Sony's New Motion Controller

[joker454]

Ah, OK. Can you say, are they a 2nd or 3rd party?

3rd party.

 

[Crossbar]

I think the PSMC's potential for "casual" game types is kinda a forgone conclusion. With cross platform potential with the Wii that stuff should just come naturally,...

Yeah, this is so obvious and that is why I expect the PSMC to get a far faster adoption rate than Natal in third party studios. In multiplat titles they can just offer dual control systems: pad style and wiimote style. The wiimote control scheme that they develop for the Wii will just be carbon-copied to the PS3. The amount of extra development will be limited to the porting and after a couple of titles it should be standard procedure. Suddenly a lot of PS3 games have a more casual friendly control scheme, that may come natural for people who want to upgrade their Wii to a HD video consol or casual people who bought the PS3 as mainly a blu-ray player.

Someone may say dual control systems make competetive online play unbalanced, but that shouldn´t be an issue as it has been handled in games before like UT3.

I´ve read Grahams posts several times and I seriously have a hard time to follow him. I can see that people are excited over Natal because it´s a brand new technology with lots of potential, but where is the economy in making games for it initially? It will initially have a very small installed base and the first party games will dominate it initially. For comparison look at how much third party support the Eyetoy got and still it had multi-million installed base. Microsoft is of course putting a lot more effort behind Natal than Sony ever did to the eye-toy, but it is a comparison that could be kept in mind as a reality-check.

Introducing Natal in multiplattitles will not come easily as it is so different from the Wii-mote control scheme, buttons are really handy. Having Natal as an extended control scheme to the pad in multiplat titles may be a more realistic approach, but still will require a lot of extra development efforts to get right. Just look at how many companies have bothered to use the Sixaxis in a meaningful way and still that functionality is included in every PS3 sold and don´t just say the Sixaxis is stupid because there are some good implementations out there and several have been suggested in this forum. Another reality-check.

As I said before I believe Natal will do well in first party Eye-toy type of titles and really shine in "In the movies" type of titles and that will have a big market and new scenes as dlc will make a lot of money. It is pretty obvious why Microsoft brought Spielberg to E3.

 

[patsu]

This may be old news by now (I am working through the GAF thread). New, and free !, addition to the PS3 SDK:
http://www.developmag.com/news/32075/Sony-adds-LiveMove-2-to-PS3-SDK

Sony revealed its motion-sensing 'wand' controller at its press conference yesterday - and already it's revealed its plans to help developers get the most of the peripheral.

AiLive – company which co-developed the Wii MotionPlus together with Nintendo – will make its LiveMove 2 motion recognition software available for free to all licensed PS3 developers thanks to a new deal with Sony Computer Entertainment.

Developers can 'train' the software to understand key gestures, which can then be accurately recognised when performed by players, taking the burden of motion interpretation away from game teams.

"The phenomenal precision of SCE's new controller combined with the effortless expressive power of AiLive's LiveMove 2 will dramatically expand the frontier for natural motion control in games," said Dr.Wei Yen, chairman of AiLive

"Our market-proven LiveMove 2 software gives developers the power to inject freeform player motion directly into their game. We're thrilled to help deliver this unparalleled level of control to PLAYSTATION® 3 players."

Masayuki Chatani, executive vice president and chief technology officer at Sony Computer Entertainment, added: "The LiveMove 2 development tool is the premier solution for recognizing human gestures in games. I am very happy to be able to make this revolutionary tool available to all PLAYSTATION® 3 developers. We eagerly look forward to the new and engaging game experiences developers will create for the PLAYSTATION® 3 with the help of LiveMove."

EDIT:
Someone did a rough mock up of the "transformer" Dualshock in the patent:

If they want to explore in this direction, the symmetry of Dualshock seems to work well for them here.

Cover up the sticks and we have almost a mouse.
Most people would buy 2 "waggle" controllers for co-op anyway.

 

[patsu]

This is even more interesting. Old presentations using Cell for computer vision, speech processing, head tracking, face detection, gesture recognition, video processing, etc.:
http://emsys.denayer.wenk.be/?project=EmCel&page=cases&id=23

Dr. Richard Marks has an entry in the list too (brief mentions of libvision, libcamera). libvision library seems to work in standalone fashion on a SPE (No DMA. Hm....). At the time of the presentation, the library was not released to the developers because Dr. Marks was unhappy with the design.

