Post by: Wiiner on May 18, 2006, 08:57:47 PM
Interesting stuff I think - sheds a little light on how things work...
http://www.xgaming.com/newsletter/Wii%20Dupe.shtml
Post by: wandering on May 18, 2006, 09:18:21 PM
Anyway. Interesting that it seems IGN was half right: if you're too close to the sensor bar, you may need to point the wiimote at the bar:
Quote
But then a Nintendo® rep came to my rescue and told me to take 2 steps back and aim the controller at the sensor bar stuck to the bottom of the TV set...whoah!... that's better.
Also interesting that there are 3 accelerometers in the wiimote (and 2 in the nunchaku.)
Post by: Wiiner on May 18, 2006, 09:26:10 PM
Post by: TrueNerd on May 18, 2006, 10:09:08 PM
Post by: thejeek on May 18, 2006, 11:44:06 PM
I think that, in practice, you'll either be aiming at a specify game object, which will be on the screen and so in range of the bar from most practical angles where you can confortably see the TV, or you'll be making gestures where absolute positional accuracy is not required.
Finally, developers will be aware of the issues and will play test everything extensively before release - everything at E3 still has months of development and testing to go before release.
Post by: trip1eX on May 19, 2006, 03:12:35 AM
The article is a bit off about the boxing arm thing. The wiimote is a gesture system. And it can determine location in 3d space away from the sensor bar and direction and angle and acceleration/speed.
Here's a couple paragraphs about the Wii and sensor bar from another article.
"The sensor bar emits an infrared field out directly in front of the TV. In order to accomplish this, it is necessary to place the sensor bar in the same plane as the TV, either on top of or below the TV is recommended. As the remote is pointed towards the TV, it interacts with the infrared field. Using triangulation logic, the remote is able to determine location, angle, and distance; as you move the remote around, the change in location/angle/distance is calculated. The remote then communicates with the console via Bluetooth, sending this information to it, and from there the software determines how your movements relate to the game world/application.
The sensor bar doesn't receive any type of signal from the wireless remote and is there purely to generate the field. The remote also has a gyrometer and accelerometer to allow it, for instance in the tennis game, to know the difference between a back hand or regular swing. You need to maintain a proper distance from the sensor bar as well; if you come to within about three feet or less, it will not work precisely."
http://www.anandtech.com/tradeshows/showdoc.aspx?i=2757
Post by: Wiiner on May 19, 2006, 05:51:38 AM
Post by: thejeek on May 19, 2006, 05:53:06 AM
[EDIT] triangulating from two points only allows you to determine your position on a circle centered on a line passing through those two points, and in fact triangulation from an arbitrary number of points along the same line in space (e.g. the sensor bar) does not improve this situation.
Post by: UltimatePartyBear on May 19, 2006, 05:54:55 AM
Post by: thejeek on May 19, 2006, 06:00:33 AM
Quote
Nintendo probably won't ever publically announce everything about how it works, sadly.
I think this is probably because it's capabilities are indeed more limited than they've led people to believe: roughly accurate sensing of linear acceleration and rotation plus fairly accurate sensing of angle relative to sensor bar within a limited range of distances and angles. This likely won't afffect gameplay, so long as developers stick with what works well, but it could be a stick for Nintendo's competitors to beat them with
Post by: Wiiner on May 19, 2006, 06:02:13 AM
If you had a projector and you had the sensor bar at the bottom of the screen near your floor, then you calibrated it to point at the screen, when you pointed at the top of your screen near the ceiling the IR signal could be lost...maybe...
Post by: zakkiel on May 19, 2006, 06:16:56 AM
QuoteIt doesn't use those sensors for triangulation, at least not in the way you think. It doesn't sense the absolute distance to the sensors, just the relative distances between the sensors and the ports on the remote. So they won't help locate it in space, just sense orientation.
Exactly. Triangulation from only two points is not sufficient to determine location in 3D space - hence the need for both accelerometers and the sensor bar. Whether this is a problem in practice though remains to be seen.
Post by: Infernal Monkey on May 19, 2006, 06:27:02 AM
Post by: KDR_11k on May 19, 2006, 06:42:23 AM
Solution: Don't sit too close to the TV
. Seriously, if you are close enough to have such viewing angle problems you're too close.
