A dedicated backtop computer for VR - worth it?

This is for discussion and development of non-commercial open source VR/AR projects (e.g. Kickstarter applicable, etc). Contact MTBS admins at customerservice@mtbs3d.com if you are unsure if your efforts qualify.
User avatar
V8Griff
Sharp Eyed Eagle!
Posts: 450
Joined: Fri Mar 01, 2013 11:22 am
Location: UK

Re: A dedicated backtop computer for VR - worth it?

Post by V8Griff »

Very interesting discussion on a technical level but have you considered what happens when you've pulled on your HMD and fully immersed yourself in your 3D fantasy world?

As you stumble around the real world and bang into walls, knock the TV off the wall, trip over the cat and even worse maybe fall down the stairs? :shock:

I mentioned in a post elsewhere that I used to operate Virtuality equipment (both 1000SD and 2000SU units) for Corporate Hospitality and Marketing for more than 10 years so I've probably put more people into virtual environments than anyone and it was amazing how many people became completely dis-orientated and were amazed when we took the HMD off to find they had turned through 180 degrees when immersed in an environment.

If the Rift is as good as we're all hoping it is I think many of you will be surprised just how immersed you will be and as a result be locked out form the real world. We rely very heavily on peripheral vision to orientate ourselves in the real world and how losing those subconscious references confuses the brain.

Obviously everyone is different but in my experience 95% of new players are initially disorientated and suffer from what I called PC monitor syndrome in that they don't tend to turn their head to look around and even when they do it's something they find unnatural.

All these systems that people are discussing to simulate or facilitate walking will be totally unnecessary (certainly in the early days) and in my opinion it is quite dangerous to be able to free roam.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: A dedicated backtop computer for VR - worth it?

Post by geekmaster »

V8Griff wrote:... Obviously everyone is different but in my experience 95% of new players are initially disorientated and suffer from what I called PC monitor syndrome in that they don't tend to turn their head to look around and even when they do it's something they find unnatural. ...
I noticed that in one of the posted Rift prototype demo videos. The guy complained a bit about dizziness, but after he was handed the controller, we quickly stopped turning his head and just looked straight ahead, while turning his viewpoint with the controller. I think that moving his head to control turning might have helped prevent the dizziness, but I have not been in VR enough yet to know for sure. Perhaps head turning without vestibular feedback could actually be worse than turning with a controller.

I am certain from experience that the queasiness can be overcome with repeated sessions of slowly increasing exposure to immersion, being careful to not go past that point each time or the adaptation to VR will be reduced.

I agree that unconstrained free roaming will be dangerous for some people who are not prepared for it. One way to avoid tripping over obstacles is to make sure your VR world incorporates them into it, so you naturally avoid it. A nearby stairway could appear as a pit (or even a real stairway) in a VR environment. Having a head mounted depth sensing camera (like a Kinect) that controls the position of obstacles in the VR world would increase its safety. But even without that, you can model your VR environment to match the Real World around you, and you can put props in your Real World to match virtual content. Such intersections between VR and RL can also provide a sort of haptic feedback, which increases both immersion and safety.

I plan to free roam in my home, but only after I have modelled it sufficiently and incorporated obstacles in the game to protect me from bumping my head. The VR world will tell me when to duck in the Real World...

Another place to free walk in VR would be on a football field. I would not carry a realistic looking gaming weapon though, to prevent unwanted intrusion from SWAT teams or other immersion-reducing personnel.
User avatar
V8Griff
Sharp Eyed Eagle!
Posts: 450
Joined: Fri Mar 01, 2013 11:22 am
Location: UK

Re: A dedicated backtop computer for VR - worth it?

Post by V8Griff »

Queasiness is not something I saw a lot of surprisingly and the Virtuality systems had much bigger latency issues than the Rift tracker is reporting iirc 60Hz updates with 12 millisecond latency which is way slower.

I'd say physically turning your head in the environment with the HMD is the right way as you'll be likely to make yourself throw up and quickly become disorientated if you used a hand controller rather than your head. I think keeping it 'real' is the way to go, so allow your head to do what it does in the real world and the same for your hands with tracked peripherals such as the Razer.

I still question the validity of free roaming and I think you'll see what I mean once your strapped into your Rift, that said I like your idea of mapping your home environment into your virtual worlds but I'd still question why when there's all these marvellous fantasy worlds to interact with?
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: A dedicated backtop computer for VR - worth it?

