question on oculus rift implementation

[SR-F_Winger]

Hello devs. A possible issue that came to my mind when looking at the specs of the devkit 2.0 is how will positional tracking react once the head is turned around 180 degrees to check six and the camera looses track of the LEDs that are mounted at the front of the DK? I know rotation will nnot be kep track of by the LEDs so this is covered. Maybe adding some LEDs manually on the headstrap that keeps the rift in place?

Any thoiughts on that? Maybe can be adressed softwarewise by how the view reacts once track of LEDs gets lost?

 

[DD_bongodriver]

at this stage we don't even know it needs addressing, at least on the crystal cove prototype it was clear that IR led's were mounted on the side of the unit that would stay in sight of the camera, given that when checking 6 the body is not fully turned 180 and a large portion of the side is still exposed to the camera.

[SR-F_Winger]

at this stage we don't even know it needs addressing, at least on the crystal cove prototype it was clear that IR led's were mounted on the side of the unit that would stay in sight of the camera, given that when checking 6 the body is not fully turned 180 and a large portion of the side is still exposed to the camera.

Youre right. I think we´ll see if it becomes an issue.

[Bando]

You guys think it will make a customers version this year?

[SR-F_Winger]

You guys think it will make a customers version this year?

Guess not. At the very best at the beginning or middle of 2015.

But hey. We get the DK2, dont we:)

[DD_bongodriver]

You guys think it will make a customers version this year?

 

Almost certainly, I think Oculus said themselves that the CV would follow soon after DK2 and I think the timeline was always for Q4 2014 and it seems like they are making the deadlines.

Edited March 19, 2014 by DD_bongodriver

[JG27_Chivas]

I think they want to have the consumer version available for the end of the year, but can't commit for obvious reasons.   The DK2 is suppose to allow developers design their games with a prototype close to the consumer release version.   That said everything will have to go right to make the end of the year release possible.   Anyone who has ever developed anything new, knows, there are always setbacks.

[vapor_tales]

LED's will be on the sides of the rift, This will allow a large sweep of head rotation, if not quite fully 180.  Think about how often one turns his head a full 180 degrees in a fighter plane:  Never.  We can't see through the seat or armor plating.  About 120 degrees is the limit. Need more? Rudder back and forth to swing the tail side to side. This will not be a problem for flight sims, IMHO.

[HagarTheHorrible]

I'm confused, as far as I know, the headset does all of the movement tracking, up, down, left, right, the whole nine yards.  The camera and LED's are nothing more than a check base device to correct for any drift from the starting position, a bit like the re-center key on early TrackIR.  Because of this the LED's and camera don't need a continuous connect, high refresh rates are nice but not essential to the smooth re-centering of the device and continuous looking backwards over one shoulder for anything but very short lengths of time is very unlikely, if only because it would be physically uncomfortable.  Why do people not think this to be correct, what do you know that I don't or have I just being following the development of the OR with false assumptions ?

[DD_bongodriver]

No, the camera and IR led's are for the positional tracking but will have a usefull side effect of providing fixed reference for calibration.

[JG27_Chivas]

The OR has internal tracking, BUT the internal tracking had a tendency to drift without a perfect reference point.   The OR development found the best solution "so far" to be the external camera.  Now they combine the data from the internal tracking and the external camera tracking to provide the best possible 6DOF tracking at this point.  This could still change, but there has been no rumor recently to suggest they've found a better solution.

Edited March 22, 2014 by JG27_Chivas

[FuriousMeow]

6dof isn't done internally, just up/down and left/right is done by internal gyros. The lateral and vertical strafe and zoom require external reference points as gyros are incapable of doing those axis.

 

You can physically turn around more than 180 degrees in a plane. The shoulder straps aren't anchoring the pilot down. They aren't strapped down like in a racing car/truck. A lap belt is all that's needed to keep a pilot in his seat in inverted flight, that's how it was done in ww1. The pilots only fell out when they didn't fasten that belt.

Edited March 23, 2014 by FuriousMeow

[DD_bongodriver]

Rotational movement is done by the internal gyro/magnetometer, positional movement is tracked by the IR camera and sensors.

 

you cannot turn around more than 180 degrees in a plane, shoulder straps or not your ass is still pinned to a seat that does not rotate, your head might make something close to 140 degrees from forward axis and you use a component of peripheral vision to check six, inside a small single engine fighter cockpit you still have a canopy and structure that may affect a full range of motion, if you try to turn your head as far round as you can in the computer chair at home you will notice just how much your body needs to lean, in a cockpit you don't get that luxury.

[FuriousMeow]

I have a torso, I've done it in a plane. So yes, you can rotate your head back more than 180 degrees.

[DD_bongodriver]

I have the full complement of male human anatomy so I can't, my legs still have to point forward, my shoulders cannot rotate beyond approx. 70 degrees, my neck can do more or less 90, 70 + 90 = 160, studies have shown the average to be 160 degrees for women with men achieving less.

