Support NeoGAF

[PkunkFury]

OK, here is a great site showing how one can find exact position in 3D space using only 2 receptors. It is about the P5 glove which I mentioned only had two receptors in the sensor bar earlier, and it is an interesting read. I thought I remembered some of the revo impressions saying Nintendo placed two sensors by their TV, so I'm thinking they they are using similar tech.

http://www.mts.net/~kbagnall/p5/p5 dissassembly.html

Scroll about halfway down the page.

It appears the P5 sensor bar uses two receptor bars in order to triangulate and determine the distance from the sensors to the glove, but as mentioned before, you need 3 sensors to triangulate in 3D space. Therefore, the inclusion of two receptors is only being used to find the glove's Z coordinate.

Now here is the tricky/clever part. The receptors themselves are made up of 4 photodiodes, two for each direction being tracked. I gather that the photodiodes can read the amount of light from each LED in sequence. If the glove is slightly off to the side in the measured direction, one of those photodiodes will not receive 100% of the light signal (due to the ridge around the diodes occluding the light. Each direction has 2 diodes so that the machine can tell which direction the glove at based on which of the two diodes is missing light information. If both diodes receive 100% of the LED signal, the glove is directly in front of the receptor, and the hand is pointing straight forward. If the glove is off to the right, the diode on the left misses a percentage of the LED signal based on the glove's disance to that side, thus the computer can figure out how far to the right the glove is. Once the glove goes so far right that neither photodiode receives light, the glove is out of range.

The page explains it a little better, with pictures.

This is a clever way of finding the location of these LEDs in space, and judging by what we know of the rev controller, I'm thinking it also would work with the remote. As I said before, the P5 glove tracks orientation based off of the correspondence between multiple LEDs. I hope the Rev controller has an internal gyro to do this if it uses this unique tracking method, cuz it would make rotations a lot better.

 

[heidern]

Can it/do you think it can sense strength of motion?

Yes. It can detect you flicking the controller a couple of millimetres. It can detect you leaping from one end of the room to the other wildly winging the controller about. Say it's set to read at 120 frames per second. What will happen is every 1/120th of a second it will capture the position of the controller. As long as it is within range(which will probably be as high as 15-100 feet) it can detect ANY movement and positioning. With this thing you get TOTAL flexibility. It's just up to the developers to make use of it. Take baseball. You could do it mouse gesture style where the player flicks the controller to swing. Or you can go all out and get the player to stand and make a full swing.

 

[Shao]

No you don't get it. You can do this both ways. By giving the developer the positions of sensor 1, 2 and 3 the developer can interpret those any way he wants. If the controller is turned upside down the developer can account for that and invert the processing. Or not. The choice is there to work this any way you want. Because the interpretation is done in software you have total flexibility. And because this is the equivalent of processing one limb the hit on the processor is pretty much negligible.

I know what you're saying but I guess something just strikes me as odd since I'm not all clued up on the technology. I'm gonna assume the sensors are somehow assigned corresponding positions as follows.

Say we have the 3 sensors and in the default position i.e. flat and pointing at screen. Sensor 1 = forward, 2 = left and 3 = right. First of all how would the game know which sensor was which direction? If you designate the directions, this suddenly limits your options as said previously, if you don't - whats to stop the game mixing up directions?

i.e. I'm playing a racing game which involves pointing left or right to turn. Say I turn particularly sharply to the right. Sensor 2 is now facing forward, not left. Would it get confused? What if I boot up in that position? Would the directions be all wonky? Whats to stop the game thinking you want to play an FPS using the NES controller position?

I know this sounds fussy but this is something Nintendo would consider.

i.e. I'm sitting playing fishing, I decide I want to lie down on my right side instead of sitting. Now pulling towards my head would feel unnatural, so ideally sensor 2 would change from being left to up.

PkunkFury - are LEDs really a reliable gaming tool? I was always led to believe they were not very robust in terms of home entertainment, or maybe I'm thiking of IR.

 

[Sweedishrodeo]

I'm excited to see this implemented into RPGs. FF7 remake with battles carried out with virtual weapons controlled by the controller, as well as puzzle solving elements pushed to a whole new level ..... I would die happy if such a thing were to occur.

 

[DrGAKMAN]

If something gets in the way of the IR's LED's light path (kid brother jumping in front of the screen) wouldn't it be rendered unreadable? That's like a major limitation and why IR wireless controllers were so crappy in the past...so why use something so unreliable (IR) now for something that would require even more technological prowess?

I know in Iwata's keynote he pointed out that the window on the front of the controller used a new technology (he named it something...like DPD or PDP or something) to track movements, but certainly it's not IR, right? A newer form of RF is obviously the better bet and I'm not trying to guess what it is, but I don't think (just 'cos there's an IR-like window on the front of the controller) that means IR.

