How will the Revolution sense the controller?

[_xxx_]

What I am thinking is, how easy would it be to throw of the calibration of the controller by simply changing the rev-mote to a different position either by standing up, stepping a step either to the left or right... It may not make a big difference, except for games that require very high precision...

These sensors are very accurate, much more so than the hand moving them.

Oh, and Shifty: the 10-20$ range might be for a single consumer buying these in shops, but in a huge bulk and as a low-cost version (which it will undoubtfully be) it's more like 1-2$ a piece.

 

[Phil]

I know _XXX_ which is why I am pointing it out. If the precision of the input device is too high, it's also more prone to errors. As an example, I was wondering what kind of an effect moving the position of the "rev-mote" would have on the game itself -> would it assume the action was triggered on purpouse or by accident by the player?

To be more specific, I was wondering if a light-gun-game would be possible with it and if yes, what some of those events noted above would affect the game and how, especially considering the input-device is a high-sensitive one...

 

[NANOTEC]

The revmote contains 2 gyros..one at the front and one at the back...

It also contains a camera..that's actually the secret that will be revealed at E3...

 

[_xxx_]

To be more specific, I was wondering if a light-gun-game would be possible with it and if yes, what some of those events noted above would affect the game and how, especially considering the input-device is a high-sensitive one...

You can filter the signal just like any other, apply some value tables, interpolation, weighted mean, low-pass or whatever. But that is surely a very important thing, I bet they'll invest lots of time into fine-tuning that for all the possibilities.

EDIT: and if you were purely referring to gameplay, you can for example choose to get the input from the revmote only where needed and otherwise ignore the movement. Like, play that Jedi game normally and only get the input from the controller when light sabre fight occurs or some such.

These sensors are not directional, just the relative movement is measured, so changing the player's position should have no real influence.

 

[Squeak]

What about a baseball came where you're pointing away from the screen ready to bat?

Well you wont be able to do that, just like a mouse won't work outside the desk surface.
But the controller will still be able to measure tilt and yaw, so driving games and golf games will still work, but no baseball, ever.

the only thing missing is altitude or Y position of the controller, which I guess would be worked around using controller tilt to determine the input you'd associate with this. eg. Instead of lowering the controller in a snowboarding game to duck under a branch, you'd tilt the controller down..

I hope you can hear how stupid that sounds yourself. One kind of measuring for x and another for y!?
How well do you think that is going to work? Esp. in games that require high precision like fps. shooters?

 

[pc999]

Well you wont be able to do that, just like a mouse won't work outside the desk surface.
But the controller will still be able to measure tilt and yaw, so driving games and golf games will still work, but no baseball, ever.

But there is basebal (at least in the TGS video).

 

[Squeak]

But there is basebal (at least in the TGS video).

I. WAS. KIDDING.

 

[OtakingGX]

I. WAS. KIDDING.

About everything, or just the baseball comment?

I'm with Ingenu and Shifty on how it would work. A transmitter at the front and rear of the revmote. You do need three points for triangulation. Nobody ever said they had to be stationary points. The third point being in the remote would always be a set, known distance from the fourth. It's a lot of math to do by hand, but processors chew through 4x4 matrices like you do through simple algebra.

I'm almost positive that IR is not used for communication. Bluetooth would be cheaper, more reliable, and much faster for transmitting data from the controller to the console. It's probably just there for on/off functions and may be useful for light gun style play.

 

[Squeak]

About everything, or just the baseball comment?

I'm with Ingenu and Shifty on how it would work. A transmitter at the front and rear of the revmote. You do need three points for triangulation. Nobody ever said they had to be stationary points. The third point being in the remote would always be a set, known distance from the fourth. It's a lot of math to do by hand, but processors chew through 4x4 matrices like you do through simple algebra.

I'm almost positive that IR is not used for communication. Bluetooth would be cheaper, more reliable, and much faster for transmitting data from the controller to the console. It's probably just there for on/off functions and may be useful for light gun style play.

