Old Discussion Thread for all 3 motion controllers

Status
Not open for further replies.
Sony also hasn't mentioned how much memory will need to be reserved for Move.

Yes they have

http://www.engadget.com/2010/03/13/playstation-move-requires-only-2mb-of-ram-developers-breath-sig/

Quizzing Sony's David Coombes, they found out that the advanced image processing required to make sense of your wild, flailing movements will take only 1-2 MB of RAM.

Given they've been able to retroactively add Move support to games that already used all the SPUs they had access to, I'd imagine Move runs on the one that was already reserved.

SOCOM seems to support that theory, I doubt they'd take an SPU from such a complex game

As well any system used will have to accurately take into acount the distance between a person's iris' otherwise

I've heard a few devs say they just use 6.5 cm as the difference
 
Yes they have

http://www.engadget.com/2010/03/13/playstation-move-requires-only-2mb-of-ram-developers-breath-sig/

Given they've been able to retroactively add Move support to games that already used all the SPUs they had access to, I'd imagine Move runs on the one that was already reserved.

That does seem quite likely.

SOCOM seems to support that theory, I doubt they'd take an SPU from such a complex game

Killzone 2 had an average load of 60% on the SPEs. Even if the Move required games to surrender some SPE power, I don't think there are many games that don't have 15% of SPE power to spare.

I've heard a few devs say they just use 6.5 cm as the difference

Currently, about 1 in 5 people have trouble watching 3D. That the 6.5 cm setting is not appropriate for them is one of the factors there. For games it's probably going to be quite easy to allow calibration, since the two views are dynamically rendered already anyway - you just adjust the camera positions slightly closer to each other or wider.
 
The Wiimote style with camera in the controller is far superior in some ways. It doesn't require players to stay in the camera's FOV for instance. So you can easily have 4 people flailing around without having to make sure they stay in the FOV of the camera.

On the flip side it does have drawbacks, as you obviously won't be able to do any head tracking for instance. But for 4 player active party games, it's by far the superior method of input.

Both Sony and MS on the other hand need to require players to stay within a limited area.

They have slightly different use case.

Wii requires the Wiimote to remain pointed at the sensor bar. If you point the Wiimote "out of screen" for too long, it will complain. It also requires more frequent recalibration (masqueraded as "Do you want to take a rest now ?" in WiiSuports Resort).

PS Move allows you to swing your stick freely. You can point it at any direction. It will only complain if the bulb is obscured for more than 5 seconds.

PS Move can support up to 4 Move controllers. If you use a subcontroller, then you're limited to 2 Move controllers + 2 subcontrollers. As you said, Wii does have a wider playing angle.


Move will also be resource hungry, requiring 1 SPE to be reserved for it's use. Isn't 1 SPE already reserved for the gameOS? Sony also hasn't mentioned how much memory will need to be reserved for Move.

So it could well end up at 15% or more of total system resources similar to Natal.

In the interviews, Anton mentioned that PS Move uses very little resources.

Although 1 SPU is reserved for PS Move, I don't think it'd be fully utilized for tracking the light bulb and controller orientation. The latter is similar to SIXAXIS tracking, whereas the former should be very suitable for vector math.

However, PS Move's scope also includes controller-free gaming (e.g., head tracking, facial recognition). More of the SPU would be used in these complex use cases. I remember Dr. Marks had a prototype image processing library that runs on only 1 SPU.
 
In the interviews, Anton mentioned that PS Move uses very little resources.

Although 1 SPU is reserved for PS Move, I don't think it'd be fully utilized for tracking the light bulb and controller orientation. The latter is similar to SIXAXIS tracking, whereas the former should be very suitable for vector math.

However, PS Move's scope also includes controller-free gaming (e.g., head tracking, facial recognition). More of the SPU would be used in these complex use cases. I remember Dr. Marks had a prototype image processing library that runs on only 1 SPU.

Indeed, very few games max out the SPUs, so IMHO we're likely to see no drop in quality with Move support (Socom being a great example). If devs are clever they will stick to adding value rather than noise.
 
I am not sure if I got the last part with the owl :smile:

Can you explain? :p

Is it because it looks so much like a lolipop that you cant notice that it is holding the controller as a lolipop?

You dang yougins, no sense of history. :p It's the owl from the old Tootsie Pop commercials. "How many licks does it take to get to the center of a tootsie pop? Lets ask Owl." They knew how to make commercials back then. :LOL:
 
There's a big difference between pointing at the screen and having that trajectory mapped to a point on the display, and having a pointer on-screen which is controlled by arm gestures.