If there is any sort of IR, like someone mentioned before, it could just be for powering the system on & off or maybe for DVD controls.

 

[Gahiggidy]

This will prolly be the first console that Lik-Sang cracks open the controller before the main unit. :lol

 

[PkunkFury]

If something gets in the way of the IR's LED's light path (kid brother jumping in front of the screen) wouldn't it be rendered unreadable? That's like a major limitation and why IR wireless controllers were so crappy in the past...so why use something so unreliable (IR) now for something that would require even more technological prowess?

I know in Iwata's keynote he pointed out that the window on the front of the controller used a new technology (he named it something...like DPD or PDP or something) to track movements, but certainly it's not IR, right? A newer form of RF is obviously the better bet and I'm not trying to guess what it is, but I don't think (just 'cos there's an IR-like window on the front of the controller) that means IR.

If there is any sort of IR, like someone mentioned before, it could just be for powering the system on & off or maybe for DVD controls.

Thanks, this is the information I was asking for. I'm interested in finding out more about this DPD/PDP tech. I'm only speculating the infra-red or similar tech because the P5 glove is using that tech in a way that would be affordable for a consumer level game pad. However occlusion will be a big problem. Please note that ultrasonic tracking suffers from occlusion as well. I think the only tracking techs available that don't suffer from occlusion are magnetic tracking, which is very expensive (and very accurate) and mechanical tracking, which requires some form of mechanical robot arm like the haptic phantom device. Any way, don't be sruprised if you need to keep a path cleared between the remote and the sensors.

 

[DrGAKMAN]

Yeah I think he called it the "Direct Pointing Device" and he compared it to using a stylus on one's TV screen. But it also determines depth, rotation, tilting as well as pointing on a up, down, left, right plane.

My theory, again, is that there's 2 sensors in the bar and 2 in each remote-controller (one in the front end and the other in the back end). Plus there's a pedometer/accelrometer (sp?) as well as a tilting gyro detection device (in each remote-controller) for the plane of the remote-controller itself. I guess there could be IR involved that adds to the precision of aiming/pointing dirrectly at the screen for some games (hense why they've only shown the remote-controller pointing towards the screen & why the remote-controller will slide into the top of the retro shell (so as to keep some pointing function even when plugged into it)), but I imagine most of the motion tracking is done like sonar tracking between the sensors in the bar & each remote-controller.

 

[heidern]

i.e. I'm playing a racing game which involves pointing left or right to turn. Say I turn particularly sharply to the right. Sensor 2 is now facing forward, not left. Would it get confused? What if I boot up in that position? Would the directions be all wonky? Whats to stop the game thinking you want to play an FPS using the NES controller position?

That's the thing, each sensor has a unique id. So in your example the system knows the "home" position is Sensor1 being the closest. If you boot up with it to the right then Sensor2 is the closest. The software will recognise this and say to itself "if the controller was facing forward Sensor 1 would be the closest, but actually Sensor2 is closer so actually the controller is being pointed to the right". How it deals with it is up to the software. It could interpret Sensor2 as the forward sensor(i.e. calibrate it) or it could say Sensor1 is still forward but the player is pushing right. Your fishing example would also work. Like I said, if I'm right then you can basically do ANYTHING with this.

 

[PkunkFury]

Yeah I think he called it the "Direct Pointing Device" and he compared it to using a stylus on one's TV screen. But it also determines depth, rotation, tilting as well as pointing on a up, down, left, right plane.

My theory, again, is that there's 2 sensors in the bar and 2 in each remote-controller (one in the front end and the other in the back end). Plus there's a pedometer/accelrometer (sp?) as well as a tilting gyro detection device (in each remote-controller) for the plane of the remote-controller itself. I guess there could be IR involved that adds to the precision of aiming/pointing dirrectly at the screen for some games (hense why they've only shown the remote-controller pointing towards the screen & why the remote-controller will slide into the top of the retro shell (so as to keep some pointing function even when plugged into it)), but I imagine most of the motion tracking is done like sonar tracking between the sensors in the bar & each remote-controller.

The way intersense does ultrasound tracking is by emitting sound from the sensor bar which is recieved by microphones on the device being tracked. The time it takes to recieve the sound is what determines position. Nintendo's remote could be doing the same thing, but you would be able to see small microphone ports on the outside of the remote.

Check out the pic of the intersense wand here:

http://www.5dt.com/products/pis900.html

The little silver hole on the front corner of the wand is one of the microphones (it looks like a screw). Although I think sonic tracking is better at dealing with occlusion than infra-red, it is still a line of sight tracking system, and blocking the tracker with your body would cause the sensors to lose track of those little microphones.

I'm willing to bet Nintendo is simply assuming that people can't play games when something is blocking their view of the television anyway, so yes, your little brother could stand in front of you and you'd lose control (but how much control did you have anyway if you couldn't see the screen?), and when he moves away you'd get control back almost immediately.

