Official Nintendo Fanboy Smugness & Gloating Thread

[xav]

My guess is that the "sensor bar" doesn't contain any sensors at all, but rather a couple of IR transmitters. With a camera behind the IR filter at the end of the Wii remote, it should be possible to keep track of position in a vertical plane by following the position of one of the sensor bar's transmitters in the camera's image.

The second transmitter allows it to calculate relative distance to the screen based on the distance between the two lights in the image. Without some calibration it wouldn't know how near or far you are, but could determine you moving forwards or backwards.

Throw in some tilt sensors and accelerometers in the remote itself and you can track it to some extent even if it's not facing the sensor bar (i.e. for the golf swing example). Assuming one of the tilt sensors can tell it how far off the horizontal the remote is then it should be able to calibrate the camera system even if the two reference points aren't on a horizontal line (which would fit with the claims that they can be removed from the bar if necessary).

[zbandito]

well if it was IR, it would show on any videocamera footage

[Rod_S]

Anyone seen this? X360 gesture recognition:

http://gizmodo.com/gadgets/e306/e306-cl ... 172978.php

[xav]

well if it was IR, it would show on any videocamera footage

Fair point. In that case it must be voodoo. I'd hate to be a chicken living near Nintendo HQ right now.

[Roobarb]

Neat

What they need now is a rumble glove to go with it. I'm a great believer in force feedback

[Roobarb]

well if it was IR, it would show on any videocamera footage

Fair point. In that case it must be voodoo. I'd hate to be a chicken living near Nintendo HQ right now.

The cameras could be on the sensor bar instead of course, with the wiimote acting as the beacon ...

[xav]

well if it was IR, it would show on any videocamera footage

Fair point. In that case it must be voodoo. I'd hate to be a chicken living near Nintendo HQ right now.

The cameras could be on the sensor bar instead of course, with the wiimote acting as the beacon ...

I considered that option: it would mean that each remote would have to send a distinct code via its IR beacon. Sending a distinct code would be simple enough, but decoding that information becomes a harder problem. Consider that the beacon would have to be off some of the time (otherwise it's not much of a code) so you end up with some frames in which the light is switched off, and some with it switched on.

Now consider that the light source is moving; you would end up with a frame with a light on, a frame with a light off, and another with it on but in a different location. Multiply by the number of remotes. Tracking the beacons in this case becomes harder.

Oh, and you've got to do this for two sets of images so that you can track depth as well.


None of this is impossible, but it does have a lot of scope for problems and would require some bloody clever algorithms. The engineer in me says that two static beacons using one camera is comparatively easy, whereas tracking four coded beacons using two cameras is comparatively hard.

Ergo, it must be voodoo.

[Ratsoap]

Anyone seen this? X360 gesture recognition:

http://gizmodo.com/gadgets/e306/e306-cl ... 172978.php

The fun part is the in and out.

[zbandito]

a few observations;

there could be a third sensor in the console itself, though I doubt it. I don't think it would help since the setup between sensor bar and console can vary.

There are probably other ways to do pointing; for example old light guns used to track a point on the screen. Perhaps the Wiimote does something similar; a timing sequence could be transmitted to the remote so that it knows where to looks for a certain pixel display that is shuffled around the screen to random locations at 60hz. For example, sending code 1234 means the pixel pattern will appear at locations 1-2-3-4 onscreen...

ultrasonic emittors of two different frequencies might be cheaper than IR in some places

chances of interference have gone up; aside from many 2.6ghz wireless devices at e3, there are IR and ultrasonic considerations. playability could be disrupted by atmospheric conditions at times

[Atty]

Considering that sites that got a more intimate pre-e3 play didn't report any of these calibration problems, it seems likely that the problems were caused/exaggerated by the rather unique conditions at e3.

It's also worth nothing from the look of the videos that most people were standing too close to the screen, plus they were, well, standing, which doesn't help when you're trying to make controlled movements with your hand without fatigue creeping in.

I'm not trying to be Mr Nintendo apologist, and I think Ninty were rather stupid for repeatedly promising the controls would immediately be completely intuitive, but considering we are still atleast 4 months away from launch and developers freely admit they've not done any real fine tuning on the pointer sensitivity yet, things are generally looking very positive.

People are naturally going to feel like the bubble's been burst a little bit after all the claims Nintendo have made, but they do seem to be delivering a genuinely workable control system. And that's good enough for me.

[Darkflame]

a few observations;

there could be a third sensor in the console itself, though I doubt it. I don't think it would help since the setup between sensor bar and console can vary.

