Support NeoGAF

[Gahiggidy]

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.

If there is no type of tilt-sensor, then how could the sonsor-bar detect twisting along the axis of the remote?

 

[sc0la]

If there is no type of tilt-sensor, then how could the sonsor-bar detect twisting along the axis of the remote?

tandem points in the controller. If there are transmital points in the top or bottom as well as the front and back it can use those to extrapolate the roll. It could have tilt sensors inside the controller, but it doesn't need them, it can just do the same process that it is already doing with more points of transmission.

I am not sure which is better, that is for Nintendo to decide.

 

[maharg]

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.

At some point, if not already pending, they will probably patent it. Certainly the patent will probably be pending by the time 3rd parties get actual controllers. When they patent it, how they did it will be public (as is generally the case with patents). If they don't patent it, they open themselves up to easy copycats. This is, however, assuming that the existing patents from Gyration don't basically cover the patentable parts.

If they don't, we won't know until someone gets one and opens it up and lets everyone know how it looks like it works inside.

 

[Shao]

Jesus Christ people.

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

OK I never said it shoots lasers, but its quoted everywhere as functioning like a bloody laser pointer. NOT A 3D MOUSE. But hey thanks for being so condescending *rollseyes*

I'm sure all these technically minded people would be able to differentiate between mouse-like capabilities and laser pointer-like capabilities. Of course thats where things stop making sense. If there is no calibration, if you can place the sensors arbitrarily anywhere around your TV and you can move around freely as shown in the demos, how the hell does this baby work?

My only conclusion is that there simply has to be some form of calibration for times when it wishes to perform like a laser pointer or otherwise accessible in the system menu.

 

[sc0la]

OK I never said it shoots lasers, but its quoted everywhere as functioning like a bloody laser pointer. NOT A 3D MOUSE. But hey thanks for being so condescending *rollseyes*

I'm sure all these technically minded people would be able to differentiate between mouse-like capabilities and laser pointer-like capabilities. Of course thats where things stop making sense. If there is no calibration, if you can place the sensors arbitrarily anywhere around your TV and you can move around freely as shown in the demos, how the hell does this baby work?

My only conclusion is that there simply has to be some form of calibration for times when it wishes to perform like a laser pointer or otherwise accessible in the system menu.

I wasn't calling anybody out in particular, relax.

It can re calibrate itself any time it wants to "re-zero." It should be able to decern when the controller is pointed straight at the TV just based on what ever locational system it uses. From there, depending on where you point the controller the cursor will move. All the pointing showcases described used some kind of crosshair indicator of where you where pointing, thus your movement is all relative to that. Any program can have its own level of sensitivity based on the programming.

I don't think the thing is capable of an acurate detection of what location you are pointing at on the screen. I just don't think it is realistic, if they are going for a seamless gaming experience, true pointing would require manual calibration almost everytime. Relative pointing does not. Based on the existence of cusors in most of the demos I believe it relies on relative pointing. But I could be wrong. (and that isn't to say that developers couldn't impliment true pointing on a game by game basis, but it would require manual calibration)

I think they already patented the technology (there was an old patent flying around that showed a gameboy being used physically in a lettuce chopping simulation, like the revolution controller teaser movie)

 

[Mike Andrade]

Sorry if old.

Game system with tilt sensor and game program including viewpoint direction changing feature

 

[SpoonyBard]

Jesus Christ people.

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?)

Actually, wouldn't that be an intersection of two spheres, not two circles. So instead of two points in space you'd get quite a bit more points.

 

[sc0la]

Actually, wouldn't that be an intersection of two spheres, not two circles. So instead of two points in space you'd get quite a bit more points.

You are correct. You would actually get a circle with every point on that circle a possible location. But that is if there is only one point of detection in the controller.

My brain is too tired to figure out what the possible outcomes of more are.

 

[SpoonyBard]

These are the 3 technological possibilities that we have came up so far:

1. IR.
+ cheap
- limited number of places where the sensors can be placed on the controller
- unreliable (needs visual connection)
- position tracking requires accurate timing

2. Ultrasound.
+ very accurate
> not as reliable as RF, but sligtly more than IR
- limited number of places where the sensors can be placed on the controller

3. RF/Bluetooth.
+ very reliable.
+ the controller can hold as many transmitters and receivers as needed.
+ transmitters and receivers can also be placed inside the console.
- position tracking requires accurate timing

I think it's pretty safe to say that the Rev controller uses one of these. I also think the controller has tilt sensors, accelerometers, and/or gyros to increase accuracy and measure tilting.