 

[Arwin]

Through gaf, I came across this:

http://audioboo.fm/boos/27700-e3-final-day

In it, Kaz says that they eventually dropped the ultrasound bit and just went with optical recognition only, using the size of the ball for depth. The only thing I can think of now is that the precision comes from that you know what you are tracking is a perfect circle. In that case what the software needs to do is:

1) find the circle in the image
2) map the digital image to a perfect circle of a certain size to determine x/y/z.

I think given a digital input image you don't actually even need a very high resolution to still get very precise input, as luminence is going to help also?

If Kaz has it right, it's still interesting though and it has a number of implications.

 

[Arwin]

Also, I came across this LiveMove 2 demonstration, which is also available in Sony's SDK:

 

[Quaz51]

as luminence is going to help also?
.

probably, the absolute luminence of the LED-sphere is known and fixed. relative luminence give information on the distance

 

[Quaz51]

Also, I came across this LiveMove 2 demonstration, which is also available in Sony's SDK:

i see he reset the sequence and recalibrate (put his arm at a specific neutral position) after every movement because lack of absolute tracking position like sony controller (and motion+ use only a 2 axe gyro, not really 6DOF). not true 1:1, impossible to redo the sony demonstration with this i think .

 

[Shifty Geezer]

Serious lag too. I don't want any motion control that is that laggy, at least not for anything other than casual games.

 

[Asher]

Through gaf, I came across this:

http://audioboo.fm/boos/27700-e3-final-day

In it, Kaz says that they eventually dropped the ultrasound bit and just went with optical recognition only, using the size of the ball for depth. The only thing I can think of now is that the precision comes from that you know what you are tracking is a perfect circle. In that case what the software needs to do is:

1) find the circle in the image
2) map the digital image to a perfect circle of a certain size to determine x/y/z.

I think given a digital input image you don't actually even need a very high resolution to still get very precise input, as luminence is going to help also?

If Kaz has it right, it's still interesting though and it has a number of implications.

If this is true, then there's nothing the PSMC can do that Natal cannot? All it is is a camera that detects lighted spheres and calculates distance and position based off of them?

I'm curious what the cost is for using PSMC -- 1 SPU?

 

[Shifty Geezer]

If this is true, then there's nothing the PSMC can do that Natal cannot?

As I understand it, PSMC also has motion detectors in the wand to determine tilt, rotation, etc.

 

[Crossbar]

If this is true, then there's nothing the PSMC can do that Natal cannot? All it is is a camera that detects lighted spheres and calculates distance and position based off of them?

You mean beside the accelerator sensors and the buttons and knobs of the wand?

I'm curious what the cost is for using PSMC -- 1 SPU?

That is probably the case. Marks powerpoint presentation mentions a lib that can be kept on an SPU without requiring DMA loads.

http://emsys.denayer.wenk.be/?projec...ge=cases&id=23

 

[Arwin]

Serious lag too. I don't want any motion control that is that laggy, at least not for anything other than casual games.

Then I recommend going for the PMC rather than WiiMotionPlus, as the latter is what he's demonstrating. Mind you I don't know if Nintendo have managed to improve this here, but I've heard in general that PMC is a lot more precise and quick than WiiMotionPlus, and it has to show up somewhere!

 

[Shifty Geezer]

Is PSMC proven yet? I remember one vid recently posted that showed the trouble PSMC was having with pointing, and the cursor wasn't following the wand particularly well, but this was with a split video-feed I think. At the moment I have high expectation but need ot be shown the working product is as good as it seemed. I expect it to be. Sony have been working on this stuff for ages and I don't expect them release a product that doesn't work properly. They haven't yet with other devices (no matter how lacking their software may be!)

 

[Asher]

You mean beside the accelerator sensors and the buttons and knobs of the wand?

Forgot about those. (It's early and I just got my coffee)

Yes, I suppose that's a big component. I'm just no waggle fan, sadly.

 

[Crossbar]

Noticed that some of the patent links didn´t work properly, so I made a new search.

Here is a bunch of patents Sony may be using. Enjoy!

