Meant to be Seen

I'm kind of a little sad that sixense are ditching the razer brand... I really like razer, I feel what they have done for PC gaming has been great and making peripherals is really tough in the PC market but really important for it to compete with consoles.

Oh well

I hope this isn't the case:

Concerned about Sixense Wireless because it's just a 3D rendering and no real specs:
http://www.reddit.com/r/oculus/comments ... _its_just/

Parallaxis wrote:I hope this isn't the case:

Concerned about Sixense Wireless because it's just a 3D rendering and no real specs:
http://www.reddit.com/r/oculus/comments ... _its_just/

ahhh, the choosing of the exact same font style and color as the Oculus Rift SDK for the Sixense Development Kit, tells me that there is a lot of enthusiasm for VR in general hidden.
Also the joke of anouncing a 'Revolution' one day too early implies that there's more too it than just
checking out new marketing options.
'Fear is the path to the dark side. Fear leads to anger. Anger leads to hate. Hate leads to suffering.'

This does look good. I hope they have vibration capability as well.

Parallaxis wrote:I hope this isn't the case:

Concerned about Sixense Wireless because it's just a 3D rendering and no real specs:
http://www.reddit.com/r/oculus/comments ... _its_just/

Did anyone on the internet actually think this was a picture?

Reddit to the rescue!

STEM
Sixense Tracking Electromagnetic Motion?

The STEM device will enable us to explore all these things:
http://en.wikipedia.org/wiki/STEM_fields

I was guessing pretty close to this, but i also imagined it would have 2 ankle staps with trackers as well. Could they be an add on later?

Would it be necessary to have trackers on both feet? Wouldn't tracking one foot be enough? With this type of device, what would be the optimum /efficient body model?

JoeReMi wrote:Would it be necessary to have trackers on both feet? Wouldn't tracking one foot be enough? With this type of device, what would be the optimum /efficient body model?

If you would 'only' track the gait direction, it would be sufficient to mount one tracker onto the pelvis(torso).
But anyway you would have to know if your feet are moving or not.
On only one foot would mean that you would have high latency in the case you raise the foot that has no tracker. And you cannot predict if you are going to raise the foot without tracker unless you use the Myo.
Perhaps in addition with some primitive electrodes on the soles or on the floor that sense when your foot is up or down.

With regard to what they (sixense) are actually saying in terms of motion sensors, I read the following sentence in their press release:

Sixense will release development kits that will support three wireless motion trackers AND will include fully playable controllers compatible with games and applications developed for existing Sixense-powered products

To me that says there are to be 2 hand controllers (as in the current Hydra) AND 3 extra motion sensors.

So if I my conjecture he some what...

Oculus will create a head translation tracker (The head can do a LOT of movement using just the neck)
Sixence will create a "Hydra 2" with 3 motion sensors 1 for torso and 1 for each foot (along with the motion trackers for hands in the controllers)

This sounds good to me, although I guess one more tracker for the waist would be better (I know, never bleeding satisfied !! lol)

Future looks good

That's a bit optimistic, I'm betting two of those trackers are within the "playable" controllers.

Ideally, the new wireless controllers could be used WITH the existing Razer Hydra controllers. You could mount the Hydra base in a backpack, and mount a Hydra controller on torso and foot, perhaps. And still use the full Sixense Wireless controllers. Only a dream, for now...

Getting game developers to use even just realistic head/hand controls is a huge enough advancement. People also don't have unlimited funds for VR controls. With rift positional tracking, and wireless hydras you can manipulate things very well, things like leg controls etc. increase cost and setup time for your average gamer(they need to have it cheap to try it).

I think what would work well for "full body" VR would be after the consumer Rift is commercially successful, Oculus could release their own VR plug & play system, where you buy an Ouya type box (better gpu of course) that runs on a VR operating system and plays *only* VR apps, or apps that convert content to VR. Not only does this kit come with the gaming box, but also comes with Rift gloves, leg-straps, and torso all with at least current controller level force feedback, if not actual haptics.

All Rift-box games offer positional/body/hand/head tracking, high performance, minimal bugs, open source ,etc.

This system could easily cost under $1000, including Rift.
$300 Rift/headphones
$200 VR box (ouya is only 100, but we want better GPU)
$100 gloves
$250 torso
$100 leg straps


It would be " true virtual reality" and destroy the console market, while costing roughly the same, only you're indie game developers so games cost 15 dollars, not 60.