Post by: UltimatePartyBear on May 19, 2006, 06:48:28 AM
Quote
Originally posted by: thejeek
I think this is probably because it's capabilities are indeed more limited than they've led people to believe: roughly accurate sensing of linear acceleration and rotation plus fairly accurate sensing of angle relative to sensor bar within a limited range of distances and angles. This likely won't afffect gameplay, so long as developers stick with what works well, but it could be a stick for Nintendo's competitors to beat them with
How's the competition going to beat them with it when their own controllers are far more limited? Even if Sony's motion sensing was identical to Nintendo's, Nintendo's can be used as a pointer, and with the nunchuck you get two motion sensing controllers in one. Besides, you're downplaying the Wiimote's capabilities too much. It's very accurate sensing of acceleration on three axes simultaneously plus tilt plus rotation plus almost too-accurate by some accounts pointing.
Quote
Originally posted by: Wiiner
"there's no conceivable reason why you couldn't calibrate the thing so that you could point at the TV" - That's what one Nintendo rep told me at E3 too, but all of the Wiimotes I played there that used the pointing you pointed at the bar.
Then I would say that the Wiimotes you used were improperly calibrated or the games you played weren't working right. Some of the impressions I've read explicitly stated that you aim at the screen.
Quote
If you had a projector and you had the sensor bar at the bottom of the screen near your floor, then you calibrated it to point at the screen, when you pointed at the top of your screen near the ceiling the IR signal could be lost...maybe...
If my rough estimates are correct, the maximum screen size exceeds anything you're likely to find in any home theater. However, I think that after the novelty wears off, playing games on a huge screen is more trouble than it's worth. It can be a pain to actually have to turn your head to see incoming enemies, for example. Anyway, way back when Nintendo announced the Wiimote they said they tested it with 100" screens. You wouldn't have to put the bar in the middle of the screen, like with the LCD TopGun, if the viewing angle was big enough, either.
Of course, I wouldn't take anything in the article on xgaming at face value. It's written in such a way that I question the writer's grasp of technology.
Post by: thejeek on May 19, 2006, 10:47:42 AM
Quote
How's the competition going to beat them with it when their own controllers are far more limited? Even if Sony's motion sensing was identical to Nintendo's, Nintendo's can be used as a pointer, and with the nunchuck you get two motion sensing controllers in one. Besides, you're downplaying the Wiimote's capabilities too much. It's very accurate sensing of acceleration on three axes simultaneously plus tilt plus rotation plus almost too-accurate by some accounts pointing.
It sounds like the wiimote is quite accurate enough for any reasonable game use. I simply see a potential problem with people getting worked up if Nintendo have described it as having accurate 3D positioning when it (in a very narrow, technical and probably unimportant sense, admittedly) doesn't - simply because some people would like to see Nintendo fail and it's easy to get the public upset by letting them think they've been lied to.
The fact that competitors controllers perform worse may not be the issue, if they haven't made claims for them that aren't strictly accurate
Post by: Smash_Brother on May 19, 2006, 10:52:08 AM
Quote
Originally posted by: thejeek
It sounds like the wiimote is quite accurate enough for any reasonable game use. I simply see a potential problem with people getting worked up if Nintendo have described it as having accurate 3D positioning when it (in a very narrow, technical and probably unimportant sense, admittedly) doesn't - simply because some people would like to see Nintendo fail and it's easy to get the public upset by letting them think they've been lied to.
The fact that competitors controllers perform worse may not be the issue, if they haven't made claims for them that aren't strictly accurate
Sony lies their goddamn head off and still won the last console war. They promised "Toy Story" graphics for the PS2 and other such nonsense and they clearly didn't deliver on that.
The only punishment for "exaggerating" in the gaming industry seems to be that a few die-hard fanboys get pissed at you and everyone else goes on buying your product.