 

[DrGAKMAN]

I don't think we should worry to much about how/if it works.

Two things Nintendo has already drilled into our minds before it was revealed was that it was going to change the way games are played and attract non-gamers. If it wasn't gonna change the way games are played (meaning, it wasn't gonna work the way they're saying/showing it will work) then they've staked their whole future on a lie that will be found out as soon as anyone plays it and finds out for themselves that it doesn't work. And if it's meant to attract non-gamers and be simple (Nintendo, also drilled into our minds that it has a quick start-up and is easy to use, etc.) than I doubt they're gonna ruine that philosophy with a technology that requires people who don't really even understand games to begin with to calibrate, set-up and hold their tongue just right to make it work.

After they've unveiled it they show a video, not of exaggerated unbelievable motions that can't really be done with this technology, but of things that can/will be done! Already, the early hands-on impressions tell us that much. And in interveiws they assure us that all we need to do is place the sensor bar near the TV and turn it on. They've told us that it works on any TV, doesn't need to be set-up or calibrated and no matter if you change seats while playing it'll still work.

They're putting their future out there and I believe them. There's no way they're lying about it working and how easy it is...otherwise they'd be found out pretty quick and they'd *really* be up shit-creek without a paddle when they're caught up in their own lies!

 

[PkunkFury]

That's the thing, each sensor has a unique id. So in your example the system knows the "home" position is Sensor1 being the closest. If you boot up with it to the right then Sensor2 is the closest. The software will recognise this and say to itself "if the controller was facing forward Sensor 1 would be the closest, but actually Sensor2 is closer so actually the controller is being pointed to the right". How it deals with it is up to the software. It could interpret Sensor2 as the forward sensor(i.e. calibrate it) or it could say Sensor1 is still forward but the player is pushing right. Your fishing example would also work. Like I said, if I'm right then you can basically do ANYTHING with this.

As long as three sensors are "visible" to the tracking system, software can determine a plane based on the coordinates and can figure out the exact orientation of the wand in 3D space (as well as the position). Basically, the software can figure out anything you'd like to know about the wand. However, 3 sensors need to be visible for this to work, if the software only gets data from sensor 1 it has to guess the position of sensors 2-7 (or however many are included) based on the wand's previous oreintation and use that guess to estimate the current orientatin. That is a bad situation for an intense game, therefore I'm hoping Nintendo has a gyro inside that remote in addition to multiple positioning sensors. That way if the sensors are failing the gyro can do the oreintation guesses.

I'm also thinking the wand may only have one position tracking point, and might rely completely on the gyro for orientation. I'm not sure if this would be a bad thing, as I know less about how gyros gather data and what makes them fail than I do about how position tracking gathers data. Based off of WarioWare, I'd imagine just a gyro would work well enough, but I did have calibration problems with Wario Ware in some of the mini games that lasted long. I think a checks and balances system using multiple tracking points as you have described, plus a gyro would really help make certain orienation isn't lost.

 

[Shao]

How it deals with it is up to the software. It could interpret Sensor2 as the forward sensor(i.e. calibrate it) or it could say Sensor1 is still forward but the player is pushing right. Your fishing example would also work. Like I said, if I'm right then you can basically do ANYTHING with this.

I see what you mean, still sceptical because it relies on the coding of the software but also think you could be right. But again if I'm on a sofa pointing straight but decide I wanna sit on the floor, and therefore the screen becomes abit raised won't it think that I am constantly tilting upwards? I suppose it could realise that my perspective has changed and adjust.

This problem I suppose also applies to my theory of seperate pointing and gyro. Mainly because there is supposedly no calibrating and the system does not know precisely where your screen is, only its direction and distance. i.e. how does it know where the screen is in relation to the sensors? Does it need to know? Think I'm gonna give up guessing!

Oh some people still believe this will use stereoscopic 3D! (inc. that broken saints guy)This is why Merrick has said that there are still more surprises to come and that the sensors determine your distance from the screen. Imagine a helmet/glasses attachment lol. Some guy on IGN is now 99.9% sure its 3D,

Rev real name revealed! Nintendo 3D!

 

[johns all like]

ok guys, this sensor is easy to program for.

it uses a gyroscope right? segways use gyros, all you do is write simple code to just to determine the location of the sensor with relation to another point. thats for the tilt mechanism. its a simple loop that anyone could write in C++. and i'm sure that the triangulation of the sensors for 3d space is just another for loop. with a processor dedicated to this, and it being simple calculations, this thing will have no lag. well, no lag as compared to the processor speed. i realy dont see what the big deal is, and i dont think nintendo does either. they said themselves its simple technology but the way they've applied it is revolutionary.