PC would struggle against it as well, but less so due to flexibility.

I'm pretty sure there will be 2 controllers and a head tracker (STEM). This is both necessary and sufficient for a good VR control scheme. Torso tracking would be nice but it doesn't really add much for a seated vr experience, and that's what Oculus and Sixense are aiming for right now.

Why destroying the console market. Nate Mitchell already stated that the new consoles can run 1080p SBS images. And what about the new reached photorealism. I wanna definitely experience
those new games with the Rift.
About full body tracking. I dont know much about electronics, but wouldnt it be in general be possible
and also cheap if one would build adaptable exoskeletons (telescope mechanism) in which joints
potentiometers would be installed? So that you can measure an alternating analog value for the resistance each time you bend your arm or leg. Could this be a cheap alternative to track the whole body?
I remember Novint has build an arm exoskeleton. https://www.youtube.com/watch?v=lV3j2Yxv7jY

http://www.melexis.com/Hall-Effect-Sens ... s-760.aspx
A contactless magnetic sensor capable of measuring a magnetic field in 3 axis. For only 4 €.

colocolo wrote:http://www.melexis.com/Hall-Effect-Sens ... s-760.aspx
A contactless magnetic sensor capable of measuring a magnetic field in 3 axis. For only 4 €.

The specs say intrinsic linearlity error is 1-degree on XY, but 20-degrees on XZ and YZ. It does not say anything about measuring position, and it says it can measure a MOVING magnet (unless that is just unfortunate wording in that sentence). Being Hall sensor based should be able to read stationary magnets too.

boone188 wrote:I'm pretty sure there will be 2 controllers and a head tracker (STEM). This is both necessary and sufficient for a good VR control scheme. Torso tracking would be nice but it doesn't really add much for a seated vr experience, and that's what Oculus and Sixense are aiming for right now.

Maybe. Sixense was presenting alongside Omni, so who knows what they're doing.

Anyways, there's no reason it can't be both.

geekmaster wrote:

colocolo wrote:http://www.melexis.com/Hall-Effect-Sens ... s-760.aspx
A contactless magnetic sensor capable of measuring a magnetic field in 3 axis. For only 4 €.

The specs say intrinsic linearlity error is 1-degree on XY, but 20-degrees on XZ and YZ. It does not say anything about measuring position, and it says it can measure a MOVING magnet (unless that is just unfortunate wording in that sentence). Being Hall sensor based should be able to read stationary magnets too.

Wouldn't it be enough to have only one positional sensor on the exoskeleton for being a reference point? From there you would only need rotation data of the joints.
But 20 degree would be way too much. Would only work for knees and elbows.

colocolo wrote:

geekmaster wrote:

colocolo wrote:http://www.melexis.com/Hall-Effect-Sens ... s-760.aspx
A contactless magnetic sensor capable of measuring a magnetic field in 3 axis. For only 4 €.

The specs say intrinsic linearlity error is 1-degree on XY, but 20-degrees on XZ and YZ. It does not say anything about measuring position, and it says it can measure a MOVING magnet (unless that is just unfortunate wording in that sentence). Being Hall sensor based should be able to read stationary magnets too.

Wouldn't it be enough to have only one positional sensor on the exoskeleton for being a reference point? From there you would only need rotation data of the joints.
But 20 degree would be way too much. Would only work for knees and elbows.

20-degree linearity. You could calibrate for that. The Hydra also has linearity problems, such that you can measure a position with repeatable precision to within 1mm, but the absolute linear accuracy can be way off. Because it is repeatable, you could measure a series of calibration points and interpolate between them, so not really a problem. You just need to be aware that you cannot use it as simply as you might expect, if you want absolute positioning accuracy.

geekmaster wrote:The Hydra also has linearity problems, such that you can measure a position with repeatable precision to within 1mm, but the absolute linear accuracy can be way off. Because it is repeatable, you could a series of calibration points and interpolate between them, so not really a problem. You just need to be aware that you cannot use it as simply as you might expect, if you want absolute positioning accuracy.

ah, would that be why the sixense tuscany demo didn't actually have 1:1 hand positioning when it was in 1:1 mode?