Post by: thejeek on May 19, 2006, 11:04:41 AM
Post by: willie1234 on May 19, 2006, 11:24:33 AM
Post by: thejeek on May 19, 2006, 11:28:20 AM
Post by: willie1234 on May 19, 2006, 11:48:13 AM
Post by: Switchblade Cross on May 19, 2006, 01:26:47 PM
Post by: Wiiner on May 19, 2006, 01:32:45 PM
Post by: IceCold on May 19, 2006, 01:50:11 PM
QuoteI didn't believe that, and I don't know how anyone would jump to that conclusion after the NRC was revealed. What we have is exactly what I envisioned..
This lead me, and likely you to believe that this was like some sort of 3D grid in the room that the console could detect exactly where it is, and in what position it is in.
Post by: Michael8983 on May 19, 2006, 02:15:09 PM
If the technology at work can indeed detect the tilt/rotation of the of the controller, where it's pointed, and how far away from the sensor bar it is - it can more or less do everything a 3D mouse could even if it technically isn't one.
As for the example of it not being able to detect where in the room the Wii-mote is upon boot-up. Does you're computer know where on your desk your mouse is when it boots up?
Post by: Kairon on May 19, 2006, 02:27:44 PM
Quote
Originally posted by: Michael8983
Didn't a developer of Trauma Center say the game would involve "depth" and gave the example of thrusting the controller forward and backwards to simulate a defibrillator?
If the technology at work can indeed detect the tilt/rotation of the of the controller, where it's pointed, and how far away from the sensor bar it is - it can more or less do everything a 3D mouse could even if it technically isn't one.
As for the example of it not being able to detect where in the room the Wii-mote is upon boot-up. Does you're computer know where on your desk your mouse is when it boots up?
But the Wii controller detects depth by detecting and comparing several signals from the Sensor Bar, probably measuring the angle between two IR rays is my guess.
This means that it can function like a full 3D mouse AS LONG AS it can detect the sensor bar out of it's front window.
~Carmine M. Red
Kairon@aol.com
Post by: Michael8983 on May 19, 2006, 03:10:13 PM
This means that it can function like a full 3D mouse AS LONG AS it can detect the sensor bar out of it's front window."
I see what you mean.
It just seems like it's been suggested that the remote has to ALWAYS be pointed at or near the sensor bar for it to work. I'm just not sure if that's really the case.
It's been confirmed that pointing the Wii-mote too far away from the bar will lose your cursor position in games like Metroid and Red Steel. When that happens, it apparently resets itself after a second or two. People act like it's a technical limitation but it could simply be in the programming. Afterall it has to take into account that in real life situations you might move from one end of the couch to the other, switch chairs, or even pass the remote to another player. If the sensor bar could indeed detect where the remote is pointed even if its pointed far from the bar (like straight down or even towards the player) imagine the frustration of having to try and figure out where to aim the remote to get the cursor to reappear on screen. Having it reset itself when the Wii-mote is pointed too far out of range just makes sense.
Post by: ThePerm on May 19, 2006, 05:02:50 PM
if you turn the controller away from the sensor bar it cannot detect depth...it detects depth by measuring the amount of time it takes the ir to get to the sensor. kinda like echo location..but with light.
as far as light gun style games...they can be done....but are done in a different way. The controller understands tilt. left right, up and down, and "wrist barrelrolls". if tilting the front up and down it would change the y position up or down on the screen, if turning left and right it changes the x position. "barrelrolling" the controller changes the circular position of the "site".....moving your hand in 3d motion affects a 2d cursor..which can then affect a 3d plane....
so what we know....your movement is controlled by your wrist, but probably not your elbow...doing elbow motions would mess up the whole thing
Post by: nemo_83 on May 19, 2006, 06:23:03 PM
Post by: Athrun Zala on May 19, 2006, 07:48:26 PM
Quote64DD games on VC FTW!
Originally posted by: thejeek
Only 4D? For backwards compatibility they should probably have 34D. Or maybe 34DD, just to be double sure...
Post by: wandering on May 20, 2006, 12:57:47 AM
Quote
Originally posted by: nemo_83
I'm confused, didn't Nintendo make an investment in Gyration (demo).
Perhaps that was just for Wario Ware?
Post by: The Omen on May 20, 2006, 07:14:53 AM
Quote
t just seems like it's been suggested that the remote has to ALWAYS be pointed at or near the sensor bar for it to work. I'm just not sure if that's really the case.
