'Wii' stumped: the mystery of the optical sensor

whats an optical sensor?

it lets the game know where the thing is pointed at, otherwise all you can know is motion and speed of the controler but not where it is releative to anything.

thx

The Wiimote does not sense x, y, and z positions all on it's own... only x and y.

While I believe it still has an accelerometer for the Z access (depth), the optical sensor is able to determine exactly how far away the wiimote is.

Like T-Man said, it's probably so the console is aware of where the controller is relative to itself. Without it, it has no real way of knowing the distance between the Wii and the controller.

how could you know where "you" are relative to in-game objects if you don't have a point of origin? sure, you can sense changes in x, y, and z....so what? those changes in position are completely arbitrary. Without a point of reference, they'd be useless.

The sensor bar lays out an infrared grid and lets the Wii sense depth (aka distance from the console). It does not receive any data.

I think the optical sensor is designed to give the system a frame of reference as to the current location of the sensor. The accelerometers and gyroscopes can tell you all about the acceleration of the device, and from that can determine its velocity, but the initial condition is unknown, so you couldn't determine its exact position in space, in fact, it could conceivably be anywhere (within wireless range). The optical sensor probably sets this initial condition, just my guess.

it has no concept of the starting position of the remote, only the movement. at least that's what a lot of people have reported. so when playing warioware, you don't have to hold the remote over your head to do the weightlifting game. if you started around backwards, how would the controller know you weren't pointing it at the screen?

this likely also means that with the sensitivity on certain levels, it's probably possible to overflow the movement counters by moving the remote incredibly fast, meaning that it couldn't properly track the location.

the optical portion is required to place the front in relationship to the screen for those reasons.

I imagine for a game like DuckHunt, the optical sensors are needed... It needs a fixed point to calibrate to.

For example, if the wiimote detects a 15degree angle when you point at the tv sitting down... how can it tell the difference between that and some one standing up. The second (using the wiimote at a different angle) person pointing at the screen will be a different angle.

Why not, With the sensor BAR, which is probably a whole string of multiple sensors, it should be able to triangulate exact position and orientation of the wiimote in 3D space. The most likely reason for the optical sensor for the pointer is that, while the triangulation is acurate enough to detect changes in orientation of the wiimote, it probably isn't accurate enough to predict the exact alignment to work as a lightgun.

Or it could be there for redundancy. Who knows.

As near as I can figure, the sensors inside the controller determine only the rotation of the actaual controller. So you get a roll, like a plane doing a barrel roll, pitch, like a plane going up and down, and yaw like a plane twisting(think flat serface, rottating like a clockhand). These sensors do nothing to detect the physical location of the controller, just its angle. This is essentially the PS3 controller.

To detect location, you need a reference point. That's where the optical(IR) sensor on the Wiimote combined with the sensor bar comes in. The sensor bar will have at least 2 sensors(or transmiters-not sure which), one one either end, facing the player. With the 2 points on the bar, and one one the Wiimote, you can triangualte the location(geometry time, people!). That's it!

The sensor determines movement on the z-axis, so the Wii can sense you push forward or pull back. That's what separates it from the PS3 controller.

all that notwithstanding, doesn't the wii-mote also act literally as a remote control, turning the console on / off (or standby that is)?

It does, jojonyc, but that could be done over Bluetooth I assume...

Fuzz is right, though, it's the IR triangulation that requires the sensor bar and optical sensor on the Wiimote.

Look at that, mystery solved in less than half an hour thanks to the wonders of blog posters! Good job, guys!

Actually, it's so you can point it at a mirror and cheat at duck hunt. How else were you supposed to beat your friend's high score?

Slightly Off Topic: My big screen is the type that has the speakers on the bottom so there is about 1-2 feet between the actual screen and the floor, what i'm wondering is how well is this going to work with so much space between my sensor bar and the actual "game area"? I mean I understand that it will work, regardless, but I don't want to have to point at the speaker to do something on the bottom of the screen.

I think it's been pretty well covered here, but to summarize:

The controller itself senses rotation, the optical sensor if for position relative to the TV. So, if the wiimote moves left or right, forward or back, up or down, all without rotating, only the optical sensor can pick that up accurately.

As far as I can tell, this sensor bar is the difference between the Wii and the Ps3. Each controller (as far as I know) has both lateral and rotational accelerometers, 3 axis for each, giving the 6 axis sensitivity quoted by Sony. With just this setup it would be possible to have pointer style play, setting up the conroller at the start of each game with a g-con style calibration screen, but i suspect that it would not be terribly accurate and would probably slowly fall out of synch as the game progressed.
The addition of the wii sensor bar would be there for the reasons of increased accuracy, redundancy, and ease of setting up so that little or no calibration is needed.

Maybe it's to map peoples faces onto characters and other nifty eyetoy things.

http://www.thegamefeed.com/storage/images/News/WiiReggie.jpg
http://www.thegamefeed.com/storage/images/News/WiiIwata.jpg
http://www.thegamefeed.com/storage/images/News/WiiMiyamoto.jpg

It's not used for position sensing. It's used to determine the size of the screen you’re pointing at and to know when your pointing at it. That's my guess anyway.

