Previous
Log in or register to reply