 

[PkunkFury]

ok guys, this sensor is easy to program for.

it uses a gyroscope right? segways use gyros, all you do is write simple code to just to determine the location of the sensor with relation to another point. thats for the tilt mechanism. its a simple loop that anyone could write in C++. and i'm sure that the triangulation of the sensors for 3d space is just another for loop. with a processor dedicated to this, and it being simple calculations, this thing will have no lag. well, no lag as compared to the processor speed. i realy dont see what the big deal is, and i dont think nintendo does either. they said themselves its simple technology but the way they've applied it is revolutionary.

I don't think it is the programming we are wondering about, but the method of detecting the position/orientation data itself. It's been said by a couple people here already that once you have 3 position points for the wand (or an internal gyro and one position point) you can do calculations to find how the wand is moving just as easily as a GPU rotates a simple triangle in 3D space. The reason you might experience lag is because of the speed and accuracy of the tracking system itself, which all depends on the technology being used. Motion tracking hasn't developed a consumer level standard yet, and multiple diffeent methods are currently being researched. Some track faster/more accurately than others. Don't forget that a gyro alone won't tell you where the remote is located relative to the sensors. If the remote were using just a gyro\, the sensor bar wouldn't be necessary.

 

[DrGAKMAN]

I know I just got thru saying not to worry and just be happy with the fact it works, but I read this over at Nintendo-Centrum:

Bluetooth is a way for two systems to share data; it's the radio equivalent to using a cable to link to devices. You're correct in thinking that the devices don't need line-of-sight in order to connect via Bluetooth. Infra-Red (IR) is the same sort of thing, but making an optical link instead of a radio link, implying that the devices do need line-of-sight. While your DVD and TV remotes probably uses IR to communicate, my guess is that's not exactly what the Rev. controller is up to...

The "hands-on" reports have stated that some sort of device was placed along the TVs that were running the Rev. demos. The IR port on the front of the controller is probably used in conjunction with that device to help fix the controller's location in 3D space. Imagine that there are two "IR eyes" on a strip along the top of your TV. Each Rev. controller could send a unique set of (very fast) flashes to identify it (allowing for more than one controller to be "seen" at once). The stereo eyes could now detect the controller's movement just like a human can put two eye's worth of information together to see in three dimensions. The job would be even easier if there were multiple IR sources on the front of the controller to help detect the distance from the screen--it would be like watching cars drive around a dark parking lot with their headlights on. All other input from the controller (button presses, left/right tilting, etc.) could be sent to the Rev. base unit via Bluetooth radio signals.

So, in the end, the controller's IR beam would only need to reach the television (or, more precisely, the device ON the TV) to work. The Rev. base unit could be tucked away out of sight, where the radio Bluetooth signal would still reach it.

It's quite possible that there are also accelerometers in the controller to sense the forces on the controller as it moves around. Cars with navigation systems currently use a combination of the Global Positioning System (GPS) and built-in accelerometers to track the car's position. The GPS updates the car's position a few times a minute, and the accelerometers are used to estimate the car's motion in between GPS updates. This also let's the nav system know where the car is when you're going through a tunnel (when you wouldn't be able to detect the signals from the GPS satellites).

The Rev. controller might use accelerometers in the same way. The video of the Rev. controller in action clearly shows people swinging the controller so that the IR port does not point toward the TV. By coupling a stereo IR system with a few accelerometers, the Revolution could track controller movements anywhere, much like a car's navigation system does. The accelerometer information would just be one more piece of data that is sent to the console via Bluetooth.

This has probably already been discussed, but I think BlueTooth + accelerometers/pedometers + IR + gyration *altogether* could be how this works. I once head that ComCast (who's working with Nintendo on the WiFi workings of the Revolution deck's connection to the internet) was also working on the communication to the controller. So Nintendo has alot of wireless types with Revolution...WiFi for internet/NDS communication, BlueTooth for controller communication & possibly IR for motion tracking and/or powering on/off the system remotly. All of these probably so that each can handle each of their duties without overworking the others. BlueTooth (since it would only be used to really communicate button presses, gyration, accelelerometer/pedometer detection) also opens up the opportunity for wireless shortrange headsets which would be nice. So, together with some form of IR, precision pointing is added on top of what BlueTooth & gyration is already doing. This really is alot of stuff, but I think all of these working in tandem is what really makes it so intuitive. In the hands of a non-gamer it's REALLY going to be a magic wand.