DETERMINATION OF CONTROLLER THREE-DIMENSIONAL LOCATION USING IMAGE ANALYSIS AND ULTRASONIC COMMUNICATION

HAND-HELD CONTROLLER HAVING DETECTABLE ELEMENTS FOR TRACKING PURPOSES

GAME CONTROL USING THREE-DIMENSIONAL MOTIONS OF CONTROLLER

DETECTABLE AND TRACKABLE HAND-HELD CONTROLLER

SCHEME FOR DETECTING AND TRACKING USER MANIPULATION OF A GAME CONTROLLER BODY

SCHEME FOR TRANSLATING MOVEMENTS OF A HAND-HELD CONTROLLER INTO INPUTS FOR A SYSTEM

METHOD FOR MAPPING MOVEMENTS OF A HAND-HELD CONTROLLER TO GAME COMMANDS

METHOD FOR AIMING AND SHOOTING USING MOTION SENSING CONTROLLER

CONTROLLER CONFIGURED TO TRACK USER'S LEVEL OF ANXIETY AND OTHER MENTAL AND PHYSICAL ATTRIBUTES

MAPPING MOVEMENTS OF A HAND-HELD CONTROLLER TO THE TWO-DIMENSIONAL IMAGE PLANE OF A DISPLAY SCREEN

ENHANCED GAME CONTROLLER

 

[Shifty Geezer]

Forgot about those. (It's early and I just got my coffee)

Yes, I suppose that's a big component. I'm just no waggle fan, sadly.

I'm no fan of Wii waggle, but I do like the idea and potential of proper matched motions (Wii+ and PSMC and whatever). eg. the bow shooting example was very nice. Would be good in a more realistic RPG where you don't shoot an arrow every 2 seconds and need 30 to fell a monster, but needs accuracy and management of resources (hold of firing until the last moments). Incidentlaly in the demo I don't know if it was just me, but it seemed that the aiming was more erratic with the bow pulled further back. Was this a deliberate technique to simulate the difficulty of a higher draw on the string - more power = less accuracy, and the player could draw back less for less aim wobble but to deal lower damage? Or was it just the tracking having issues?!

 

[Crossbar]

Scanned some of the patents briefly. This is a fairly new one (October 6, 2008) which may be used to get that sub-millimeter detection.

ENHANCED GAME CONTROLLER

[0040]Using a multi-sensor module controller allows a user's input to map to gestures in the game. For example, a distance between two sensors can be used in a digital signal processor (DSP) along with how the sensor modules move, to improve mapping a user's gesture and to convert that gesture into an input for a game movement. Using multiple sensors in an enhanced game controller provides many advantages. For example, using DSP techniques, the input from the multiple sensors can be combined subtractively, additively, by averaging, or other combination to enhance the signal level as well as better detect the user's gestures. In other words, multiple sensors provide multiple forms of data from the user's gesture, and these multiple forms of data can be combined so as to produce a better signal to mimic the user's gestures.

[0041]Use of multiple sensors along with the knowledge of the structural coupling between the sensors can improve gesture mapping of a user's motion. For example, two sensor modules can be coupled using a support structure such that the two sensor modules are at a known off-set distance from one another, and the movement of the sensor modules relative to each other is bound by the constraints of the coupling of the support structure. The input signals from the two sensors and the known constraints of the support structure, for example, a rigid support structure or flexible support structure or resonant support structure, can be used to better detect signals from the sensors and map user's gestures.

[0042]FIG. 10 is a block diagram of another embodiment of an enhanced game controller 1000. As shown in FIG. 10, a sensor 1002 can be attached to a user, for example, to a user's hand 1004 by an attachment mechanism such as a strap 1006. In the embodiment shown in FIG. 10, the user's hand motions and gestures can be detected and input to a game as game movements or gestures. For example, if the sensor 1002 is a 6-axis sensor and is strapped to a user's hand, then in a game such as a boxing game, the motion as well as the rotation of the user's hand can be detected and input into the game to better mimic the user's actual motion.

Does anyone know how to view the images within the patents?

 

[AzBat]

I use this viewer...

http://www.internetiff.com/

Tommy McClain

 

[catisfit]

Incidentlaly in the demo I don't know if it was just me, but it seemed that the aiming was more erratic with the bow pulled further back. Was this a deliberate technique to simulate the difficulty of a higher draw on the string - more power = less accuracy, and the player could draw back less for less aim wobble but to deal lower damage? Or was it just the tracking having issues?!

I noticed the same. As the tracking for aim would be simple XY, and it happened just after he pulled back a long way to show how the system handled it, I'd be tempted to say it was just him fighting an aimkick. But it did look odd and it could have been a tracking anomaly.