 

[DSN2K]

its bluetooth.

 

[ImNotLikeThem]

Will it work under water?

Yes Gahiggidy, just like all good revolutionary wireless video game system controllers, the Revolutions tilt and positioning functions of it's controller will work fully and accurately when under the surface of deep water.

And also, Nintendo has already stated the controller is using Bluetooth.

 

[SpoonyBard]

And also, Nintendo has already stated the controller is using Bluetooth.

I know what was said in the interview, but that doesn't necessary mean that the location measurement part of the controller uses it. It could be used only for communication between the controller and the console. Also standard unmodified Bluetooth doesn't seem to be very good for accurate measurements anyway...

 

[Shao]

I wasn't calling anybody out in particular, relax.

It can re calibrate itself any time it wants to "re-zero." It should be able to decern when the controller is pointed straight at the TV just based on what ever locational system it uses. From there, depending on where you point the controller the cursor will move. All the pointing showcases described used some kind of crosshair indicator of where you where pointing, thus your movement is all relative to that. Any program can have its own level of sensitivity based on the programming.

I don't think the thing is capable of an acurate detection of what location you are pointing at on the screen. I just don't think it is realistic, if they are going for a seamless gaming experience, true pointing would require manual calibration almost everytime. Relative pointing does not. Based on the existence of cusors in most of the demos I believe it relies on relative pointing. But I could be wrong. (and that isn't to say that developers couldn't impliment true pointing on a game by game basis, but it would require manual calibration)

I know but you were still pretty arrogant considering you hadn't read the thread properly and said something we'd more or less already figured out although not confirmed - which isnt possible right now. But you seem to know you're stuff so thats cool.

Personally I think its natural for it to have on-screen cursors even if you DO point directly at the screen. Example, if you happen to have a laser pen or torch handy, the visual cues are vital in the accuracy and take nothing away from the experience or perspective.

All the signs suggest it is this relative pointing but I can't help thinking that wouldn't be a solid gaming option alot of the time for all sorts of reasons that make it not so user-friendly, as well as it not being that all that revolutionary. so I hope they found some easy way to do this laser pointing because that would indicate a shift similar to the touch screen rather than adapting a 3D mouse.

 

[PkunkFury]

These are the 3 technological possibilities that we have came up so far:

1. IR.
+ cheap
- limited number of places where the sensors can be placed on the controller
- unreliable (needs visual connection)
- position tracking requires accurate timing

2. Ultrasound.
+ very accurate
> not as reliable as RF, but sligtly more than IR
- limited number of places where the sensors can be placed on the controller

3. RF/Bluetooth.
+ very reliable.
+ the controller can hold as many transmitters and receivers as needed.
+ transmitters and receivers can also be placed inside the console.
- position tracking requires accurate timing

I think it's pretty safe to say that the Rev controller uses one of these. I also think the controller has tilt sensors, accelerometers, and/or gyros to increase accuracy and measure tilting.

Great summary.

magnetic transmitters/receivers are also possible, but I think they are costly.

Bluetooth sounds like a good bet, but I find myself wondering why I've never seen a VR tracker that uses it. Here is a 6DOF project called BlueTrak (also from 2003) which relies heavily on the use of bluetooth:

http://hep1.physik.uni-bonn.de/MySQL/Files/bluetrak_ISMAR03.pdf

However, once again the bluetooth is being used to send data about the position in space of each module to the computer, and ultrasound is what actually determines the position. I gather maybe the reason I'm not finding stuff about bluetooth trackers is the same reason I'm not finding papers from beyond 2003? Technology is too new, and companies guard their secrets? I've also never seen one at Siggraph or IEEE VR, but I wouldn't really have been looking either. Most applications of blutooth tracking seem to be for gross tracking of bluetooth devices over wide areas (children at a zoo, location of a cell phone) like GPS. Therefore I'd like to question the "very reliable" status of a bluetooth device being used as a tracker. One paper I found (http://www.princeton.edu/~msimsir/perv/project/BTlocation.pdf) described a precision of within 2.08m when detecting a bluetooth device's position which is within 8m of each receiver. This group was using bluetooth receivers to detect the location of a standard bluetooth device, so it is possible a device designed for tracking could improve these numbers, but they'd have to improve a lot for accurate human gesture tracking.

BTW most arcade games that are doing motion tracking are using IR. particularly games like MoCap boxing, Police 911, that sword fighting game, and other konami motion tracking arcades.