Post by geekmaster »

My Virtuality experience (a Mech War simulation) was that I had a lot more problems with "wall humping" than with motion sickness. Just as I FINALLY figured out how to use the controls with any degree of usefulness, my time was up, and it was too expensive to try it again. This was back in the 90's and I was the only person in our group willing to cough up the $35 it cost to play the game at that time. It was sadly disappointing, mostly from not having time to learn the controls first.

Wall humping (NPC in this example):
[youtube]http://www.youtube.com/watch?v=nL18INzP2ps[/youtube]
User avatar
V8Griff
Sharp Eyed Eagle!
Posts: 450
Joined: Fri Mar 01, 2013 11:22 am
Location: UK

Re: A dedicated backtop computer for VR - worth it?

Post by V8Griff »

geekmaster wrote:My Virtuality experience (a Mech War simulation) was that I had a lot more problems with "wall humping" than with motion sickness. Just as I FINALLY figured out how to use the controls with any degree of usefulness, my time was up, and it was too expensive to try it again. This was back in the 90's and I was the only person in our group willing to cough up the $35 it cost to play the game at that time. It was sadly disappointing, mostly from not having time to learn the controls first.
$35!!! Ouch that was expensive. :o

Exorex or Heavy Metal was the name of the game, I loved that one but as you say could be confusing on first encounter, especially if the controls were not explained properly. We always explained teh controls and tried to help players as we weren't in an arcade situation and always wanted players to have a good time.

The controls on the joystick were forward and back and spin left or right, the triggers fired the laser weapons and the top buttons brought up a plan view overlay of the arena. The head controlled the weapons and your view, if you kept your gaze steady on a target it launched a rocket. Some machines had vanity covers with steering wheels and pedals, even more confusing.

The confusion you encountered was typical as you weren't necessarily going the way you where looking so if the joystick was pushed forward and you looked to the left you had the situation akin to you looking through the front screen of your car while looking out of the side window. We often had people walking away from the main arena while screwing their heads round trying to keep the weapon sight on the targets. Again funny to watch but frustrating inside the game.

It's all related to the confusion I mentioned.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: A dedicated backtop computer for VR - worth it?

Post by geekmaster »

V8Griff wrote:The confusion you encountered was typical as you weren't necessarily going the way you where looking so if the joystick was pushed forward and you looked to the left you had the situation akin to you looking through the front screen of your car while looking out of the side window. We often had people walking away from the main arena while screwing their heads round trying to keep the weapon sight on the targets. Again funny to watch but frustrating inside the game.
It certainly was frustrating enough that I never went back to try again, although I would have tried again if I wasn't afraid of wasting money in frustration again, enough to buy groceries for a week!

However, with my level of hard-wired technolust, I never lost my DESIRE for good affordable VR. I am looking forward to my Rift, with more than a little PTSB.

This time around, brought to the world by the affordable Oculus Rift, we absolutely MUST prevent VR novices from having the frustrating experience I had with my first hands-on experience in VR. It is up to us who provide VR content to be sure we do it right this time, now that we have the tools to actually do that.
User avatar
brantlew
Petrif-Eyed
Posts: 2221
Joined: Sat Sep 17, 2011 9:23 pm
Location: Menlo Park, CA

Re: A dedicated backtop computer for VR - worth it?

Post by brantlew »

VR control interfaces are a tough nut to crack. From my experiences so far, the biggest contributors to queasiness are "disconnected" spinning. When my body (not my hands) is the cause of my spin then I experience no uneasiness - even when stomping around and shaking the view without any form of positional tracking. That is far more tolerable than wildly swinging the view with a controller. VR developers may be forced to "slow down" the content and dial down controller sensitivities to cope with the limitations of the human body.

Ultimately I think a 360 setup feels much more comfortable and will be highly desirable. It's just a whole lot less confusing to use a single intuitive interface for turning instead of two. At first it seems people either want to spin around and yank the wires out of your computer or they look straight ahead and just use the controller - but not both. It takes practice to learn to use the two together effectively. But 360 requires either tricky wiring, wireless solutions, or worn computers.
Last edited by brantlew on Tue Mar 05, 2013 11:31 am, edited 1 time in total.
User avatar
V8Griff
Sharp Eyed Eagle!
Posts: 450
Joined: Fri Mar 01, 2013 11:22 am
Location: UK

Re: A dedicated backtop computer for VR - worth it?

