Old Discussion Thread for all 3 motion controllers

Status
Not open for further replies.
So what, did Turn 10 just suck at optimising Forza 3? Would seem so if they can retain the same graphical fidelity + more AI while taking the 10-15% Kinect overhead hit.

Either that or the 10-15% rumor is with every single feature and control of Natal active, and it's significantly less if you only utilise part of it. I don't think we'll ever know for sure though as I'm sure companies will be under NDA if they have that info.

Also, I'm watching the EA presentation right now, and from rewatching the EA Active 2.0 demonstration, it looks like Kinect has far less lag than the EA designed motion detectors for PS3 and Wii. The lag for the PS3 and Wii with the EA arm and leg bands is just really REALLY bad.

Regards,
SB
 
Here's something else I don't get.

Forza team has shown that Kinect is perfectly fine tracking only part of a persons body.

http://e3.gamespot.com/video/6265584/?tag=top_stories;title;8

You can see there that often only the head and shoulders and hands become visible to Kinect when he moves right up to the screen while kneeling down.

I don't see why sitting would be such a problem if you don't need to track the lower body while sitting. It doesn't matter if Natal loses track of those parts because they wouldn't be used for...say movie watching.

I'm sure MS would love to get the calibration such that the legs could also be tracked accurately while sitting, but when sitting I just don't see that as entirely necessary for a couch experience.

Regards,
SB
 
Either that or the 10-15% rumor is with every single feature and control of Natal active, and it's significantly less if you only utilise part of it. I don't think we'll ever know for sure though as I'm sure companies will be under NDA if they have that info.

Also, I'm watching the EA presentation right now, and from rewatching the EA Active 2.0 demonstration, it looks like Kinect has far less lag than the EA designed motion detectors for PS3 and Wii. The lag for the PS3 and Wii with the EA arm and leg bands is just really REALLY bad.

Regards,
SB

From what i saw from the Rare games they showed where that 10% rumor came from(?). They use all of natal. The child of eden not sure if that was 3D tracking or webcam tracking he had white gloves on so it's probably easier to work with.
 
The EyeToy technology may have been limiting but the crazy thing is that Sony has been working on various Motion Control schemes since the PS2 days.

I barely remember them even demonstrating a Move/Wii alike controller on the PS2 around in 2000.

I remember seeing a very primitive, 2D version of an EyeToy-like experience on the Amiga using NewTek's DigiView and a genlock. :smile:
 
Either that or the 10-15% rumor is with every single feature and control of Natal active, and it's significantly less if you only utilise part of it. I don't think we'll ever know for sure though as I'm sure companies will be under NDA if they have that info.

Also, I'm watching the EA presentation right now, and from rewatching the EA Active 2.0 demonstration, it looks like Kinect has far less lag than the EA designed motion detectors for PS3 and Wii. The lag for the PS3 and Wii with the EA arm and leg bands is just really REALLY bad.

Regards,
SB

Perhaps, but I don't think it scales down linearly, there's probably a significant minimum overhead to use Kinect, it definitely seems so, as Kinect support of any kind can't be patched into existing titles unlike Move, the fact that Forza Kinect required an engine rewrite (and they already built a new engine for Forza 3) would seem to confirm this.

With Forza, upper body motion tracking is still going to incur a significant cost, I would say 10% rather than 5%. We really need Digital Foundry to do a tech analysis of Forza Kinect, I reckon the physics wouldve taken a hit, no more 360hz refresh.

I'm not sure if this is possible, but could they configure the engine in such a way that when you're playing Forza with Kinect (and so with assists enabled) the physis and graphics are pared back, but if you play full simulation with a wheel or controller, Kinect functionality is disabled and you get the full experience?

Or do they have to build the Kinect processing cost into the entire engine so that even when Kinect is not being used, there's still 10% or whatever reserved?
 
This is the first I've heard of you being a moderator. I didn't know until now.

Ha ha, I meant I deleted his comment from my post. ^_^

Look at all these previous eyetoy games that failed with the casual...

They didn't fail with casual. Those EyeToy games were successful. I know casual EU and Asian non-gamer friends who bought a PS2 because of it. I don't know if SCEA focused on EyeToy at all.

