I primarily take it as a good sign that Valve went pretty deeply into the subject of meaningfully measuring CPU and GPU performance in previous talks I've seen.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
The High-end VR Discussion Thread (HTC Vive, Oculus Rift, Playstation VR)
- Thread starter Cartman86
- Start date
Probably the Rift but if this is really what you're interested in then just spend your money on a nice high res monitor
Do we have a list of the exclusive ones, and ones that are likely 'first on OR'? I'm probably going with the vive because of motion controllers now vs motion controllers in six months, but if too many of the OR launch window games will take months to arrive on vive, that'd be frustrating.
What I mean is, this plugin only affects GPU stuff, so if X is making your frametimes longer, dial back graphics settings until X goes away.
It doesn't matter if X = virus scan started, X = 50 explosions just went off on screen, or X = Unity cpu weirdness. The problem is "my frametimes are going up" the solution is "turn down the graphics"
Adaptive CPU loading would be more difficult because that is stuff you can't usually skip like AI or physics.
The slides from the presentation talked about GPU load rather than frametime so I dunno if it works the way I am describing.
Thanks for the quick answers, folks. I have no experince with VR at all besides watching every youtube video I could find since the Rift was announced. In the latests months I wasn't keeping up with the news so I'm little lost right now, and after watching the Virtual Desktop trailer I wanted to catch up because I really, really liked it.
So as far as I can see, for what I'm asking the Rift is the way to go right now, however I will try to find more about the Vive and what other benefits it could bring me. Neither of them are selling on my country anyway so I'll have to import and after taxes and shippment costs the Vive won't be that much more expensive than the Rift.
Thanks a lot you all.
I already have a nice monitor. I want a VR headset now.
Yes, but that was my point. This kind of simplistic approach would scale down the graphics uselessly in cases like X1 and X3. Only case X2 can actually benefit from e.g. lowered resolution.
There are some tables out there, but none which I know of that I'd describe as complete or up-to-date.
OmarTheHippo
Member
If a game slows because your virus scan starts and takes up 75% of your CPU, dialing resolution back won't help your FPS
If a game slows because Unity hiccups and maxes your CPU, dialing resolution back won't help your FPS
Even if the game magically turned your GPU into a 980Ti 4-way SLI, CPU bound is CPU bound
If the problem is "my frametimes are going up because of CPU" the solution is NEVER "turn down the graphics that GPU controls". The game will have to find ways to scale CPU load
Croatoan
They/Them A-10 Warthog
has Striker VR been talked about here yet?
https://www.youtube.com/watch?v=H2A9xmMZuu4
Imagine a pistol and AR version and holsters for each that show up in VR.
or depending on weight and how they design it, one unit could work like a pistol or AR
https://www.youtube.com/watch?v=H2A9xmMZuu4
Imagine a pistol and AR version and holsters for each that show up in VR.
or depending on weight and how they design it, one unit could work like a pistol or AR
wait, I just remembered that valve addressed this in their gdc talk...
http://www.24liveblog.com/share/193787729?url=http://v.24liveblog.com/live/?id=1322392
If the CPU misses its deadline, the previous frame can be resubmitted to the GPU with updated headset position/rotation information.
I don't know if this was implemented or if it is just a suggestion.
Oculus said they are funding something like 2 dozen titles exclusive to the Rift.
Figure 6 months on the the games that make it to the Vive if somebody doesn't make it happen first. Like all launch software though, you have to figure there's a few gems and a bunch of stuff to hold you over.
It can.
Think of it this way:
During normal gameplay, for each 11ms vsync window, assume the CPU takes 2ms to do its work and the GPU takes 8ms, leaving 1ms of headroom.
If there is a CPU activity spike and the game suddely takes 50% longer to do its CPU work (3ms), the gpu will still take 8ms and hit vsync with no headroom left. The adaptive settings algorithm will see this and dial back the graphics.
The next frame the CPU spike worsens and takes 4ms, but we dialed back the graphics so GPU only takes 6ms. We still do not miss vsync.
The adaptive settings alone could not account for extreme single frame spikes or spikes that drive the CPU slice of frametime to over 11ms minus the time it takes for the GPU to render the frame at the absolute lowest adaptive setting.
Croatoan
They/Them A-10 Warthog
So I have been thinking about current controls (Vive/Touch) and I think certain sport simulations are going to need something different.
Take golf,or baseball bat swinging, for example. I have played golf my entire life, and baseball up until 18 years of age, and I can tell you; swinging a controller with no weight is not only a terrible experience, but can also hurt you and your real life game.
What is missing? That is easy, weight.
Now, you don't need a realistic amount of weight (particularly in the bat case) but you do need a moderate amount.
In my mind you could take something like this (remove the weird bottom part)
run the end through the hole in the vive controller and then securely fasten the vive controller to the shaft. After that you would just have to add weight to taste.
A custom version would have fully adjustable weights (sort of like weighted computer mice) and the vive controller built in. This would make swinging any type of object, from golf clubs to swords, immensely better, and without having to use the real things (which are tool long and dangerous).