Wow, just wow.
Now you are kidding right? Right?!

 

[OtakingGX]

Wow, just wow.
Now you are kidding right? Right?!

No, I'm not. I'm convinced that no information will be transmitted via infrared. Why? Well, let's look at the specifics of infrared communication:

IrDA 1.0 allows for data transfer rates of 115.2 kbps at plus or minus 15 degrees at a distance of 1 meter. IrDA 1.1 allows for a greater bandwidth of 4 Mbps. While IrDA 1.1 does have more bandwidth than Bluetooth (721 kbps) it is rather restrictive in use. It is designed for communication between two devices (not five). It also has the angle dependency. Not only must your controller face the sensor, but it must be within 15 degrees of parallel to it. This wouldn't be a problem when pointing at the TV screen from far away, but it would be if your sensor bar were below the TV and you were within the maximum distance of 1 meter (3.3 feet) from it.

Now that we've established that IR can in no way be used for data transfer, how is it that the sensor bar can locate the remote? As pointed out, you need three points to locate a fourth. If you have two radar stations you can only locate a target on a large circle. Adding a third station will resolve the location of the target to two possible points. The sensor bar ostensibly only has two sensors though (one at each end). If a third is required for location of the remote, where could it be? We're guessing it's also in the remote.

 

[Squeak]

Now that we've established that IR can in no way be used for data transfer, how is it that the sensor bar can locate the remote? As pointed out, you need three points to locate a fourth. If you have two radar stations you can only locate a target on a large circle. Adding a third station will resolve the location of the target to two possible points. The sensor bar ostensibly only has two sensors though (one at each end). If a third is required for location of the remote, where could it be? We're guessing it's also in the remote.

Where did anyone suggest that the IR window is used for the datatransfer? The discussion was about whether it was used as an emitter or receiver of an IR positioning signal.
It would be very weird and expensive to have an IR receiver/emitter just for the hell of it, like you are suggesting, it has to play a very important role.

What you are suggesting with the radio wave triangulation, which takes advantage of doppler effect and/or delay, might work on a large scale, but is impossible or prohibitively expensive for consumer equipment.
And even if they did do this, why not just have the third sensor seperately (in the Revolution even), instead of wasting battery life?

 

[Ty]

How much data do you really need to transfer?

Sure the sampling rate will be high but the width would seem to be pretty thin.

 

[ninzel]

hehe looks I finally asked a good question.

 

[Teasy]

Single chip Gyros and accelerometers bought in bulk would cost in the region of $10-20 I expect, looking at publicly available prices of components.

Also I doubt Nintendo would have to buy at a normal bulk price anyway given that they own a large stake in Gyration inc.

 

[_xxx_]

No, I'm not. I'm convinced that no information will be transmitted via infrared. Why?

I don't think it'll be IR. IR is directionaly bound and could never allow you for any movements outside the range of +/- 20° or so. Not even mentioning the lag which would probably occur.

I'm pretty sure it'll be done the way I described it, it's the only technical solution which would make sense.

While I'm not 100% sure what they will do, I've been working with yaw rate and accel sensors for years, so I think I know what I'm talking about

 

[Bohdy]

IR bounces...

Depeding on the sensetivity, a sensor could pick up IR input from anywhere in the room.

 

[[maven]]

Back in the day after the announcement, there was some discussion on that here.

 

[_xxx_]

IR bounces...

Depeding on the sensetivity, a sensor could pick up IR input from anywhere in the room.

Maybe, but the sender side would have to be a real monster. Won't work.

 

[_xxx_]

Back in the day after the announcement, there was some discussion on that here.

Oh, ultrasound sounds nice. This also makes sense. Hmm, now I'm curious...

EDIT: but I still think it'll use yaw/accel sensors as well, maybe all this combined

 

[ninzel]

Well after thinking about it more,I think they will try and make the bar itself do most of the work. The more tramsmitting and sensing they can offload to the bar the better for reliability and battery life.