 

[johns all like]

I don't think it is the programming we are wondering about, but the method of detecting the position/orientation data itself. It's been said by a couple people here already that once you have 3 position points for the wand (or an internal gyro and one position point) you can do calculations to find how the wand is moving just as easily as a GPU rotates a simple triangle in 3D space. The reason you might experience lag is because of the speed and accuracy of the tracking system itself, which all depends on the technology being used. Motion tracking hasn't developed a consumer level standard yet, and multiple diffeent methods are currently being researched. Some track faster/more accurately than others. Don't forget that a gyro alone won't tell you where the remote is located relative to the sensors. If the remote were using just a gyro\, the sensor bar wouldn't be necessary.

i've heard it uses bluetooth to transmit the data which it could transmit the data from the gyroscope. but if it uses some sort of laser to sense position, if it emites light with a certain phase when directly in front of it, the sensors could maybe determine position by shifts in the frequency depending on where the light is emitted? the difference in the frequencys in the system and the sensor could be used to calculated 3 dimensional position by triangulation and the frequency shift.

 

[heidern]

I see what you mean, still sceptical because it relies on the coding of the software but also think you could be right. But again if I'm on a sofa pointing straight but decide I wanna sit on the floor, and therefore the screen becomes abit raised won't it think that I am constantly tilting upwards? I suppose it could realise that my perspective has changed and adjust.

This problem I suppose also applies to my theory of seperate pointing and gyro. Mainly because there is supposedly no calibrating and the system does not know precisely where your screen is, only its direction and distance. i.e. how does it know where the screen is in relation to the sensors? Does it need to know? Think I'm gonna give up guessing!

According to my theory it doesn't need to know. Arrgh, it's hard to explain. Let's see...

Regarding your first example here, again it depends. You could code it so that you read the controller in a literal way so that if you sat on the floor it would be constantly tilting up. You could also code it relatively. So say you boot up and standard is 0 degrees is forward, 90 degrees is up. If you code it absolutely and then sat on the floor it'll read it as 30 degrees up. In some games you may want to do this. In others you might want relative controls where you boot it up at 30 degrees which would be forward and 120 degrees(up and slightly back) would be up. Both are possible. See, the choice is up to the developer.

You could for example go upstairs and face the opposite direction to the tv and use it like that(assuming the sensors could work at that range). It's up to the software to translate the real world information into game controller co-ordinates. Example, all the keys on your keyboard face up. Doesn't matter if they're on the left or the right. You can work out their direction. Same applies, you can calculate relative distance at any point.

Also the key point is with recentering. Because it can work relatively in terms of acceleration, it could automatically recentre. You can stand up, sit down, lie on your side. The controller could face any direction, towards the ceiling, towards the tv, away from the tv and that could be interpreted as the centre, i.e. it is forward, forward, forward respectively. Or you could take it absolutely so that it is up, forward and backward respectively. The choice is up to the individual game. See there's no calibrating to the tv because it works anywhere. It can calibrates it dynamically in the middle of a game. If you sit down across the room it will automatically dynamically recalibrate.

 

[Gahiggidy]

The way intersense does ultrasound tracking is by emitting sound from the sensor bar which is recieved by microphones on the device being tracked. The time it takes to recieve the sound is what determines position....

Ultrasound?!

If so, then PUPPYS AM CRY.

 

[PkunkFury]

I know I just got thru saying not to worry and just be happy with the fact it works, but I read this over at Nintendo-Centrum:



This has probably already been discussed, but I think BlueTooth + accelerometers/pedometers + IR + gyration *altogether* could be how this works. I once head that ComCast (who's working with Nintendo on the WiFi workings of the Revolution deck's connection to the internet) was also working on the communication to the controller. So Nintendo has alot of wireless types with Revolution...WiFi for internet/NDS communication, BlueTooth for controller communication & possibly IR for motion tracking and/or powering on/off the system remotly. All of these probably so that each can handle each of their duties without overworking the others. BlueTooth (since it would only be used to really communicate button presses, gyration, accelelerometer/pedometer detection) also opens up the opportunity for wireless shortrange headsets which would be nice. So, together with some form of IR, precision pointing is added on top of what BlueTooth & gyration is already doing. This really is alot of stuff, but I think all of these working in tandem is what really makes it so intuitive. In the hands of a non-gamer it's REALLY going to be a magic wand.

Yeah, this is the theory I've been banking on. I know the Bluetooth is for communicating the data to the system itself (unless it is possible to use bluetooth for motion tracking, which I have never heard of), and I imagine tracking is light based as opposed to sound based due to the desgin of the controller. Still wondering about the two receptor comments for determinig tracking position; if the rev is using an approach similar to the P5 photodiodes. Also wondering how much cost adding gyros and accelerometers would add to the controller price, as all of the tracking could be done with just the trackers, only it would not work nearly as well.

 

[koam]

ok guys, this sensor is easy to program for.

it uses a gyroscope right? segways use gyros, all you do is write simple code to just to determine the location of the sensor with relation to another point. thats for the tilt mechanism. its a simple loop that anyone could write in C++. and i'm sure that the triangulation of the sensors for 3d space is just another for loop. with a processor dedicated to this, and it being simple calculations, this thing will have no lag. well, no lag as compared to the processor speed. i realy dont see what the big deal is, and i dont think nintendo does either. they said themselves its simple technology but the way they've applied it is revolutionary.