There are probably other ways to do pointing; for example old light guns used to track a point on the screen. Perhaps the Wiimote does something similar; a timing sequence could be transmitted to the remote so that it knows where to looks for a certain pixel display that is shuffled around the screen to random locations at 60hz. For example, sending code 1234 means the pixel pattern will appear at locations 1-2-3-4 onscreen...

ultrasonic emittors of two different frequencies might be cheaper than IR in some places

chances of interference have gone up; aside from many 2.6ghz wireless devices at e3, there are IR and ultrasonic considerations. playability could be disrupted by atmospheric conditions at times

I dought theres any screen single issues.
Nintendo said it worked fine on all display types includeing projectors.
Plus, if there was any screen thingys at all, you wouldnt need a bar.

[aureole]

Is there anything definative on what degrees of freedom the controller can actually sense?

[Roobarb]

I considered that option: it would mean that each remote would have to send a distinct code via its IR beacon. Sending a distinct code would be simple enough, but decoding that information becomes a harder problem. Consider that the beacon would have to be off some of the time (otherwise it's not much of a code) so you end up with some frames in which the light is switched off, and some with it switched on.

Now consider that the light source is moving; you would end up with a frame with a light on, a frame with a light off, and another with it on but in a different location. Multiply by the number of remotes. Tracking the beacons in this case becomes harder.

Oh, and you've got to do this for two sets of images so that you can track depth as well.

None of this is impossible, but it does have a lot of scope for problems and would require some bloody clever algorithms. The engineer in me says that two static beacons using one camera is comparatively easy, whereas tracking four coded beacons using two cameras is comparatively hard.

Yes it probably would be harder. Just trying to account for zbandito's observation re: IR on videofootage Another possibility is the Wii IR transmitters and receivers use a longer wavelength IR than Nightshot style cameras are responsive to.

I doubt what Nintendo are doing even needs to be even as elaborate as a camera in the controller though to be honest. Since the mems devices can detect acceleration and tilt in all 3 axes, there's no need to reproduce any of that functionality with the sensor bar.

It looks likely the sensor bar is only there to determine roughly where the controller is located (and is pointing) relative to the tv screen.This would make the the 'screen pointing' aspect of the controller appear more realistic than just using the mems device alone (which has no idea if you're pointing at the screen or not).

This could be done with a camera in the controller for sure, but it may be even simpler/cheaper to just use two simple IR motion sensor systems (e.g. one responding to horizontal angle, the other responding to vertical angle, such as this). The sensors could be attached to a bar for neatness, or detached and placed in more convenient locations on the TV.

One thing that hasn't been made clear yet is whether you can 'point' using the sensor bar with more than one controller at a time (as opposed to using just the mems device to detect wave type motions).

If you can, then it would start to get more complicated for the reasons you outlined.

From an old interview with Jim Merrick:

How is movement of the controller detected?
We use Bluetooth technology to communicate between the controller and what we call a 'sensor bar', which has two little sensors on it that are maybe a foot apart. These sensors can be detached from the bar and they can be above the TV or below the TV - it doesn't really matter.
There's really no set-up other than just putting the bar by the TV. There's no calibration for size or type of TV or anything like that.

So presumably the controller will be compatible with all TVs?
It works with any kind of television set - it doesn't matter if it's LCD, plasma, projector or CRT.

And the gamer won't have to sit in the same position every time they're playing a title that uses the motion-sensing function?
Absolutely not - the relative angle in front of the sensor bar is about the same as the viewing angle of the TV. So if you can see the TV, the controller is going to work.

Latest sensor bar pictures

Ergo, it must be voodoo.

Nah, Nvidia already use that

[Flabio]

Nah, Nvidia already bought that

Fixed that for you

[Muttley]

During an extended technical demonstration for a new, unnamed Xbox 360 and PlayStation 3 Star Wars title, LucasArts staff have confirmed to Gamasutra the company's intense interest in creating a Wii lightsaber game, if not yet its explicity existence.

The X360/PS3 demonstration, named 'Force Powers Pre-Vis', featured a CGI trailer showing various Jedi defeating stormtroopers in a variety of imaginative and amusing ways.

The second part of the demonstration was in real-time on the Xbox 360, and showed the console simulating a number of real-world materials such as wood, stone, crystal, ice, bone and the Star Wars metal "carbonite" using Pixelux Entertainment's Digital Molecular Matter (DMM) technology. The DMM system was being used in conjunction with NaturalMotion's euphoria technology in both the new Star Wars game and LucasArts's forthcoming new Indiana Jones game.

At the end of the demonstration, Gamasutra inquired as to whether the company planned on creating a lightsaber game for the Wii, after many commented on he suitability of the system to the concept - especially after an internal speaker was revealed in the controller being used to demo the concept.

This question produced a number of knowing smiles around the room from LucasArts employees, followed by the comments: "We know" and "We are looking into it", as possible concepts for the game were discussed. However, the firm has not yet made any official announcements regarding planned Wii titles.