Post by V8Griff »

Yes I agree, he best way to induce nausea would be to give head control to the joystick.

Virtuality managed to more or less sort the 360 degree thing with the SU units by having a long(ish) cable for the HMD mounted behind you with the handcontroller to the front with a 'curly' cable to allow for the greater movement.

In my observations very few people ever tried to do the full 360 and tended to turn back the the way they came after going through 180.
geekmaster
Petrif-Eyed
Posts: 2708
Joined: Sat Sep 01, 2012 10:47 pm

Re: A dedicated backtop computer for VR - worth it?

Post by geekmaster »

brantlew wrote:... When my body (not my hands) is the cause of my spin then I experience no uneasiness - even when stomping around and shaking the view without any form of positional tracking. That is far more tolerable than wildly swinging the view with a controller. ...
I did not realize that. I get queasy with any head motion with my HeadPlay HMD, just watching a movie. But there, it uses sequential color. And moving your head leaves a perceptual trail of the individual color planes. This is because sequential color displays each color at a different time, but they were taken from a full color frame from ONE point in time. Really, the individual color planes need to be INTERPOLATED for a much higher framerate, so you brain can see continuous motion as the sequential planes appear in their correct intermediate POV locations.

A quick experiment with your "body controlled spin" idea: I just moved my head around quickly while gazing at a portion of my eyeglass frames. Those could well have been video content instead of real objects stuck to my head. But no uneasiness at all.

These to experiences taken together say that my HeadPlay queasiness was caused by sequential color flicker problems, with color plane content not tracking my head movement.

The the LCD in the Rift (no personal experience with that), the video in it would be just like looking at my eyeglass frames while turning.

That means that my ideas for decoupling head-tracking from rendering will not be that important to suppress motion sickness, but they would still be useful to increase perceived immersion.

EDIT: It seems that motion sickness is a learned experience that can be remembered and felt again. Lately, while writing about it and envisioning in my mind some motion sickness inducing experiences, I felt on onset of that very symptom. In fact, I feel it again now, and all I have been doing is typing. Unless this is residual from looking at my eyeglass frames while moving my head, but if so, it is a delayed reaction...
User avatar
love2scoot
Two Eyed Hopeful
Posts: 56
Joined: Fri Jan 06, 2012 4:43 pm

Re: A dedicated backtop computer for VR - worth it?

Post by love2scoot »

Looking at the reality of our research, it was apparent that to achieve a level of performance required for 60Hz gameplay, we would need to assume a slightly higher total system power usage. This really comes down to one item, the battery. We need to affordably expand battery capacity if we were going to reach the goal of 3 hours of gameplay without making this system impractically expensive.

After a bit more research, I was able to find this battery:
http://www.amazon.com/BiXPower-Capacity ... 090&sr=8-1
At a slightly higher capacity of 153Wh it has a reported 28Wh more capacity than our PPoP above but at a higher cost per Wh. After brainstorming with some peers, I think an alternate solution is within reach that is both cost effective and offers additional benefits.

For a first iteration, I propose using two PPoP battery packs, running in parallel with simple diodes on each line. This provides at least 3 advantages:
  • 1) We almost double battery capacity, while being able to keep our dollar per Wh cost low by using the PPoP packs.
    2) Since one of these packs will likely have a *slightly* higher regulated voltage, one of these packs will discharge first, and the second pack should seamlessly take over when the primary battery hits a cutoff voltage on its internal pack. This opens the door for a redundant / hot swap system.
    3) Having two smaller battery packs would simplify mounting as it will be easier to mount two smaller packs than one monolithic pack.
On the downside, the diodes will drain a small amount of power based on the current drawn from the system. This has two ramifications:
  • 1) Our overall battery life will be reduced slightly when using this system
    2) We will need to ensure the diodes are mounted in such a way that they can shed their excess heat. This should be as simple as surface mounting them on a small bit of aluminum or putting them within the path of a fan.
One other item to note here: although the bixpower battery specifies a max current of 4.75A, we do not appear to have a current limit specified on our PPoP. Assuming the small loss incurred by the diodes, (2) PPoP packs would allow us a total capacity of ~240 Wh total in a best case scenario. If the PPoP could maintain ~4A of current, we can expect 3 hours of gametime with a peak system power of ~75W (including HMD). I think this gets us to our goal, while allowing us to purchase one battery at a time during the initial system build. This is certainly a less expensive route that allows for more moderarte expenditures while building and testing our untethered solution, and allows us to scale to greater gameplay time in the future.