I strongly doubt you'd need to use a loop. Normaly, you'd just use a built-in #define that the Dev Kits provide you with to poll the info you need.

i.e

KEY_A
TILT_X


or whatever.

Nintendo wants to simplify the lives of their devs, i'm sure they've defined them in the Dev Kits, if not, you could IRQ poll them (which I doubt is the case).

 

[johns all like]

I strongly doubt you'd need to use a loop. Normaly, you'd just use a built-in #define that the Dev Kits provide you with to poll the info you need.

i.e

KEY_A
TILT_X


or whatever.

Nintendo wants to simplify the lives of their devs, i'm sure they've defined them in the Dev Kits, if not, you could IRQ poll them (which I doubt is the case).

ok i'm sure taht would work, but then again it would need to be accessing that information at so many times a second, at a certain step size, which i think would need some sort of looping function. its not complicated by any means, a guy in new york built his own segway with a simple loop, nothing fancy.

 

[PS2 KID]

Just thinking out loud: I wonder if RFID chips are accurate enough to gauge distance, speed, and xyz location.

 

[maharg]

(unless it is possible to use bluetooth for motion tracking, which I have never heard of)

Any time you have a radio signal (or really any signal of any sort) you can figure out position to some extent. All positioning technology is based on some form of triangulation.

 

[PkunkFury]

Any time you have a radio signal (or really any signal of any sort) you can figure out position to some extent. All positioning technology is based on some form of triangulation.

Yeah, I know how to do the math to get the data from the positions, what I'm saying is that I've never seen a device that specifically uses bluetooth to register the distance from a remote to a receiver. It's possible I have and I'm just calling it ultrasound tracking. I guess the problem I have is that I'm not sure exactly how bluetooth works. Bluetooth is basically just radiofrequency transmission correct? So if the remote used bluetooth for position tracking, would the remote need a microphone input like the intersense tracker I posted above, and would it need to be line of sight to track properly?

Most of the articles I'm finding relating bluetooth and motin tracking are describing blue tooth as being used to transfer the data, not to pin point any location information.

 

[Jive Turkey]

This will prolly be the first console that Lik-Sang cracks open the controller before the main unit. :lol

Holy crap I never even considered that. Has Nintendo found the best way to foil pirates to date?

 

[Gahiggidy]

Yeah, I know how to do the math to get the data from the positions, what I'm saying is that I've never seen a device that specifically uses bluetooth to register the distance from a remote to a receiver. It's possible I have and I'm just calling it ultrasound tracking. I guess the problem I have is that I'm not sure exactly how bluetooth works. Bluetooth is basically just radiofrequency transmission correct? So if the remote used bluetooth for position tracking, would the remote need a microphone input like the intersense tracker I posted above, and would it need to be line of sight to track properly?

Most of the articles I'm finding relating bluetooth and motin tracking are describing blue tooth as being used to transfer the data, not to pin point any location information.

Well, companies are already using WiFi, a claose cousin of BlueTooth, to do gloabal positioning systems for areas within cities.

http://www.infomaticsonline.co.uk/vnunet/news/2138522/wi-positioning-system

 

[sc0la]

I feel like a full 2/3 of the people in this thread have no clue what they are talking about.

 

[DDayton]

I'm fairly sure the Revolution controller won't be using IR signals for any purpose... as that would result in the Revolution controller being less capable than the Wavebird. Nintendo is still in the hands of a bunch of software and hardware geeks -- they aren't going to remove the "play anywhere" functionality. Radio signals make more sense.

 

[SpoonyBard]

Well, companies are already using WiFi, a claose cousin of BlueTooth, to do gloabal positioning systems for areas within cities.

http://www.infomaticsonline.co.uk/vnunet/news/2138522/wi-positioning-system

Not very global then, innit?

 

[sc0la]

I'm fairly sure the Revolution controller won't be using IR signals for any purpose... as that would result in the Revolution controller being less capable than the Wavebird. Nintendo is still in the hands of a bunch of software and hardware geeks -- they aren't going to remove the "play anywhere" functionality. Radio signals make more sense.

exactly. that is why I think half the people in this thread are retarded. If anything the IR is used for the power on function of the system, because that thing isn't on the front of the controller for nothing. The system might not be able to receive the controllers normal signals in a powered down state.

 

[PkunkFury]

exactly. that is why I think half the people in this thread are retarded. If anything the IR is used for the power on function of the system, because that thing isn't on the front of the controller for nothing. The system might not be able to receive the controllers normal signals in a powered down state.