Very true, but, except when his arms exceeds the canvas, it feels like its tracking his arm direction, not just moving the pointer in the direction his arm moved... It feels like a pointer, the very few times on the video when he points in the center of the canvas with a slower speed than the needed to throw the paint, the pointer seems a projection from where the avatar is pointing.

Do you know if Ms allowed anyone to test this demo too?
 
:LOL: Kotaku maximizing mileage from Dr. Marks:
http://kotaku.com/5502008/the-possi...f-playstations-move-according-to-the-inventor

I didn't explain Wiimote and PS Move controller tracking differences well enough. Here's Dr, Mark's attempt. I suspect most people won't get/appreciate it until they use it for real, and there is a compelling use case:

One of Sony's most aggressively expressed talking points about the Move is that it detects movement in the Z-plane — the plane that defines a player's distance from the vertical slice of air defined by the front of his or her TV better than any other motion controller, including the Wii Remote. The Remote, when pointed at a TV and moved forward and back, can be used to determine its range in the Z-plane, but Marks emphasized the Move's ability to be detected in the Z-plane at all times and with 60-frames-per-second precision based on the Eye's detection of the position and relative size of the Move's sphere.

"When you want the absolute best tracking, you have to have the absolute best position tracking, which is the camera for us," he said. If the sphere is ever obstructed, like when a player might through their hand back while holding a Move, the controller's motion sensors kick in to approximate the position of the Move, a technique that is similar to what can be accomplished with the Wii Remote's add-on, the MotionPlus.

But what's the big deal with detecting movement in the z-plane, with detecting more than just movement of a wand controller up and down (y-plane) or side to side (x-plane)?

The answer was partially provided in Marks' latest Minority-Report-style Move demo, which he would refer to throughout our conversation.

"You can punch in Z," Marks said."The Wii does that and we do it too. But what we were just doing while I was moving the camera around and flying, the only way to do that is Z."

And what of games? "If you want to place something in the 3D world; if you want to reach into the 3D world and manipulate. .. maybe I can grab things. (Fellow Move researcher) Anton Mikhailov wants to make a game where there's things like eggs that you have to pick up softly and other things you have to pick up with ammo triggered to them. I want to make a game where I'm a Greek god and I have to smash these little evil guys and the good guys I have to pick up carefully and safely.

"Reaching into the world like that, there's no way to do that if it's (only) 2D."

Essentially, when we want to track the Z distance, Wiimote must be pointing at the sensor bar whereas with PS Move, you can hold the controller anyway you want as long as the bulb is visible.
 
img6648-3102010-3.jpg


What the heck is that 2nd connector called EXT?

Same article above:

"I think there's a lot of options with having this thing being part of another peripherel, with something bigger that you would hold" Marks said, with Move in hand. "We've had a lot of talks with our licensees abotu what would make sense for this... We do have this USB cord [at the base of the wand] which is how you charge it. We also have this external port which is proporetary. Our licensees can talk to us about it." I asked what Sony uses it for. "We don't use it right now for anything, but it has data and power, I guess I can say that much."

EDIT:
I was thinking about the analog trigger too:
As I typically do, I concluded my interview with Marks by asking if there was anything key he thought we didn't cover. It turnes out the Move has an unheralded feature that he believes could be a big deal: the analog trigger on its underbelly. As opposed to the Wii's similarly-positioned digital trigger, the Move's "t-button" is sensisitive to the same gradual pull as the analog triggers on the shoulders of traditional modern game controllers. "I still think that people underestimate the importance of this kind of controller where you're doing something and you squeeze," Marks said.

... and the light ball ! Don't ignore the light ball (Well, it kinda stick out like a sore thumb for some gamers).



No press talks about how Sony can combine both camera-based, AI and sensor-based tracking together. The puppetry demo is just one example. I am sure more can be done in this area. We don't really care whether there's a controller or not, but if the technologies are applied together, I think it will elevate the experience. e.g., Handwriting recognition for Scribblenaut PS3.

The controller is also good for sending XMB message to friends, just sending my hand-written/drawn note is enough to make a point. The computer doesn't need to recognize/transcribe it. There will be lot's of smilies and leet speak anyway. :)
 
He should make the point that the Wii with motion plus uses acceleration to determine where it is, but measuring that way is inaccurate because they cannot detect every change in acceleration, but you can detect a position in space with the move in an absolute error free way.
 
Currently, about 1 in 5 people have trouble watching 3D. That the 6.5 cm setting is not appropriate for them is one of the factors there. For games it's probably going to be quite easy to allow calibration, since the two views are dynamically rendered already anyway - you just adjust the camera positions slightly closer to each other or wider.

They could do that easily by adding a setting to your profile, along with the size of your TV, position of the camera.