At that time, the "mainstream" dancing game was Dance Dance Revolution. It's cheaper (to run). I believe Dr. Marks mentioned that body tracking on PS2 took 30% of the CPU.
 
A lot of catch up to do, but fortunately, I have seen the hundreds of beautiful women in the Ninty conference.

Now for Vitality Sensor:
http://wii.ign.com/articles/110/1100039p1.html

IGN: To start us off, what happened to the Vitality Sensor?

Miyamoto: It's actually progressing quite well. We could have shown it here at E3 if we wanted to, but the environment here isn't really suited to that game. E3 is pretty exciting, and that's a device that's all about relaxation. So it just wasn't the best time or place to show that game. But it is ready, so we'll probably find another place that's more suitable to announce it. But I'm not directly involved in that project.

IGN: What are your impressions of the PlayStation Move and Microsoft's Kinect?

Miyamoto: I haven't played all of the demos myself and haven't had much opportunity to be out on the show floor, but I think it's a good idea in general that the party game mentality and the idea of playing games in your living room is beginning to broaden even further. In general, I think that's a good thing for video games.

IGN: You mentioned the 3D camera. I think it's great that you can take photos in 3D, but obviously that feature's locked down to the system. Has there been any thought for creating a standard to get those 3D pictures into other mediums so people can share those pictures?

Miyamoto: Well, maybe somebody will start making 3D picture frames.

[size=-2]3DS better have homebrew to export the 3D images out.[/size]
 
Here's something else I don't get.

Forza team has shown that Kinect is perfectly fine tracking only part of a persons body.

http://e3.gamespot.com/video/6265584/?tag=top_stories;title;8

You can see there that often only the head and shoulders and hands become visible to Kinect when he moves right up to the screen while kneeling down.

I don't see why sitting would be such a problem if you don't need to track the lower body while sitting. It doesn't matter if Natal loses track of those parts because they wouldn't be used for...say movie watching.

I'm sure MS would love to get the calibration such that the legs could also be tracked accurately while sitting, but when sitting I just don't see that as entirely necessary for a couch experience.

Regards,
SB

I'm pretty sure the Forza demo was staged.

http://i48.tinypic.com/epk4fa.jpg

Not only does the cursor move before his hand, but it doesn't match his motions at all.
 
If it's staged then this guy practiced for a really really really long time

http://e3.gamespot.com/video/6265584/

The first part of the video is extremely sketchy to me, either he's not controlling it, or it has a lot of bugs. There are areas where the hands on screen move independently from his own, they move at the same time, or they have distinct lag. Very strange indeed.

The second portion, however, has no bugs at all. However, it's extremely likely that all that is going on there is eye tracking and hand tracking.

Honestly though, at this point most of this is guess work, and we'll never know what these products have to offer until we get our hands on them for ourselves. Time will tell!
 
The first part of the video is extremely sketchy to me, either he's not controlling it, or it has a lot of bugs. There are areas where the hands on screen move independently from his own, they move at the same time, or they have distinct lag. Very strange indeed.
I noticed those things as well.
 
I don't see why sitting would be such a problem if you don't need to track the lower body while sitting. It doesn't matter if Natal loses track of those parts because they wouldn't be used for...say movie watching.

I'm sure MS would love to get the calibration such that the legs could also be tracked accurately while sitting, but when sitting I just don't see that as entirely necessary for a couch experience.
Yes, in principle. So why could it not be working? I'm guessing because not everyone sits bum-in-seat, feet-on-floor. Some will be reclined feet in front. Some legs to the side. Some will bring their knees in to their body, especially at a scary moment, even hugging a cushion. It could be confused data like that which Kinect can't handle which makes it unstable in home use, as opposed to labratory use. For games, I'd expect a player to be sat forwards and easier to track especially combining Kinect with the controller in shooters etc.
 
I'm pretty sure the Forza demo was staged.

http://i48.tinypic.com/epk4fa.jpg

Not only does the cursor move before his hand, but it doesn't match his motions at all.

Which isn't even remotely like the behind doors Forza demo that I linked to if you had bothered to look. There would be no way to memorize all that inadvertent swaying as well as body motions (mimicked in the demo) reacting to questions asked by people in the room.

Especially when it semi-glitches out when he turns completely around to look back at the questioner to answer one of the questions.