IMO, this is needed for any future 1 to 1 golf title.
I might actually attempt this when I get my vive. I have some old golf clubs I could chop up and fit some weights/vive contorller to.
Take golf,or baseball bat swinging, for example. I have played golf my entire life, and baseball up until 18 years of age, and I can tell you; swinging a controller with no weight is not only a terrible experience, but can also hurt you and your real life game.
What is missing? That is easy, weight.
Now, you don't need a realistic amount of weight (particularly in the bat case) but you do need a moderate amount.
In my mind you could take something like this (remove the weird bottom part)
run the end through the hole in the vive controller and then securely fasten the vive controller to the shaft. After that you would just have to add weight to taste.
A custom version would have fully adjustable weights (sort of like weighted computer mice) and the vive controller built in. This would make swinging any type of object, from golf clubs to swords, immensely better, and without having to use the real things (which are tool long and dangerous).
IMO, this is needed for any future 1 to 1 golf title.
I might actually attempt this when I get my vive. I have some old golf clubs I could chop up and fit some weights/vive contorller to.
Zalusithix
Member
Well, g-sync and free-sync are easier solutions as they're implement once and work for (nearly) all. The thing is, they're really solving a different problem - how a display deals with varying frame times. Adaptive scaling solves the problem on how to reliably remain near a target framerate. Both result in the improvement to the user experience, but neither is a cold cut replacement for the other. The best experience will be obtained with both. Scaling to keep the framerate mostly consistent and a variable framerate display to keep up with small fluctuations that the scaling can still let slip through.
My whole point is that traditional games are not what you want to be playing in VR. They are not the best use of the technology and worse are far more likely to make you sick. As you say you haven't experienced VR so how do you know that you want it for traditional seated games? That's how I felt before I got my DK2 but after two years of VR use I now only want to play purpose built VR experiences.
Also, Virtual Desktop is nothing like how it looks in the video. The resolution is so low that it's useless for 80+% of the things you would do on a desktop. You are imagining that VR is just another display but it's not. If this is what you are going for you will be terribly disappointed especially if you already have a nice monitor.
plagiarize
Banned
Get one. Don't listen to people downplaying how great traditional gaming is in VR. To steal a term from Tested that really resonated with me...
VR is like headphones for your eyes.
Sleeping Lesson
Member
Someone on Reddit is reporting their Rift status has now changed to "shipped". They're a Kickstarter backer.
The Elite
BOSS
plagiarize
Banned
Gear VR is the best way to watch Netflix in bed. I can do it without the light from a screen disturbing my wife. I am isolated from other visual distractions.
The parallels to headphones are pretty clear.
The Elite
BOSS
Hmm. I guess in the long run I should consider a second Gear style headset for the iPhone. I hope someone comes out with one. I think the only option right now is to use cardboard.
do they do cardboard for iphone too?
The best hope is that oculus and htc/valve work together and add official vive support for the oculus store.
Barring that, it should be possible to trick the oculus store into working with a vive, but the added layer of software between the two would have to compromise performance to some degree.
Croatoan
They/Them A-10 Warthog
So I thought more about this and realized it wouldn't do to just to put weight on the end. Swords and baseball bats are more evenly weight distributed while golf clubs do have their weight on the end. So I think something like this quick PS mock up would be better (Not to scale).
This wouldn't be that hard to prototype if the Vive controller will still work with something metal sticking through its doughnut hole.
I am thinking the actual length to be about 1.5 feet to 2 feet.
This wouldn't be that hard to prototype if the Vive controller will still work with something metal sticking through its doughnut hole.
I am thinking the actual length to be about 1.5 feet to 2 feet.
Zalusithix
Member
So I thought more about this and realized it wouldn't do to just to put weight on the end. Swords and baseball bats are more evenly weight distributed while golf clubs do have their weight on the end. So I think something like this quick PS mock up would be better (Not to scale).
This wouldn't be that hard to prototype if the Vive controller will still work with something metal sticking through its doughnut hole.
Keep in mind the controller is where the hand is going to be in the VR world. Placing the controller on an extension pole away from the hand is bound to have unintended results.
plagiarize
Banned
I still think that's a fairly decent worst case scenario, but it will also depend on how regularly Oculus update home, and how regularly that breaks any wrapper solution.
Even without my free headset I was planning on getting the Rift because I am eager for some of the exclusives. I'll dabble in room scale with touch and go from there... but Edge of Nowhere, Lucky's Tale, Chronos, Pinball FX 2 VR and Adr1ft are selling me on the Rift.
If Valve don't want to prevent me from playing Steam VR games... I mean, I'm certainly not going to complain... and hopefully the wait for Touch won't be too difficult.
I know a lot of people are pissed off about Oculus locking their store down in principle, but then I've never really been upset by exclusives. Of course Valve want to let everyone buy VR content from their already really successful store. Of course Oculus want to give people reasons to go to their store instead.
Neither is a bad position. It's just a major bummer that both companies can't align their goals enough to make the Vive work with Oculus Home.