While few games max out the SPUs, I'd imagine a lot of them still use all of them
 
It's a bit irrelevant anyway as holding your arm out will only get you a minute or two of gameplay before it gets tired.

Actually i'm more curious on non-gameplay applications for pointing on natal, like dealing with interfaces :p

For a game i believe it will be a lot better to use just an video overlay to simulate pointing (something along some Move Party demos, like the draw inside shapes one and the one which you had to cut the hair of the creature)
 
But, didn't Kudo said that this kinda of pointing (by projecting the arms direction at the screen for instance) wasn't possible?
It's certainly possible, using the forearm trajectory and mapping to a calibrated screen-space. You wouldn't have hand/finger tracking, but given the significant distance between elbow and hand, as long as both are accurately positioned, the target trajectory should be accurately derivable. That said, it's going to require excellent tracking of the joints. If the elbow position is partially/completely obscured by the forearm, there'll be an a degree of error in its placement and hence the forearm trajectory will be out. This wouldn't affect a game with an avatar-type object/arm mapped to the player's arm, but it would cause significant inaccuracies in a cursor. Still, you could derive a general target okay with substantial smoothing, good enough for a painter program or rough point-and-shoot or unit selection in an RTS.

Move just reverses that, allowing it to work at more angles.
Yes, significantly! Patsu link to Mark's comment explains this in detail, but it should be readily apparent that with Move pointing 90 degrees to the side, it's z position is still tracked, whereas Wii would lose its accuracy. Wii also has quite a limited range from my experience, with 3-4 metres-ish being the threshold where it starts to lose positioning. Move is a significant improvement on movement tracking, even over Wii Motion+.
 
He should make the point that the Wii with motion plus uses acceleration to determine where it is, but measuring that way is inaccurate because they cannot detect every change in acceleration, but you can detect a position in space with the move in an absolute error free way.

Is it accurate to say:

When the user points the Wiimote+ at the sensor bar, Wii can determine the relative and exact location of the user (plus additional sensor data from the gyro and accelerometer).

When the IR camera is not used (i.e., Wiimote+ swinging wildly in Golf or Ping Ping game), then the accelerometer guesses the location of the player via incremental data. This part is inaccurate.

Also for the digital compass in PS Move, does it mean that PS3 can now do geo-location applications like iPhone now ? (e.g., Find nearest Pizza Hut in my neighbourhood). If so, the press has missed yet another major area to explore.

Actually i'm more curious on non-gameplay applications for pointing on natal, like dealing with interfaces :p

For a game i believe it will be a lot better to use just an video overlay to simulate pointing (something along some Move Party demos, like the draw inside shapes one and the one which you had to cut the hair of the creature)

You may not need detailed 3D orientation data for UI. Using PS Eye alone, I can navigate Creature Feature screens just fine by waving my hands -- or anything -- over the on screen buttons/objects. I think it's pretty cool, I hope PS Move keeps this part instead of throwing away the entire stack. In fact, I think they should refine it further.

OTOH, with all the talks about absolute positioning on PS Move, these additional data are meaningless if the games/applications don't use them effectively. In fact, it could harm the title. e.g., Walking in the "Minority Report" UI is inefficient, the absolute positioning Anton demoed there may not be useful as a general UI mechanism.
 
When the IR camera is not used (i.e., Wiimote+ swinging wildly in Golf or Ping Ping game), then the accelerometer guesses the location of the player via incremental data. This part is inaccurate.
To be annoyingly nitpicky, the accelerometer doesn't guess anything. :p It provides the acceleration data from which one can determine how much the Wiimote has moved and get a placement. This will be inaccurate to the accuracy of the accelerometer(s) and accumulated errors. A long period without any fixed-location calibration could see significant drift. This is no doubt why Wii golf/bowling have you point towards the screen before commencing your pull back, to reset the position to a known quantity prior to each reliance on the accelerometers.
 
True, true. Somebody answer my digital compass question please. ^_^

Ah, no... not GPS. Just orientation w.r.t magnetic north. :(
Brain not working correctly this morning.

Was playing with the iPhone Yelp app last night. If I can retrieve geo-location data from a cellphone, would be interesting to do a similar compass app on PS3.
 
Also for the digital compass in PS Move, does it mean that PS3 can now do geo-location applications like iPhone now ? (e.g., Find nearest Pizza Hut in my neighbourhood). If so, the press has missed yet another major area to explore.
Nope. It's no GPS system! The compass simply returns direction of the Move relative to magnetic north. It has no idea where in the world you are, only which way you are pointing - east is east whether you're in Hawaii or Alaska!
 
Status
Not open for further replies.
Back
Top