Well I posted info about the P5 glove to show that IR might be a safe bet for the rev controller, since the P5 glove is an example of IR tracking at an affordable price. I realize ultrasound can be used for tracking, too, but I'm not seeing the microphone slots we see on the Intersense controllers or on the good old power glove on the rev controller, which is why I've been thinking IR is the asnwer.

The Wi-fi article is intriguing, but I feel the difference in tech required for tracking something within 20-40 meters will be a lot different then tech required to track the exact position in a 10 foot square area (or whatever the rev can track)

I'm mostly curious about occlusion, and where the data is being sent between the receptors and the remote. This is a hard call since we haven't seen any receptors yet, but the remote has one section which appears to be designed for IR input, and no indication of microphone slots or speakers for ultrasound.

I'll agree that bluetooth could be doing the tracking, but I'm curious if a blue tooth device can be tracked through the remote's casing, when most ultrasound tracking devices can't deal with occlusion. The window on the front of the remote doesn't appear to function with sound.

 

[SpoonyBard]

Bluetooth is just 2.45 GHz radio signal, so it doesn't need any visible windows in the controller's casing. I'm starting to believe that the window in the front of the controller is just for the power button.

 

[PkunkFury]

But just because my laptop can talk to my phone via bluetooth does not mean my laptop can figure out my phone's position in 3D space does it? I'm fairly certain that actual position detecting using sound requires an unobstructed path between the transmitter and the receiver; otherwise the obstruction will affect the speed of the sound which results in questionable positioning data. Here is a link that specifically describes the technology behind the Intersense I-900 system:

http://www.isense.com/company/papers/IPT_EG_3D_Interactive_Devices.pdf

I haven't finished it yet but it is a great read if you are interested in this type of tech.

Of particular interest is this section where they talk about wireless tech:

"Bluetooth was able to offer enough bandwidth and enough channels while operating in the globally available 2.4 GHz ISM band, but it was unable to meet the stringent synchronization requirements for making precise ultrasonic time-of-flight range measurements. In the end, an off-the-shelf 2.4 GHz frequency-hopping spread-spectrum telemetry card was modified to provide dedicated channels for real-time synchronization while maintaining the robustness of spread spectrum frequency hopping to assure compatibility with other 2.4 GHz wire-less systems (such as Bluetooth and 802.11b) that might be used in the vicinity of the tracker."

Here they mention considering bluetooth for wireless, but not using it because of difficulty synchronizing the signals with the rest of the tracking system. However, they are referring to transmission of the data gathered by the wand/tracker to the software, and not to the actual method of determining position in space. Position, as I described earlier, is determined based on the time in which the microphones on the wand/tracker receive the ultrasonic chirps from the SoniStrips (or the sensor bar in the revolutions case). If they considered blue tooth for sending the data from the remote to the software, why not consider bluetooth for position tracking? My guess is that it just doesn't work (though I don’t know enough about bluetooth to be sure), and thus rev is not determining position using the bluetooth, and is instead using it to send the data to the console.

This article was written in 2003, so some things may have changed since then, but I think if someone found a way to do occlusion resistant radio frequency position tracking, there would be some talk about it on the net. As it stands, every sound based tracker node I’ve seen has either broadcast its position by sending a noise via a speaker, or interpreted its position by receiving a noise via a microphone, and covering up that speaker or microphone made the tracking system stop working.

 

[SpoonyBard]

"But just because my laptop can talk to my phone via bluetooth does not mean my laptop can figure out my phone's position in 3D space does it?"

If you had 3 laptops, you could. With one receiver and one transmitter you can only tell the distance of the two units. With two receivers you'd know that the phone is somewhere in the insteresection of two spheres. With three receivers you'd get the exact position of the phone (the intersection of three spheres). The receiver doesn't know what direction the signal is coming from, only the time it takes from it to travel from the transmitter to the receiver. With RF, the signal is travelling at the speed of light, with ultrasound, the speed of sound. Ultrasound would be probably better, because calculating the timing would be easier, cheaper and more accurate. The ultrasound transmitters and mics could of course be a bit more expensive.

Earlier in the thread I talked about four reveivers, but I was thinking about GPS system, where the fourth satallite is also needed for timing purposes. There's no need for that here.

 

[PkunkFury]

That makes sense, but then I'm curious why VR wands like intersense bother using the microphones outside of the device. Are they only there to improve accuracy, and if that is the case will the rev controller be inaccurate without them?

 

[Scrow]

Holy crap I never even considered that. Has Nintendo found the best way to foil pirates to date?

Huh? How would the controller prevent pirates from copying software and modifying the console to circumvent any anti-piracy measures implemented in console hardware?

As for emulation on the PC though, that would be made almost impossible because of the controller. Is that what you were referring to, PC emulation?

 

[SpoonyBard]

That makes sense, but then I'm curious why VR wands like intersense bother using the microphones outside of the device. Are they only there to improve accuracy, and if that is the case will the rev controller be inaccurate without them?