Croatoan
They/Them A-10 Warthog
cant that be programmed differently on an game by game basis? Or is it somehow hard coded?
Would you still work with metal sticking through your doughnut hole?
So I thought more about this and realized it wouldn't do just to put weight on the end. Swords and baseball bats are more evenly weight distributed while golf clubs do have their weight on the end. So I think something like this quick PS mock up would be better (Not to scale).
This wouldn't be that hard to prototype given that the Vive controller will still work with something metal sticking through its doughnut hole.
Two things, the Vive controllers have that hole to help limit occlusion, something you would be potentially introducing with your design and secondly I wouldn't trust the wands to be able to withstand the forces you'll be applying to them in that setup. If you're trying for 1:1 for a sport, both baseball and golf have a ton of force being applied to the bat or club based on your swing. I don't think the Vive controllers' little corner thing could handle that.
plagiarize
Banned
Certainly it can be programmed for. If your peripheral grip thing was standardized a game could add support for it with a toggle of some nature... but the presumption is going to be that your hands are on the Vive controller grips.
Croatoan
They/Them A-10 Warthog
What about this?
Now the vive doughnut is where a Pommel would be or where the bottom of your hand in a golf grip would be.
Aren't the sensors on the outside of the hole? Also, no stress should be placed on the Vive controller. It is attached firmly to the shaft which should hold the force.
Obviously a custom device built for this would be better, i am just trying to brainstorm a prototype.
OmarTheHippo
Member
Why would I think of it that way? Pretty sure that's not how it works at all
If the CPU takes 2ms to do its part of the job, and the GPU takes 8ms to do its part of the job... the job took 8ms
I'm sure there are technical caveats, but in general CPU and GPU perform their work in parallel (not to be confused with the type of work each is doing)
Could go too many places with this. VR role playing is the new frontier.
The Elite
BOSS
Yeah, there are a bunch of cardboard programs on iPhone. Works fine with my iPhone 6 Plus.
There is nothing out there like GearVR for iPhone though.
I hope someone creates a solution until Apple gets into the VR game.
Dreams-Visions
Member
Begun, the VR revolution has.
Except headphones reproduce music with just as good if not better quality than most speakers. VR headsets have far lower resolution than most displays.
So in terms of quality
Headphones >= Speakers
VR headsets <<< Traditional display
Is not so cut and dry
Dreamwriter
Member
To be fair, you are using a DK2 which has a much lower resolution and more wasted space than either Oculus Rift or HTC Vive.
Sleeping Lesson
Member
A few other backers are reporting their's has shipped too. I'm trying to be happy for them, but I'm jealous that mine probably won't ship until next week.
plagiarize
Banned
VR headsets reproduce environments in a much better way than looking at them on a flat screen. VR headsets are only <<< than traditional displays if they're showing 2D rectangles of content.
Say I reproduce your room on a computer. What device will more closely convey that environment to the original thing?
The increased resolution is only about 11-12% in each dimension and 25% overall. It's better but not remarkably so. I remember this time a year ago everyone thought that the CV1 would be at least 1440p.
plagiarize
Banned
I'm a backer. Mine hasn't shipped yet.
Wait peasant.
And look 'shipped' here almost certainly means in a big pile ready for the FedEx trucks to arrive. They'll likely process all the day one orders before calling FedEx to come pick up the palettes.
The Elite
BOSS
Do we know it's going to be FedEx? I have heard that previous dev units were shipped by UPS.
They're just gonna leave it at my door. The VR revolution stolen by my neighbor.
metareferential
Member
I have one of the best monitors around and I'd still trade it for VR any day in a heartbeat.
1:1 scale and presence > resolution.
And the great thing about VR is that one day we won't talk about resolution anymore.
Right, and this is my entire point. VR headsets optimal use is for purpose built VR applications and not for replicating traditional display uses. If that's what people are doing they will likely be disappointed because the experience is not nearly as good due to low resolution.
I am oversimplifying but there are some things the CPU has to do before the next frame is drawn. Updating the camera position based on user input, processing user input such as button presses, animation, etc.
AbortedWalrusFetus
Member
Are there any videos demonstrating how the Vive Chaperone reacts if another person or a pet walk into your play area?
Croatoan
They/Them A-10 Warthog
They are ejected into space.
This feature is not in chaperone as of now
plagiarize
Banned
Just about any 3D environment or model is going to look better on a VR headset whether you can walk around it in 1:1 scale or not. Ravenholme wasn't made for VR, but I found it much more impressive when I played HL2 in VR. Heck everything in pCars and Dirt Rally was made for traditional displays too.
I have no idea what they are using for shipping, no.
plagiarize
Banned
chaperone doesn't monitor what's inside your tracking volume. I could sneak in and put a shin high coffee table in front of you and you'd smash into it without knowing it was there. What it does is keep on eye on the tracked headset and controllers and let you know when they are reaching the edge of that predefined tracking volume.
You set up the limits of your volume when you first setup Steam VR.