I think this article will clear things up a bit:

http://www.xgaming.com/newsletter/Wii%20Dupe.shtml

The wii-mote uses accelerometers and a gyrometer to determine how you're holding it and how you're moving it, but it doesn't actually transmit a location in 3d space. Because of this, the optical sensor is needed for any game where you have to point at the screen. For games that just use motion sensing, you can actually turn the optical sensor off.

But that's only true of the hardware shown at E3. The article notes that Nintendo may be working on a true 3d tracking system for the Wii, so by release time it could be different, and thus the sensor bar might not end up on the final unit.

I'm the accelerometer. I just got back from some Wii-controlling. It was great. I laid out an infrared grid and the controllers just surfed along it. I'd like to move elsewhere, but there's nowhere that's really accommodating.

Thats a good thought about triangulation guys. The only problem is a triangle is only a 2D shape. So having a wide sensor bar would only be useful for see changes left/right and fwd/backwards. In order to "triangulate" position in 3D space you would need an additional sensor, so that overall shape would be that of a pyramid. And it doesn't seem like a long flat bar would pull off a pyramid very well.

The sensor bar lays out an infrared grid and lets the Wii sense depth (aka distance from the console). It does not receive any data.

Posted at 4:37PM on Jun 14th 2006 by vidGuy
-------------------------------------------
Exactly right. The position and movement of the remote in the IR grid is sent by the remote to the console direct(via blootooth)

its for calibration of location

To repeat everything everyone said and clarify more (I hope)...

I have no idea what sensor the wiimote does or does not contain, but it is possible for it to, at most, have 3 rotation sensors and 3 accelerometers, or some combination that can acheive the same effect.

This can tell precisely which way is down, relative motion on 3 axis, and relative rotation on 3 axis, all without the sensor bar.

From the pictures from E3, you can see the bar has two groups of 2-3 IR LEDs at either end of the bar. From the sensor bar, it can determine exactly where you're aiming on the screen (if the wiimote camera sees the center of the LEDs 2" down, 5" left, then the wiimote is pointing at 2" up, 5" right), and roughly how far away you are (using the distance between the groups of LEDs).

By using a combination of the forward tilt and where you are aiming at the screen, it should be possible to triangulate how high above or far below the tv the wiimote is to an accuracy of probably no more than (entirely my random guessing) half a foot.

Unless the wiimote camera is crazily high-res, I do not think that the wiimote can tell if you're standing to the left or to the right of your television. The ONLY way I can think of to determine this would be to determine the relative percieved widths of the two groups of LEDs on opposite ends of the bar. (If the 2 or 3 left LEDs are 3 pixels apart, and the right LEDs are 5 pixels appart, you're standing a few feet to the left). This is possible, but I can't imagine it'd be accurate at all.

So I don't think the wiimote knows its *absolute* position in the world. But it probably knows enough to do anything you'd want to in a game. Relative motion is what's important, and it's most likely very very good at that.

I don't understand this pointless debate: the wiimote has 2 technologies - it's like the ps3 controller + a new technology. Everyone is happy with the possibilities of the controller compared to the one of the ps3.

Why now discussing about that? It's like saying, when Ford made it's first vehicule, "why can't you make it faster and more safe". Why? Because what you see is the maximum that our technology can do! (for that price)

its for using the wiimote as a replacement for your receiver/dvd/tv/electronics equipment remotes. duh!

Well Nintendo said it wouldn't be easy for their competitors to replicate all the controller secrets.

Odds are the apparent optical sensor is tied into preventing their competitors from duplicating the functionality of the wiimote.

The PS3 controller is just a tilt sensor that has been used in gbc games like Kirby Tilt and Rumble.

PS3 contoller cannot be used for 95 of the games coming out for the Wii. You couldn't do cooking games or most dogfighting games or even a lot of racers.

Lets face it most PS3 games using the tilt will suck if you have to use any analog sticks or reach to press the buttons at the same time. The ps3 controller has been done before on consoles. Those controllers btw had rumble and the controllers were considered awfull to use.

What about the PS3? It is claiming to do the same as the Wiimote but does not come with an eye toy! There are still some mysteries about Wii Nin has not let out of the bag yet!!?!

I thought about this, too and my conclusion was, that there really is nether real positional-information available, nor triangulation going on.

Acceleration is a derivate of the movement but this accelerometers don't have enough precision to integrate positional-information. And triangulation can also be excluded, because the sensor-bar is in essence a “line” and you need at least tree linear-independent points to triangulate. From that i would guess that the sensor-bar only contains two or three IR-LEDs and the optical-sensor works similar to an optical-mouse, but only sees this IR-light. With that you can sense where the Wii-mote is pointing, as long as the optical-sensor “sees” the sensor-bar. If this setup is calibrated -- so you know where the screen bounds are in relation to the light of the sensor-bar -- you can calculate pointing-information from the position of IR-LED “dots” in the image the optical-sensor sees.

This is a bit sad but on the other hand all the demos we saw from Nintendo, are possible under these assumptions.