 

[Jared Goodwin]

I know what was said in the interview, but that doesn't necessary mean that the location measurement part of the controller uses it. It could be used only for communication between the controller and the console. Also standard unmodified Bluetooth doesn't seem to be very good for accurate measurements anyway...

For some reason I want to say BT 2.0 has some optional more-rigorous timing specifications that Nintendo might be using, but I could be totally off base on this one.

 

[SpoonyBard]

This article from 2002 says there weren't any working RF tracking technologies available, but things could have been changed since then:
http://www1.cs.columbia.edu/~drexel/CandExam/Motion_Welch_Foxlin.pdf


Also ultrasound tracking could have it problems:
"Depending on room acoustics and tracking volume,
it may be necessary for the system to wait anywhere from
5 to 100 ms to allow echoes from the previous measurement
to die out before initiating a new one, resulting in
update rates as slow as 10 Hz."

That's way too slow for gaming.

And:
"Finally, we have yet to
see a purely acoustic tracker that doesn’t go berserk
when you jingle your keys."

Just imagine the possibilities in multiplayer gaming

 

[PkunkFury]

Wow, great article! This is just what I was looking to post here! A concise description of every method used for position tracking! I haven't had time to read it yet, but it looks like a very thorough paper. It's already confirmed a lot of my thoughts about the tech Nintendo may or may not use.

I'm still banking on optical tracking with internal gyros/accelerometers.
The major disadvantage of optical tracking is that oreintation has to be determined from multiple light sources on the tracker (like the P5 glove) but the gyro would fix that problem. The other disadvantage is that line of sight is necessary, but I think Nintendo's "pointing device" was meant to be pointed at the TV at all times anyway. This means games like baseball which use full range swinging motions would not be looking for position data (and wouldn't have to). The gyros and accelerometers would take care of that. Couple this infor with the IR style window on the front of the remote, the price and design of the P5 glove, and the description of the revolution's receptor bars, and I think optical tracking is a pretty safe bet.

Finally the benefits for optical tracking (Infra-red and the like) are beyond perfect for gaming. As quoted from the paper: "Analog optical PSD sensors combined with active light
sources offer the combination of relatively high spatial precision and update rates."
What more could you want for a video game??

I'll check out the key jingling with the IS-900. I'm almost positive it doesn't mess up the newer acoustic trackers.

 

[KonVex]

I was rethinking the trilateration approach and it seems unlikely now.
You not only have to know the distances to the sensors but also their positions,
that could only be solved with some major calibration.

Does anyone know for certain how the P5 gloves work? I'm not certain how
they would cope with detachable sensors.

 

[fr4nz]

The french Website Gamekult said in their handon of the controller that it stop to work if something is put between the remote and the sensors ....

http://www.gamekult.com/nextgen/actus/articles/A0000043594.html

...On remarquera juste que le faisceau est coupé lorsqu'un corps étranger passe devant, en l'occurrence une personne ou un bloc-notes, puisque c'est tout ce que nous avions sous la main....

 

[SpoonyBard]

Oh well, it's just IR then. With gyros and other stuff like that thrown in for measure. The front of the controller may house 3 IR-LEDs, which would give it more accuracy than having just one...

 

[PkunkFury]

I was rethinking the trilateration approach and it seems unlikely now.
You not only have to know the distances to the sensors but also their positions,
that could only be solved with some major calibration.

Does anyone know for certain how the P5 gloves work? I'm not certain how
they would cope with detachable sensors.

Yeah, good point. I'm not too sure about set ups working with detachable sensors. Seems you'd need to know the positions of the sensors to get good data, at least in relation to one another. There is a great website that tells how the P5 glove works. I posted a link to it earlier in this thread

The french quote about obscuring the front of the remote causing the tracker to lose position enforces the idea that this must be IR optical tracking.

 

[Shao]

Merrick also says how "the viewing angle in front of the sensor bar is similar to a TV, if you can see the TV, the sensor will work"

Which again suggests it depends on light, at least for some functions. But he also said it uses blutooth, what if positioning used both methods? Would that be worthwhile? IR could pinpoint the location of the remote on 3 axis, but something else would be needed for angle, tilt and so on (since there is only one IR window).

In terms of gyros, if the remote is stationary, does it still detect current angle? For example my remote is upside down on my bed when I turn the system on, will it detect that or do gyros only report movement rather than position?