You're incorrect sir. The wiimote is pointed at the screen. When shooting in, say a fps like Bond, you would aim at the person you're trying to shoot, not the sensor bar. The fact that the wiimote points in the direction of the sensor bar, but not directly at it/'into it' doesn't matter. I tried a Madden 07 demo, and you aim at the receiver when throwing a pass, certainly not at the bar. Nobody has told me that this isn't the case. If I stood so close to my tv that the wiimote was 6 inches from the glass, I may have a problem(though I'm not sure) because the sensor bar would be just about directly above or below it. But nobody is going to play a game like that anyway. I tried it from 3 feet away, which is closer than I would have liked , and it worked fine.
Post by: zakkiel on May 20, 2006, 08:31:17 AM
Quote
But the Wii controller detects depth by detecting and comparing several signals from the Sensor Bar, probably measuring the angle between two IR rays is my guess.
Quote
That's a very weird way to put it. The remote doesn't "interact" with a "field." Depending on its position, the bright spots from the IR LEDs will show up in different spots on its CMOS sensor.
That would require sophisticated optical equipment which I doubt the controller has, since it's unnecessary.
QuoteIf the sensor bar were used in that way, you would notice that up/down orientation was much less accurate than right/left. If you have two receivers on the remote, you can use compare four imputs - two to the front of the controller, two to the back - and get an accurate orientation around two axes, which is all you need for pointing.
Exactly. Triangulation from only two points is not sufficient to determine location in 3D space - hence the need for both accelerometers and the sensor bar. Whether this is a problem in practice though remains to be seen.
QuoteCompletely wrong, but points for confidence. In order for the controller to sense its distance to within, say, a tenth of a meter (the maximum distance I would say you could use and still have it be useful for gameplay) the controller and sensor bar would both have to have clocks accurate to within a third of a nanosecond, or close to the precision of the most advanced modern atomic clocks. And you would have to synchronize those clocks to start with. So unless the Wii comes with a personal industrial physics lab, absolute signal time will not be used to calculate distance.
if you turn the controller away from the sensor bar it cannot detect depth...it detects depth by measuring the amount of time it takes the ir to get to the sensor. kinda like echo location..but with light.
Post by: The Omen on May 20, 2006, 08:36:32 AM
Post by: Jensen on May 20, 2006, 09:16:56 AM
Quote
Originally posted by: zakkielQuoteQuote
Depending on its position, the bright spots from the IR LEDs will show up in different spots on its CMOS sensor.
That would require sophisticated optical equipment which I doubt the controller has, since it's unnecessary.
That's basically how an optical mouse works, nothing new...
The article has some stuff that I haven't read of before, such as the mention of multiple accelerometers in each controller, but there are some errors in the article, so I can't really believe any new information.
Post by: zakkiel on May 20, 2006, 03:09:18 PM
QuoteNot at all. Do you mean 3d mice? If so, I have no idea, except that they weren't precise enough to succeed.
That's basically how an optical mouse works, nothing new...
Post by: Michael8983 on May 20, 2006, 05:34:16 PM
Post by: Kairon on May 20, 2006, 05:41:10 PM
But if you wanted to ask a pointer being held "classic" style what angle it's long axis makes with the screen, then you can't do that.
~Carmine M. Red
Kairon@aol.com
Post by: BlkPaladin on May 20, 2006, 05:51:18 PM
The problems with "going off the sensor bar" can be explained by simple software problems. How the Wii detects the controller isn't the problem its the problem with the software that translates it to in game movement. For each sensitivity settings the software has maximum threshold for movement as well as minimal thresh holds. When the Wii detects a gross movement and the software can makes heads or tails about how it should handle it it either wont or do something else. For example if you up to close to the bar and moved a lot with a high sentivity, the bar would read it like you were moving the controller 5 or more feet if the programming was not there, so normally if a reading goes over the maximum thresh hold the programmer whould make sure that the game whould ignore the input.
So most of the controller problems at E3 would more than like be programming issues because the programming for reading the controller movements haven't been fine tuned or flushed out to include reading how close the controller is to the bar and make automatic sensitivity changes.
Post by: Michael8983 on May 20, 2006, 07:20:08 PM
Can it detect which direction it's being thrust in?