Long term, there is probably a more elegant solution than the diodes that would incur a smaller overhead. At this point, however, the diodes should be an efficient enough solution to get us past a proof of concept.
STRZ
Certif-Eyed!
Posts: 559
Joined: Mon Dec 05, 2011 3:02 am
Location: Geekenhausen

Re: A dedicated backtop computer for VR - worth it?

Post by STRZ »

@ love2scoot

Nice find, the battery looks interesting.

Btw, AMD said that they'll sell a cut down version of the PS4 APU later this year http://www.theinquirer.net/inquirer/new ... tion-4-apu

PS4 = 8core Jaguar @ 2ghz + GCN GPU on the same DIE in the range of a 7870. If they cut it down to 6 or 4 cores paired with a GCN GPU in the range of a 7770 to 7850 with low power consumption, this could be what whe're waiting for to put in our backpacks!
User avatar
love2scoot
Two Eyed Hopeful
Posts: 56
Joined: Fri Jan 06, 2012 4:43 pm

Re: A dedicated backtop computer for VR - worth it?

Post by love2scoot »

I did see a bit of news recently on this, and it certainly does look interesting. Bobcat cores are a little underpowered (compared with Steamroller cores), but it appears that the Jaguar cores offer a significant jump in performance over Bobcat. Here was an interesting thread that goes deep into this: (http://beyond3d.com/showthread.php?p=1668471). Assuming they can pack 4-6 of them on a die with a robust GPU and the TDP held to 65W (45W preferably) we might have something here.

Based on AMDs current offerings, we certainly could see this within the power level we need. The current Brazos 2.0 APUs (http://en.wikipedia.org/wiki/List_of_AM ... .282012.29) are manufactured on a 40nm process, are dual core and have a 18W TDP. Doubling or tripling the cores and ramping the clock should be counteracted by the HUGE shrink (holy oxymoron Batman!) down to 28nm. It also appears that the Jaguar cores scale well on performance when adding cores, but I think the key here will be hitting a high enough clock speed such that single threaded code (like games) runs well. Exciting stuff indeed!

There was some banter earlier in this thread regarding real world collision avoidance while using untethered VR. This is an area that I can speculate on but I was hoping to gather wisdom from the likes of brentlew et al. Basically, I envision using a joystick for forward / back & strafe left / strafe right and using the head tracker for a 360 degree view. To contain human drift I was considering a slightly raised circle with a ~1.5m diameter (like a big hula hoop with beveled edges) stuck to a wood floor. I don't expect any large steps while using the platform, so I think it's safe to assume that the small lip would not be stepped over. Also, if a circle is used it should be easier to recenter one's self if an edge is encountered. I was considering a small bump (like a subtle dome) at the center of the circle to ease re-centering w/o line of sight. While significantly more elaborate solutions exist, I'd like to explore something simple and elegant that would require low materials cost, a short setup time, and above all would ensure the safety of the user. The circular floor lip was my best candidate. Any untethered vets (or anyone else for that matter) care to comment?
El_Fitz
One Eyed Hopeful
Posts: 1
Joined: Sat May 23, 2015 3:33 pm

Re: A dedicated backtop computer for VR - worth it?

Post by El_Fitz »

Hi !

It's quite funny to notice that we've been having the same idea, at the same time, ending up with the same conclusions...
Only that it now seems the CV1 will require a GTX 970.

So I don't know what you guys think, but it seems that both the APU and the laptops are out of question, performancewise.
But I seriously can't really figure out how to power what would be a full fledge pc on my back...

The weight isn't really an issue, as I'm looking into a way of not having to carry the weight (involves an Omni / a Virtualizer and a way of hanging the computer from the ceiling, but still attached to the back as a backpack), but power supply is (after all, a 970gtx consumes 245 watts).

Worst of all, the battery outputs DC power... A regular pc PSU requires AC input.

Oh, yes, if I'm hanging the computer from the ceiling, I could just hang a power cable from there too, true. But I like the idea that it could be moved around, and used anywhere (if the weight remains reasonable).


Anyway, I was just wondering if anyone else was still looking into this. If not, well, I guess I'll keep digging on my own !
Post Reply

Return to “VR/AR Research & Development”