 

[Gahiggidy]

The french Website Gamekult said in their handon of the controller that it stop to work if something is put between the remote and the sensors ....

http://www.gamekult.com/nextgen/actus/articles/A0000043594.html

Oh shit. Horrible news if true.

Nintendo's fucked.

 

[Beezy]

I expect to see a "Nintendo is Doomed" pic soon...

I hope that it isn't true.

 

[Gahiggidy]

Here is the bit about "breaking" the signal...

First demonstration, first appearance of the capacity of "aiming" of the lever. One tests in his turn, one carries out tests of acceleration, as with a mouse PC, one looks if the sensor takes down, but not, the technological one well developed seems, any detectable latency time. It will be noticed just that the beam is cut when a foreign body passes in front of, in fact a person or a scratch pad, since it is all that we had under the hand.

Note, that it was in reference to the 1st demo. The "light-gun" demo. Perhaps line-of-sight is only affected by the "pointer" controls... while the mouselike and orientation controls are not affected.

Otherwise, Nintendo's fucked. No serious gamer is going to stand having to worry about loosing conection just becasue somebody walked across the room.

 

[Krowley]

Here is the bit about "breaking" the signal...Note, that it was in reference to the 1st demo. The "light-gun" demo. Perhaps line-of-sight is only affected by the "pointer" controls... while the mouselike and orientation controls are not affected.

Otherwise, Nintendo's fucked. No serious gamer is going to stand having to worry about loosing conection just becasue somebody walked across the room.

i don't think it's that serious. if someone walks in front of the TV, you can't see the screen anyway and usually you stop playing until they pass by. would not bother me.

it also uses tilt sensors, so obviously you don't have to point the remote directly at the tv at all times... if you did, the fishing demo would not have worked properly. It's a new control scheme, and it will have it's quirks, people will get used to them.

 

[Gahiggidy]

i don't think it's that serious. if someone walks in front of the TV, you can't see the screen anyway and usually you stop playing until they pass by. would not bother me.

it also uses tilt sensors, so obviously you don't have to point the remote directly at the tv at all times... if you did, the fishing demo would not have worked properly.

But the French guys said the fishing demo sucked:

...With the screen a lake drawn coarsely with the pencil appears, in 2d, in which some fish batifolent. With far a tree or two, a cane with fishing, and a hand. Our host seizes the controller, who thus directs the hand and approaches the latter of the cane with fishing. He carries out then a movement of comings and goings in the air, i.e. he draws the lever towards him then the growth, always with the horizontal one, and the line will be thrown in water. Then while dropping or while mounting the lever, always maintained with horizontal, it varies the depth to which the hook is. When a fish bites, it then gives a rapid blow of wrist to the top, to shoe the animal, and here. Demonstration is thus made that the controller is more than one simple gyroscopic peripheral, since its position in space, in 3d, is well taken into account. On the other hand, on this demonstration, control was not so easy for each one among us, because not very precise, in fact.

 

[argon]

I doubt Nintendo would release something as accurate and reliable as the Wavebird, and use that technology in the revolution controller, yet rely on IR solely for motion tracking... There must be more to this.

 

[DrGAKMAN]

Oh shit. Horrible news if true.

Nintendo's fucked.

HA HA HA...you're joking, right!?!

Seriously, if something is blocking the communication between the remote-controller's DPD and the sensors near the TV...it's most likely also blocking the screen which makes games unplayable anyways!

 

[bronzodiriace]

revolution gpu
and ram



Revolution GPU

ATI Custom based RN520 core. The "N" stands for Nintendo, and is because the ArtX team is with them, that is why it’s an "N".

GPU core at 600 MHz. Will support up to 2048x1268 resolution, HD support is still being decided. Will have 256 MB’s of 1T-SRAM (the RAM is much better due to some tweaking, compared to GC’s RAM. The latency and Cells are much more efficient and faster. Around 1.2 ns is the latency, on average).

32 parallel floating-point dynamically scheduled shader pipelines.

Polygon Performance: 500 million triangles per second theoretical, average in game would be around <100 Million/sec>

Shader Performance: ~50 billion shader operations per second

Revolution memory

512 MB of 700 MHz 1T-SRAM

 

[KonVex]

HA HA HA...you're joking, right!?!

Seriously, if something is blocking the communication between the remote-controller's DPD and the sensors near the TV...it's most likely also blocking the screen which makes games unplayable anyways!