If it can't, that's a major problem but if it CAN, I don't see it being very limited at all.
Like in a sword fighting game, it could detect vertical, horizontal, and diagnol thrusts.
On top of that, with the gyroscope it knows which angle the remote is being held which would allow the game to know the difference between a slash at an enemies legs or head. Then of course a stabbing motion with the sword would be performed with the remote pointed AT the screen/bar which would allow for even better accuracy. On top of that you could hold the sword outward and move it freely and precisely where you want it. Which was obviously the inspiration for the scene in Red Steel where you spare an enemy's life by holding the sword to his neck. You could even hold the sword/remote straight up or sideways and angle it however you want with just the gyroscopes. Perfect for fending off projectiles.
Someone correct me if I'm wrong about the remote being capable of any part of that.
Post by: mantidor on May 20, 2006, 07:30:43 PM
It has to, otherwise the minigame in warioware about balancing a broomstick wouldn't work.
Post by: Dirk Temporo on May 20, 2006, 08:38:29 PM
Post by: Jensen on May 20, 2006, 09:28:55 PM
QuoteWii doesn't use IR for data transmission. Excite Truck doesn't use IR at all. Only games that display a cursor use IR.
Originally posted by: BlkPaladin
In addition it doesn't use IR since IR needs line of sight, which is not possible when you are holding it sideways in games like Exite Truck.
Post by: Jensen on May 20, 2006, 09:35:56 PM
Quote
Originally posted by: zakkielQuoteNot at all. Do you mean 3d mice? If so, I have no idea, except that they weren't precise enough to succeed.
That's basically how an optical mouse works, nothing new...
Optical mice use a small camera to track the texture of a surface. The Wiimote uses a small camera to track LEDs. The Wiimote can calculate a little more info though, like how much the controller is twisted(roll axis), and how far the controller is from the sensor bar. (I hate calling it the sensor bar, as it isn't)
Post by: BlkPaladin on May 21, 2006, 04:38:16 PM
How it can work with just what we know about:
The accelerometer are used to measure direction and "force" of the movement. (And thus you could derive a sense of speed.) And the tilt of the controller.
The bluetooth is used for transimission plus to measure depeth and relitive position. The depth is found by using a radar gun type of approch, to futher this there are only four known settings for the controllers so each has a set timing interval and fequence using this the Wii can calculate how far the controller is from the bar. And if the controler is moving away or towards the bar.
Now here is why there has to be a bar (the wand). The bar is used to sense were the leading edge of the tranmission is (the first contact, a small sensor would not be able to do this well)
Post by: Athrun Zala on May 21, 2006, 07:00:14 PM
QuoteI'm pretty sure I read somewhere that the wiimote is calibrated/reseted/whatever when oyu simply hold it completely still.....
Originally posted by: Dirk Temporo
More importantly than anything else in this thread, why in the name of God don't any of the games shown so far have some sort of calibration. Point at middle of screen, click. Point at the four corners of the screen, click each time. Voila, it now knows where on the screen you're pointing. It can't be that hard to do. And that would completely alleviate the need to point at the sensor bar.
Post by: Kairon on May 21, 2006, 07:37:03 PM
Quote
Originally posted by: BlkPaladin
I doubt that. The power consumption required for what the controller has that we know about is phenomonal plus you can do all the required functions with just the acclerometers and bluetooth, there is no need for LEDs with cameras because then the wand would be useless. (And not to mention all we can do is speculate on how it works.)
How it can work with just what we know about:
The accelerometer are used to measure direction and "force" of the movement. (And thus you could derive a sense of speed.) And the tilt of the controller.
The bluetooth is used for transimission plus to measure depeth and relitive position. The depth is found by using a radar gun type of approch, to futher this there are only four known settings for the controllers so each has a set timing interval and fequence using this the Wii can calculate how far the controller is from the bar. And if the controler is moving away or towards the bar.
Now here is why there has to be a bar (the wand). The bar is used to sense were the leading edge of the tranmission is (the first contact, a small sensor would not be able to do this well)
Why does the bar emit light then?
Oh, and what radar gun is this anyways? I thought Radar guns measured speed, not distance!