Mystify wrote:

geekmaster wrote:The Hydra also has linearity problems, such that you can measure a position with repeatable precision to within 1mm, but the absolute linear accuracy can be way off. Because it is repeatable, you could a series of calibration points and interpolate between them, so not really a problem. You just need to be aware that you cannot use it as simply as you might expect, if you want absolute positioning accuracy.

ah, would that be why the sixense tuscany demo didn't actually have 1:1 hand positioning when it was in 1:1 mode?

Yeah, you need to create a calibration map between where your hands are as measured, and where you think they are. It could be as simple as displaying a pair of virtual "gloves" and telling you to press a button after positioning your real hands inside the virtual gloves, and repeat for a bunch of positions. Then interpolate between them. After that, no problem picking up real physical props that match the shape and position of objects in VR. That is so obvious (to me) that I am surprised Rift and Hydra demos do not already support such calibration. At best, they ask you to touch your Hydras to your shoulders or something, but we can do MUCH better than that, and save the config parameters for a given Hydra base orientation and position, to not require recalibrating all the time. Ideally, it should be in an SDK that all the apps use, so as not to become such a jumbled mess as apps just do whatever non-standard method the developer decided on. Standards are good, sometimes...

The Hydra also has linearity problems, such that you can measure a position with repeatable precision to within 1mm, but the absolute linear accuracy can be way off.

New hydras are probably going to be more accurate and have better emulation and ratio accuracy. It's obviously got new hardware (wireless, different housing) and new software (SDK).

Let's hope they have good tether range (distance apart) as that is a big issue on dual controller setups.

Rift now just needs official wireless addon for commercial variant and you have fantastic setup with minimal weight.

The linearity can be off due to environmental factors, including base orientation, nearby ferrous objects, which can be compensated for with a calibration map. Electronic devices operating on a close frequency can also cause positioning errors, but these usually appear as jitter and cannot be mapped out. Better to move the base a bit farther away from interfering devices (such as some LCD panels). The new Sixense wireless controllers will still need environmental calibration, which cannot be done in advance at the factory. At a bare minimum, you need to calibrate to a pair of know points or orientations, such as controllers pointed at the base, or controllers touching your shoulders (depending on the game). Calibrating to your shoulders also provides an estimate of player body size (for more realistic VR height adjustment for measured body shoulder separation). But measuring many more known points (touch the VR targets around you with the controllers) would provide much better positioning accuracy.

You can use the SAME displacement map methods for the controller that can use for image warping (a bitmap image or texture).

looks like motion capture exoskeleton is old tech. i ve found various with google picture search.
http://www.metamotion.com/gypsy/gypsy-m ... system.htm
8000$?

colocolo wrote:looks like motion capture exoskeleton is old tech. i ve found various with google picture search.
http://www.metamotion.com/gypsy/gypsy-m ... system.htm
8000$?

Oh man I so have to get one of them before I die, I can see me strapping in and chasing my Collie pup away from our old jack 'GET AWAY FROM HER YOU B*TCH!' (she actually is a bitch btw )

Falan wrote:

colocolo wrote:looks like motion capture exoskeleton is old tech. i ve found various with google picture search.
http://www.metamotion.com/gypsy/gypsy-m ... system.htm
8000$?

Oh man I so have to get one of them before I die, I can see me strapping in and chasing my Collie pup away from our old jack 'GET AWAY FROM HER YOU B*TCH!' (she actually is a bitch btw )

You might enjoy this thread then:
http://www.mtbs3d.com/phpBB/viewtopic.php?f=140&t=16484

Or at least this video from it:
http://www.youtube.com/watch?v=p5wqDWEaPFA

Man I have to stop getting these so offtopic, that's some funny s**t there Geekmaster lol!

Any word on whether this will be licensed through Razer, or will Sixense be doing it themselves? (hopefully the latter... We might get some decent build quality if that's the case).

ElectroPulse wrote:Any word on whether this will be licensed through Razer, or will Sixense be doing it themselves? (hopefully the latter... We might get some decent build quality if that's the case).