That's not really the problem, but you can't read the position of the controller when it faces away from the screen.
This would make most uses that were implied in fact impossible (baseball, fishing, golf, ...).
These could only be approximated using the rotational information.

 

[LUNA]

On the other hand, on this demonstration, control was not so easy for each one among us, because not very precise, in fact.

"not very precise" he is clearly talking about French guys not the controller

As it has been said, I guess the IR must be there at least to turn the console on. I would say even for the navigation around Revo menus when you are not playing, as IR is much less energy consumer than the other solutions.

I am sure Nintendo won't go out with a bad controller for gaming, but what I really don't know is how they will solve the "limits" in the movement of your hand. Everytime you go further with the controller than what the game can show (wall limit, arm displacement, movement too fast for ingame movement,...), it can be a pain in the ass to go back to coordinate the control position-orientation with the onscreen image .

Hope it is an understandable comment.

 

[DrGAKMAN]

That's not really the problem, but you can't read the position of the controller when it faces away from the screen.
This would make most uses that were implied in fact impossible (baseball, fishing, golf, ...).
These could only be approximated using the rotational information.

Well, um...there's going to be more to it than simply IR. I would imagine that if IR is involved it's for dirrectly pointing/aim at something on screen...while more rigorous motions would be handled by more appropriet technology. My point still stands...if I'm aiming at something on screen and someone stands in the way it may block the signal for a moment *but more importantly* the basterd is blocking my veiw of the screen which makes the game more unplayable than anything, der!

 

[Gahiggidy]

HA HA HA...you're joking, right!?!

Seriously, if something is blocking the communication between the remote-controller's DPD and the sensors near the TV...it's most likely also blocking the screen which makes games unplayable anyways!

What if I want to curl up with a blanket while I play?

 

[Krowley]

What if I want to curl up with a blanket while I play?

yeah, thats the only problem i see with it.

the revolution will be a console that you really need to play sitting in a chair. I have a tv in my bedroom, and i spend a lot of my gaming time kicking back on my bed, occasionally i even cover up my controller, which would make playing the rev pretty damn awkward if not impossible.

I'm willing to deal with it though, i'll probably set up a new arrangement with a nice comfortable easy chair right in front of my TV. I used to have a setup like that, and it was comfortable enough.

edit//At least you don't have to stand up like playing an eyetoy, and the technology sounds so cool that it will probably be worth changing my gaming setup. i mean damn.. we finally have a controller that works like a virtual hand, that's pretty amazing.

 

[snatches]

revolution gpu
and ram



Revolution GPU

ATI Custom based RN520 core. The "N" stands for Nintendo, and is because the ArtX team is with them, that is why it’s an "N".

GPU core at 600 MHz. Will support up to 2048x1268 resolution, HD support is still being decided. Will have 256 MB’s of 1T-SRAM (the RAM is much better due to some tweaking, compared to GC’s RAM. The latency and Cells are much more efficient and faster. Around 1.2 ns is the latency, on average).

32 parallel floating-point dynamically scheduled shader pipelines.

Polygon Performance: 500 million triangles per second theoretical, average in game would be around <100 Million/sec>

Shader Performance: ~50 billion shader operations per second

Revolution memory

512 MB of 700 MHz 1T-SRAM

Link?.....Source?.....kthxbye

 

[Robobandit]

Link?.....Source?.....kthxbye

Indeed.. a source would be nice. Those specs sound higher than I would have expected.. by quite a bit.

 

[bronzodiriace]

Link?.....Source?.....kthxbye

a factor 5 developer on gamespot's forum

 

[G4life98]

revolution gpu
and ram



Revolution GPU

ATI Custom based RN520 core. The "N" stands for Nintendo, and is because the ArtX team is with them, that is why it’s an "N".

GPU core at 600 MHz. Will support up to 2048x1268 resolution, HD support is still being decided. Will have 256 MB’s of 1T-SRAM (the RAM is much better due to some tweaking, compared to GC’s RAM. The latency and Cells are much more efficient and faster. Around 1.2 ns is the latency, on average).

32 parallel floating-point dynamically scheduled shader pipelines.

Polygon Performance: 500 million triangles per second theoretical, average in game would be around <100 Million/sec>

Shader Performance: ~50 billion shader operations per second

Revolution memory

512 MB of 700 MHz 1T-SRAM

would this be good or bad?

 

[Robobandit]

That would be very good.. the R520 is Ati's unreleased answer to Nvidia's 7800 GTX. It is supposed to be slightly faster than the GTX even with less than 32 shader pipes, and I would think 1T-SRAM would be faster than GDDR3, so if this turns out to be true, the Revolution has some serious graphics power.