 

I also have tried it in the many aircraft I have flown.

Edited March 23, 2014 by DD_bongodriver

[DD_bongodriver]

I guess the point to convey here is that with my DK1 I am able to see as I would in real life the full extent of my body rotation and don't find the lack of positional tracking to be an issue, when twisted that far there is no need for positional tracking

[FuriousMeow]

I did it in a j3. Lean opposite the direction you are turning, lean slightly forward if necessary, rotate torso in the direction opposite the lean, use hand not on stick to grab leg for additional leverage if necessary, and already you will have an additional ten to twenty degrees of rotation before even turning your neck.

The more than 180 is assuming left is zero and right is 180. If you can't look more than just right or left in a plane then there is a physical handicap at play.

[DD_bongodriver]

Thanks for that.....I just went and twanged a lat, yes maybe a few more degrees doing it that way but to be honest if I had an imaginary line pointing out of the front of my face it still wouldn't have made it past 180 degrees.

 

 

 

where's that voltarol?

 assuming left is zero and right is 180. If you can't look more than just right or left in a plane then there is a physical handicap at play.

 

? no I'm assuming12 o'clock is 0 and 6 o'clock is 180

[FuriousMeow]

Most assume that the 180 degree of vision is the front half of a viewing circle, not the side half... At least those that claim you shouldn't be able to see greater than 180 degrees, that's why they argue you shouldn't be able to see behind you because of the shoulder strap argument. It appears we are not on the same page, yes, you should be unable to Linda Blair yourself. That's clearly physically impossible.

[DD_bongodriver]

Most assume that the 180 degree of vision is the front half of a viewing circle, not the side half... At least those that claim you shouldn't be able to see greater than 180 degrees, that's why they argue you shouldn't be able to see behind you because of the shoulder strap argument. It appears we are not on the same page, yes, you should be unable to Linda Blair yourself. That's clearly physically impossible.

 

I see, yes clearly different pages, if the 180 deg is the front hemisphere then nobody need be concerned by the IR part of positional tracking, the DK2 has sensors on the top and sides that should stay visible to the camera well beyond that scope.

 

Heres the DK2 naked showing the array of IR led's covering the unit both sides and top

 

Edited March 23, 2014 by DD_bongodriver

[HagarTheHorrible]

I'm not going to plump for the DK2, but will be very jealous of those who do and have it working in BoS. It will be interesting to see what, if any difference it will make to visual contact retention and if the movement of bogies will be easier to spot against the background.

I have a torso, I've done it in a plane. So yes, you can rotate your head back more than 180 degrees.

I'm glad you explained that your 180 is left, right and not front to back. I was beginning to think that maybe there was more to you than just your name and picture

[Venom]

No, the camera and IR led's are for the positional tracking but will have a usefull side effect of providing fixed reference for calibration.

The OR has internal tracking, BUT the internal tracking had a tendency to drift without a perfect reference point.   The OR development found the best solution "so far" to be the external camera.  Now they combine the data from the internal tracking and the external camera tracking to provide the best possible 6DOF tracking at this point.  This could still change, but there has been no rumor recently to suggest they've found a better solution.

That are two very accurate answers!

 

I use(d) TrackIR from version 3 to 5 and was never really happy with it. From the very first minute I found the OR's internal magnetic tracking MUCH MUCH better. Forget all your worries about checking six, turning 180° or 360°: the rift allows you to turn (or rather rotate) UNLIMITED ─ or at least until you get nausea or get strangled by its cable! (Safety hint: try or prove it by swirling the OR in your hands.)

 

Believe me, already in the DK1 check-six works PERFECTLY. In fact, I started to check six in WT regularly only after my DK1 arrived ─ unfortunately as soon as I play with monitor and TrackIR I drop the habit...

 

With DK2 we get, so to speak, «the best of two worlds»: a very accurate and unrestricted magnetic tracking for 3 DOF and a (much better and more versatile than TrackIR!) Infrared/camera system cabable of 6 DOF. In combination and when constantly chrosschecking and equlilibrating their respective inputs and thus allowing the automatic recalibration of the magnetic reference lines it will satisfy every need of the best BoS fighter pilot!

So if we speak of a «better solution», that's probably only in terms of easy setup and comfort, and not in terms of quality.

 

I can't wait until we get rift support in BoS!!

Edited March 23, 2014 by Venom

[JG27_Chivas]

You can turn past 180* with no problem..  There is still the internal tracking data and you've only momentarily lost the external tracking data.  This is not a problem.

Edited March 23, 2014 by JG27_Chivas

[SR-F_Winger]

Almost certainly, I think Oculus said themselves that the CV would follow soon after DK2 and I think the timeline was always for Q4 2014 and it seems like they are making the deadlines.

i certainly hope that too.