Given that we've seen no mention of Razer, and no hint or Razer branding, I think it's safe to say that Razer are probably not involved.

geekmaster wrote:Yeah, you need to create a calibration map between where your hands are as measured, and where you think they are. It could be as simple as displaying a pair of virtual "gloves" and telling you to press a button after positioning your real hands inside the virtual gloves, and repeat for a bunch of positions. Then interpolate between them. After that, no problem picking up real physical props that match the shape and position of objects in VR. That is so obvious (to me) that I am surprised Rift and Hydra demos do not already support such calibration. At best, they ask you to touch your Hydras to your shoulders or something, but we can do MUCH better than that, and save the config parameters for a given Hydra base orientation and position, to not require recalibrating all the time. Ideally, it should be in an SDK that all the apps use, so as not to become such a jumbled mess as apps just do whatever non-standard method the developer decided on. Standards are good, sometimes...

That is definitely ideal, but there's a challenge with the player's orientation changing as they play. Any calibration map is relative to the controller base, which doesn't move. If the player twists around or scootches a bit to the left or right, the map won't be aligned to their real-world location (this isn't so much a problem if you aren't in a wheeled, swivel chair, but that's probably not something you can put in the system requirements ).

That's one of the exciting possibilities for the Stem. If it's clipped to the player's belt, you have another point of reference, aside from the base station, to read the calibration map against.

CosbyTron wrote:That is definitely ideal, but there's a challenge with the player's orientation changing as they play. Any calibration map is relative to the controller base, which doesn't move. If the player twists around or scootches a bit to the left or right, the map won't be aligned to their real-world location (this isn't so much a problem if you aren't in a wheeled, swivel chair, but that's probably not something you can put in the system requirements ).

That's one of the exciting possibilities for the Stem. If it's clipped to the player's belt, you have another point of reference, aside from the base station, to read the calibration map against.

That's the issue I was trying to address in this thread on the dev forums. The positional tracking for the head needs to be tied to the positional tracking of the hands so that, as you said, if they player shifts around in positional space, the head will stay stay correct in relation to the hands. Otherwise it ends up looking a foot off to the left or right or top or bottom. This is why whatever positional tracking Oculus comes up with for the Rift should be incorporated into the positional tracking of any controllers being used for Rift gaming.

Tbone wrote: That's the issue I was trying to address in this thread on the dev forums. The positional tracking for the head needs to be tied to the positional tracking of the hands so that, as you said, if they player shifts around in positional space, the head will stay stay correct in relation to the hands. Otherwise it ends up looking a foot off to the left or right or top or bottom. This is why whatever positional tracking Oculus comes up with for the Rift should be incorporated into the positional tracking of any controllers being used for Rift gaming.

Definitely a good point, but unfortunately, there isn't a reliable way to infer the players head location (or look orientation) from the location of the hands. The Stem could be really useful there. The chest or waist make much better points of reference, both for placing the hands in gamespace and knowing what "forward" is for the player.

TheHolyChicken wrote:

ElectroPulse wrote:Any word on whether this will be licensed through Razer, or will Sixense be doing it themselves? (hopefully the latter... We might get some decent build quality if that's the case).

Given that we've seen no mention of Razer, and no hint or Razer branding, I think it's safe to say that Razer are probably not involved.

I hope that razer aren't involved but there was no hint of razer's involvement with the the original sixense motion controller, even at the demo stage.

This Will from Tactical Haptics and I share the same dream. From the VR meetup.

Sooo...have they said when we can preorder and when they expect to start producing the dev kits?
Maybe i missed that? Can i have i have them soon please?

There,s a bit of work now figuring out how a IK hands and head rig can be made in unity.

IK? What does it stand for

Inverse Kinematics, a common method in cg to rotate the joints making up a limb so that they reach a goal position.
Good calibration of head+hands(+torso) will be needed for a pretty result i imagine.

Inverse Kinematics are also used in Robotics applications.

IK determines how to automatically move all the joints to get the end effector to the specified target destination. A simple visualization model is to imagine how you would move the hands and feet of a rag doll, and how all the joints would automatically follow. But for robotics and motion control, you need to operate all those joints in software. For CGI, you need to calculate them to find vertices for game physics and animation rendering.

Inverse Kinematics @ WikiPedia:
http://en.wikipedia.org/wiki/Inverse_kinematics

Improved IK algorithms:
http://freespace.virgin.net/hugo.elias/models/m_ik2.htm

Comparison of different IK methods:
http://www.youtube.com/watch?v=XdMDuWD8jlk