 

[Kroole]

I think Factor 5 knows nothing, they are sucking Sony's cock right now..

I have to admit. It seems strange that Factor 5 would know all the specs if they aren't going to produce anything for it.

 

[snatches]

a factor 5 developer on gamespot's forum

LINK PLZ

 

[snatches]

I have to admit. It seems strange that Factor 5 would know all the specs if they aren't going to produce anything for it.[/QUOTE.

Not really. But I am skeptical that this guy is for real. This would put there system on par with X360 and PS3, and for months it seems we have been warned that this will not occur.

 

[SpoonyBard]

Well, Nintendo's main opponent is Sony, and even if they match the GPU and memory, the PS3 would still probably have a faster CPU.

 

[yoopoo]

Haven't we already learned that 'insiders' on blogs and forums pretending to hold info on Revolution are always wrong and know nothing...? No?

 

[Culex]

Speculation never gets you anywhere when it comes to the hardware statistics.

Let us just wait for official specs from Nintendo or Ati.

 

[Drensch]

for months it seems we have been warned that this will not occur.

By who? Prognosticating Sony and MS fans?

 

[Shao]

You know those were simply demos, with prototypes at that. While things moving in front of the sensors might block any IR function, it shouldnt block any blutooth information sent - so what do they mean by "it stops working"?

It will be noticed just that the beam is cut when a foreign body passes in front of, in fact a person or a scratch pad, since it is all that we had under the hand.

Beam? If something is blocking your "laser" in a laser demo, it would be impressive if this translated into the game as well. They also say the controller doesn't spaz out. Also wonder why they were made to stand in a precise position as the sensors should be reading this stuff on the fly, maybe some sensors were not set up properly, hence the wires.

Its evident that it functions just fine as a 3D mouse and should do so whether or not someone was in the way, although as suggested by the French this method is not altogether natural or intuitive.

Why couldn't any other site provide such indepth analysis of the demos darn it, bloody amateurs.

 

[Shogmaster]

revolution gpu
and ram



Revolution GPU

ATI Custom based RN520 core. The "N" stands for Nintendo, and is because the ArtX team is with them, that is why it’s an "N".

GPU core at 600 MHz. Will support up to 2048x1268 resolution, HD support is still being decided. Will have 256 MB’s of 1T-SRAM (the RAM is much better due to some tweaking, compared to GC’s RAM. The latency and Cells are much more efficient and faster. Around 1.2 ns is the latency, on average).

32 parallel floating-point dynamically scheduled shader pipelines.

Polygon Performance: 500 million triangles per second theoretical, average in game would be around <100 Million/sec>

Shader Performance: ~50 billion shader operations per second

Revolution memory

512 MB of 700 MHz 1T-SRAM

Oh for fuck's sake.... At least make it somewhat plausible!

Fucking Nintendo's coyness about the specs is gonna subject us to these kinds of wet pants fanboy BS speculations nonesense for months more to come!

KILL ME NOW!!!! X_X

N stands for Nintendo. Teehee!

STFU!!

 

[SpoonyBard]

You know those were simply demos, with prototypes at that. While things moving in front of the sensors might block any IR function, it shouldnt block any blutooth information sent - so what do they mean by "it stops working"?

They must have meant just the 3d mouse functions, tilt sensors, buttons and d-pad would still work.

Beam? If something is blocking your "laser" in a laser demo, it would be impressive if this translated into the game as well. They also say the controller doesn't spaz out. Also wonder why they were made to stand in a precise position as the sensors should be reading this stuff on the fly, maybe some sensors were not set up properly, hence the wires.

Its evident that it functions just fine as a 3D mouse and should do so whether or not someone was in the way, although as suggested by the French this method is not altogether natural or intuitive.

If the controller also has gyroscopes it shouldn't stop working even if the IR beam is blocked. You would notice that some of the accuracy is lost, and if the beam is blocked for longer time the "center" position of the controller would begin to drift. This is why 3D gyroscope-only mouses are shit.

Maybe the algorithms that let the gyroscopes take over were not in place yet, or maybe the controller doesn't have the necessary gyroscopes at all... only simple tilt sensors.

Why couldn't any other site provide such indepth analysis of the demos darn it, bloody amateurs.

Maybe Nintendo asked them to not be to technical, because the hardware and sotware are not finished yet.