That was definitely NOT staged.

And in that room he gets down with his head within inches of Natal and it tracks his motions perfectly fine with basically only his head being visible to the system.

Likewise with him moving forwards and backwards, disproving without a doubt that you "have" to stand in a certain spot. Although I'm quite sure with MS still calibrating the system that currently best results are achieved when used from a certain area.

Regards,
SB
 
Yes, in principle. So why could it not be working? I'm guessing because not everyone sits bum-in-seat, feet-on-floor. Some will be reclined feet in front. Some legs to the side. Some will bring their knees in to their body, especially at a scary moment, even hugging a cushion. It could be confused data like that which Kinect can't handle which makes it unstable in home use, as opposed to labratory use. For games, I'd expect a player to be sat forwards and easier to track especially combining Kinect with the controller in shooters etc.

Actually my guess is that certain fabric material may absorb and reflect the infrared light similar to the clothing worn by people. If this is the case, it could be a very difficult problem to solve. If the body is in motion then it would be much easier for the system to recognize that it's not a static part of the background, but legs and hips aren't going to move much when sitting. Your arms and torso though should move enough to give a good reference point. Maybe playing naked would solve the issue? :D

Notice how in Your Shape, the clothing/sweater is seen as just part of the body (although that could be artistic rather than technical limitation), but there's enough discrete control points that it can easily (relatively) deduce where all the control parts probably are with a good deal of accuracy. In the seated position it'll have much less known control points to work with to determine where the rest are.

But again, when seated, just how important will the legs, knees and hips be? If they can at least get arms, upper torso, and head tracked well that should be good enough for couch potato activities.

But you do have a point with sitting off-angle to the camera. The Forza behind doors demo had a glitch when the presenter turned to look at someone in the back of the room when they asked a question. If I were to guess, the tilting and panning abilities of Natal might have been introduced as much to deal with this problem as it was to keep people in the FOV.

It's too bad we can't see the point cloud images of a person seated to see what Natal is coming up with.

Regards,
SB
 
The first part of the video is extremely sketchy to me, either he's not controlling it, or it has a lot of bugs. There are areas where the hands on screen move independently from his own, they move at the same time, or they have distinct lag. Very strange indeed.

The second portion, however, has no bugs at all. However, it's extremely likely that all that is going on there is eye tracking and hand tracking.

Honestly though, at this point most of this is guess work, and we'll never know what these products have to offer until we get our hands on them for ourselves. Time will tell!
I really do think this demonstration (and the old Burnout w/ Natal) was done so that there is another person actually controlling the game with a wheel or pad somewhere where he can see the player who's using his body to control, but of course hidden from view (can be even in another room, if he has a camera view to the demonstration room).
Just try it for fun.
Ask some unsuspectful victim to play Burnout with your "new motion control" console, and when he starts to play, mimic his actions with a controller behind him (preferably a wheel, as it's easier to accurately mimic).
I did this, and managed to fool a couple of persons :D I was even surprised myself how well that worked and how the lag seemed very similar to the of Kinetcs, and how real the illusion of motion control for the fooled player was!
 
I really do think this demonstration (and the old Burnout w/ Natal) was done so that there is another person actually controlling the game with a wheel or pad somewhere where he can see the player who's using his body to control, but of course hidden from view (can be even in another room, if he has a camera view to the demonstration room).
Just try it for fun.
Ask some unsuspectful victim to play Burnout with your "new motion control" console, and when he starts to play, mimic his actions with a controller behind him (preferably a wheel, as it's easier to accurately mimic).
I did this, and managed to fool a couple of persons :D I was even surprised myself how well that worked and how the lag seemed very similar to the of Kinetcs, and how real the illusion of motion control for the fooled player was!
I guess we have lag all explained :LOL:
 
I also had the impression that the demonstration was faked....but I can understand that MS tried to minimize the risk of showing the demo, it seems to me that all of the shown stuff is in an early dev state...so why the heck risk a crash or something else...just buy better actors next time :LOL::LOL:

EDIT: if I remember correctly, MS had already some new stuff crashed episode/nightmare (was it new windows presentation with blue screen) - so they are warned!
 
Status
Not open for further replies.
Back
Top