And does signal triangulation work on such a small scale or is it innacurate at this level?
And finally, how does your theory fit in with Will Wright's comments that the controller functions differently at 5 feet than at 10 feet, or the comment I heard from Alain Tuscan, General Manager for EA Montreal, that the controller only has a small cone (~30 degrees) of pointer functionality and that if someone walks in front of you then it doesn't function?
~Carmine M. Red
Kairon@aol.com
~Carmine M. Red
Kairon@aol.com
Post by: Dirk Temporo on May 21, 2006, 08:00:21 PM
Quote
Originally posted by: Athrun ZalaQuoteI'm pretty sure I read somewhere that the wiimote is calibrated/reseted/whatever when oyu simply hold it completely still.....
Originally posted by: Dirk Temporo
More importantly than anything else in this thread, why in the name of God don't any of the games shown so far have some sort of calibration. Point at middle of screen, click. Point at the four corners of the screen, click each time. Voila, it now knows where on the screen you're pointing. It can't be that hard to do. And that would completely alleviate the need to point at the sensor bar.
No, I mean so that where you point on the screen is where the cursor is. As it stands now, to put the cursor in the middle of the screen, you have to point at the middle of the sensor bar.
Post by: Jensen on May 21, 2006, 08:05:25 PM
But Radar does measure distance. By taking multilple distance readings over time, a radar gun calculates speed.
But the Wii doesn't use anything like that.
Post by: Jensen on May 21, 2006, 08:41:07 PM
Quote
No, I mean so that where you point on the screen is where the cursor is. As it stands now, to put the cursor in the middle of the screen, you have to point at the middle of the sensor bar.
Point what at the middle of the screen? The controller has no sights, so how do you know you're pointing it at the middle of the screen? Do you hold it up to your eye and line it up as well as you can?
Let's say that you pick some arbitrary orientation that you think feels like pointing at the middle of the screen. Now you've just further limited the range of motion you can make before the Wiimote points out of range.
Here's a test if you have a laser pointer. Stand 10-15 feet from a target. Hold the laser pointer at waist level, pointing towards what you think is the target. Now turn the laser on. Your aim is probablyquite a bit off.
I do understand what you are explaining, I just don't think it will matter in practice.
Post by: wandering on May 21, 2006, 08:57:54 PM
Quote
Originally posted by: Dirk Temporo
More importantly than anything else in this thread, why in the name of God don't any of the games shown so far have some sort of calibration. Point at middle of screen, click. Point at the four corners of the screen, click each time. Voila, it now knows where on the screen you're pointing. It can't be that hard to do. And that would completely alleviate the need to point at the sensor bar.
I'm pretty sure people needed to point at the sensor bar in some demos only because they were so close to the tv. Farther back, you should be able to just point at the general area of the screen.
Post by: Bloodworth on May 21, 2006, 10:45:05 PM
Post by: ThePerm on May 21, 2006, 11:32:26 PM
Post by: mantidor on May 22, 2006, 04:56:43 AM
Quote
Originally posted by: Bloodworth
I never had to think about where I was actually pointing. You see where the cursor is and move relative to that, just like a mouse.
Yes indeed, I couldnt understand all this talk about pointing to the sensor bar, its a relative positioning system like a mouse, so I didnt understand all this people having problems. Actually a mouse is harder, mapping horizontal planes to vertical planes isnt something humans can do on the fly, we are just already very trained about it.
Post by: BlkPaladin on May 22, 2006, 04:38:24 PM
Post by: Jensen on May 22, 2006, 06:40:07 PM
It is the combination of several features that worries me. The Wiimote feeding power to the nunchuck (or classic controller), rumble, sound, IR camera, at least two accelerometers, 2-way bluetooth comunication.
Post by: wandering on May 22, 2006, 08:08:36 PM
I'm expecting at least 20 hours. 2 whole Luigi's Mansion playthroughs!
Post by: zakkiel on May 23, 2006, 06:16:37 AM
QuoteNo. Radar guns use the Doppler effect. They take no distance readings.
But Radar does measure distance. By taking multilple distance readings over time, a radar gun calculates speed.
An optical mouse is taking a series of snapshots of the surface right underneath it; very narrow angle and close distance.