The lack of obvious holes for microphones (or speakers) is one of the main reasons I'm not sure if the Revolution controller uses ultrasound for position tracking, but they could just have been removed from the publicly shown models of the controller, because they want to keep things secret. Or they could just be hidden inside the controller... but that would probably make them not as accurate.

 

[PkunkFury]

Sounds good, I guess I'll just wait to see the sensor bar before wondering any more about what is inside. I just hope Nintendo can deliver this technology in a cheap and reliable manner. If so I'm all for it

 

[KonVex]

According to Merrick both sensors can be arbitrarily placed, so it couldn't work like the P5 gloves.
Also the quality of the triangulation is highly dependent on the accuracy of the distance measurements. Since atomic clocks are not feasible they would need another method for accurate timing, but I'm not sure about the options. I think a quartz-crystal at several Ghz (20+) would be needed.

 

[Deleted member 1235]

Bluetooth is just 2.45 GHz radio signal, so it doesn't need any visible windows in the controller's casing. I'm starting to believe that the window in the front of the controller is just for the power button.

probably, part of its appeal is "looking like a conventional remote" according to Iwata also.

would be cool if the power button turns on the rev and your television at the same time. Fast startup, instant gaming.

 

[Hyoushi]

Bluetooth is just 2.45 GHz radio signal, so it doesn't need any visible windows in the controller's casing. I'm starting to believe that the window in the front of the controller is just for the power button.

Yes. I agree with this post.

 

[Leguna]

TO ALL:
Question, do you guys think I'll need a TV to play with the controller or would a front projector be ok? It looks you don't point the controller at the TV, looks like a only measures speed and motion using a laser like from the tip (my guess).

 

[SantaC]

TO ALL:
Question, do you guys think I'll need a TV to play with the controller or would a front projector be ok? It looks you don't point the controller at the TV, looks like a only measures speed and motion using a laser like from the tip (my guess).

projector has been confirmed to work as well.

 

[Ruzbeh]

TO ALL:
Question, do you guys think I'll need a TV to play with the controller or would a front projector be ok? It looks you don't point the controller at the TV, looks like a only measures speed and motion using a laser like from the tip (my guess).

Jim Merrick said it works with CRT, LCD, Projector, Plasma and stuff. So I assume you're just fine.

Edit: Beaten!

 

[Leguna]

Cool, thanks.

 

[SpoonyBard]

According to Merrick both sensors can be arbitrarily placed, so it couldn't work like the P5 gloves.
Also the quality of the triangulation is highly dependent on the accuracy of the distance measurements. Since atomic clocks are not feasible they would need another method for accurate timing, but I'm not sure about the options. I think a quartz-crystal at several Ghz (20+) would be needed.

That's one of the main reasons I think using ultrasound would be better than Bluetooth/RF. You don't need as accurate timing for the measurements as with RF.

 

[John Harker]

What ever happend to the Rev being able to hook up to a computer monitor?

Is that still a factor, or is it out of the equation now?

 

[Gahiggidy]

Will it work under water?

 

[sc0la]

Jesus Christ people.

The controller doesn't use IR for controller input, it doesn't have gyros, it doesn't shoot lasers.

There are two receivers in the sensor bar that are placed under or on the TV. These Two sensors talk to sensors in the remotes. With one sensor in the remote, a calculated distance from the sensor bar could yeild only two possible points in space (four technically but what the fuck would you be doing behind your TV?), but with four (top/bottom front and top/bottom rear) the sensors could determine distance, height and rotation about the x, y and z axis.

Theoretically, all the sensor bar would really need to do is receive regularly intervaled packets of data from the controller. If the controllers broadcast a known packet of information from known locations on the controller at the same speed and same frequency every time, then the sensor bar simply has to determine the time inteval it takes the signal to reach it to determine the various sensors' distance. If every sensor point in the controller broadcasts its location at the same time with great frequency the system can determine its position at any given moment of broadcast. It doesn't need a pedometer, it doesn't need gyros, it doesn't need lasers, i doesn't need IR; Just enough points to distance, a known speed of transmission and a small enough interval of transmission.

I don't know anymore than anybody else about how the controller actually works or what is inside it. I do know enough about geometry and physics to know that half of the ideas spit out in this thread are ridiculous.

EDIT: Okay reading through the thread there are actually some people who have a good grasp and have explained a few different plausible ways in which this could work. I really believe this is all based on knowing distances when the controller pings, i.e. no need for gyros or "pedometers." The idea of a third home sensor is nice in that it alleviates the single possibility of dual possible points in my imaginary controller set up (the controller being equally distant above or below the sensor bar and upside down).

Anyway, I doubt we will get any strong indication of what is specifically going on in these things as I imagine nintendo will want to keep those secrets locked up tight.