Absolute vs Relative

Heh

I am so enjoying this..
People consider the sensor bar an annoyance because it has to be positioned below their tv screen in order for the wiimote to function (though I've seen some nifty footage from E3 and I didn't spot the sensor bar) and whine about it, but they tend to forget that Wii is considerably more compact than the other 2 consoles, a bit larger than 3 DVD cases stacked together (my collection of the Simpsons season 3 takes up more room than that)

At anyrate the original post forgot to mention that the wiimote isn't just motion sensitive..

It also rumbles, has the imbededded speaker, can aslo brew coffee and vacuum the floor (j/k..

Anyway I have a bluetooth phone and once I tried sending via bluetooth a mp3 (a 1.2MB) to a friends device, it took it's sweet time..

I think that the optical sensor is there to make the controller more responsive,given the amount of data that have to be conveyed between controller and Wii which are considerably more than what the wireless Xbox 360 controller coneys or what the PS3 conrtoller coneys..

The sensor bar doesn't know where the controller is relative to the TV, it only knows where the controller is relative to the sensor bar. Perfect calibration would require you to point the controller at the screen while the sensor bar reads the controller's distance from the sensor bar as well as the pitch, roll, and yaw of the controller. It can then determine the position of the sensor bar relative to the TV, and therefore the position of the controller relative to the TV even when the optical sensor isn't pointed at the TV. It's true that it cannot determine the Y value of the controller in space (atleast I don't THINK it can) but it can determine the Y value relative to the TV. Combine this with vertical accelerometer input, and it should have all the information it needs to lead an army of killer robots... or something.

You're all correct, but some of you are slightly stuck on one bad point. The sensor bar is a sensor bar and not a ir light series. The "led's" are either a series of photodiodes or leds acting as photodiodes (which led's can do!). They are taking information in from an ir led in the wiimote! The wiimote is not a processor and there is no reason to make it try to calculate triangulation when the sensor bar is wired directly to the console.

it's quite simple: The accelerometers provide relative positioning. There are a few problems with this though. It will work great for something like Excite Truck where you want to know what angle the controller is to the floor, but if you want to use it for absolute position (i.e., as a light gun) it won't work well. First you'd have to calibrate it EVERY TIME you played a game. Then, the errors (while very tiny) would build up over time (drift, bit errors, etc) until pointing at the screen wasn't pointing at the screen.

Now the light sensor fixes that. As some have already mentioned it can be used to triangulate depth, but the bar below the TV also provides a static reference. Thus you can calibrate the controller, and just keep that calibration (like the DS does) until the TV or sensor bar are moved.

Now while the sensor bar is nice, you can't have JUST the optical sensor bar. While you can point light gun style, with just the bar and no accelerometers games like Wii Sports wouldn't work because you would be pointing the controller away from the TV and thus there would be no positional data.

It's the combination of these two sensors that allows the Wii to pull off it's cool controller tricks, and what set's it apart from what the PS3's controllers can do.

#22 - Sam. THANK YOU for actually referencing something factual. So much speculation gets thrown around in the comments that is SOO far off base, this thread in particular.

I'd give you TWO stars, if I could. You've also provided enough closure, I can move on with my life and out of this article now. Heh heh.

If your really interested in finding out what the heck the purpose of the infrared thing-a-ma-bob, I think it would be best for you to look at Nintendo's patent on the device. (sorry no link)

As far as I can understand, the big innovation hit was that of resources. They could have included as many chips x,y,z...f,g,h as they wanted in the thing to give it excellent accuracy, but 2 motion machines coupled with an infrared detector effectively tossed out some accuracy for a huge dip in processor requirements, which automatically means less battery weight, smaller size, and even free up the processor resources for some extra features.

Nintendo has taken what was only in concept a clunker of a device, and made it slim and sexy by using compromise. Additionally, they managed to find a way around doing a calibration dance everytime you want to play, something that I dont think would be possible without the IR grid.

(there is still a bit of calibration in use, but remeber, your joystiqs on your gamepad are calibrated automatically once-and-again to set a zero point. Ever have it go wrong on you and your character spins in circles until you restart?)

Although I do remeber not wanting to be seen in public with my virtual boy.....

> Takahashi wonders why this data isn't enough to determine where the pointer is positioned on screen.

Because even though the Wii knows where the remote is, it has no way of knowing where the screen is does it?

Simple really.

I think it's funny that Joystiq has posted this article without referencing the Xgaming info, when they said they would and that they agreed with the article which has been out since directly after E3.

http://www.xgaming.com/newsletter/Wii%20Dupe.shtml

The 'Sensor Bar' is just an infrared illuminator. The image sensor on the front captures an infrared image of the 'Sensor Bar'.

Given that the size of the sensor bar is known, with a fairly low resolution image from the Wii-mote, you can accurately calculate the distance to the bar, and the angle you're pointing at.

That's why Halogen lights might interfere with the Wii-mote, they emmit an awful lot of IR light that can effectively blind the Wii-mote image sensor, impeding the detection of the bar from the background image.

The same can happen if direct -strong- sunlight hits the image sensor.