I think with my room layout having a section with a lower ceiling I might go for a mixed use setup. For games like Selfie Tennis where I might be reaching overhead and making fast movements, I'll limit the space to the 2m x 2m area that has a higher ceiling. Then for things that are a little slower paced or don't need reaching up high, I can extend that out to 3m x 2m under the lower beam.
The High-end VR Discussion Thread (HTC Vive, Oculus Rift, Playstation VR)
- Thread starter Cartman86
- Start date
Dusktildawn48
Member
GRRR, by the end of the year for release date. this is probabaly my most anticipated VR game so far. Guess ill have to play the demo a bunch of times.
midasmulligan
Member
Dammit, I'm torn, cause eveytime I get hyped and hear that the screen door effect is nil and the resolution per eye makes for a completely immersive experience (on the vive) I later hear that playing a game in these devices is like "watching a YouTube video at 480p" which sounds atrocious. Is this going to be the case where I am blown away initially just because it's VR, and then come two months down the road I'm like "it's a nice proof of concept, but I'd rather play these games on a badass monitor instead"? Just wish I had access to a demo unit...
Zalusithix
Member
If you were in a virtual room the same size as the existing one, it should look about right for what you'd expect. If you're in a large area in VR though, there would be a mismatch as things get extrapolated out. This would be a problem with any tracking system really.
You could always use gyro stabilization of the lighthouse units, though the cost of doing so wouldn't be insignificant.
If you have a hard on for super high resolution I'd imagine you should just hold off. I haven't tried VR yet but I understand it isn't going to be 4K quality. Then again I still play Commodore 64 games on a C64 monitor, so I have lower standards when it comes to pixels per inch.
If getting yourself used to the idea of playing on a lower quality screen is too much, maybe hold off so you don't get buyers remorse. VR isn't going anywhere.
I think it's more the opposite. At first you're like "this isn't that great looking.." but within 10 seconds you forget about the image quality and you are in the experience. The screen door effect on vive is not a big deal and I only saw it when I took a second to stop and assess the image quality. Really it's just the resolution. It's not a whole lot of pixels to cover almost the entirety of your vision.
I've only had 10 minutes of VR so it's impossible to guess how my interest will hold up after hundreds of hours of use. It doesn't feel like I will get tired of it as long as there is a steady stream of games.
Gen 2 or 3 will be the turning point where there will be no reason to not play everything in VR, at the very least for perfect stereoscopic 3d on a virtual 2d screen.
midasmulligan
Member
I see where you're coming from, and I've played plenty of sub HD resolutions games--but when it comes to VR games--I'd imagine resolution is a corner stone to providing a reasonable VR experience. I'll hold off judgement until I can get my hands on a unit, just sharing my thoughts
Zalusithix
Member
The best examples of "these games" are going to be things you simply can't experience on a monitor. The ones designed around VR from the ground up and deliver a solid feeling of presence. This is where room scale / standing experiences with tracked controllers should come into their own.
plagiarize
Banned
The youtube part is definitely nonsense, because there's no compression going on. It's really hard to ball park what effective resolution you're seeing in the sweet spot though, to be honest. I know on Gear VR Carmack said any video above 720p was a waste, but the resolution you perceive when looking into a virtual world, espescially when you have motion tracking is a little bit higher than the raw pixel count might make you think.
Your head is always wobbling a little bit and so what you see is going to shift even as you're staying as still as you can. This is almost like doing multi frame sampling with temporal aliasing, because that diagonal line with jaggies on it, isn't static and locked in place... or you're seeing different pixels of the high resolution texture than you were a split second ago.
A better way to judge, imho, is in how far you can see in the virtual world before things become a mess of pixels. That was the main thing I perceived switching from DK1 to DK2 to Gear VR. The furthest thing away from me that I could still see clearly got further back, from the other side of the room, to two or so rooms away.
You're going to notice the lower resolution much more in something like a racing game where you're looking a ways down the track than you will in a game where everything you interact with is never further away than a couple of steps (like in Job Simulator say). Virtual objects within a certain distance feel tangible and solid, beyond that distance it becomes much more obvious that they're a mess of relatively large pixels.
Not as much as you'd think.
The scale is IMO, the biggest factor to making VR feel unique.
http://www.pcadvisor.co.uk/review/w...reality-headset-hands-on-review-2016-3635648/
This sounds so cool. Does already a video exist of this feature?
plagiarize
Banned
I like that I get notifications on my Gear VR, and sometimes I use voice with my smart watch to reply. The Vive's stuff sounds like a better version of that. Anything that minimizes taking the headset on and off is good.
From what I understand how much you're bothered by the screen door effect is (at this point) mostly a personal preference. Either it'll bother you or it won't. I'm not 100% sure since I've never used VR aside from 90's mall demo but I imagine if I still enjoy low resolution games on low resolution monitors that I'd be fine with a bit of screen door. Your mileage may vary.
Zalusithix
Member
You more or less never have to take it off now until you're done - or have to go to the bathroom. You can deal with the phone. You can see and interact with the desktop. You can see the real world to temporarily interact with. Outside of emergencies I don't see anything necessitating the removal.
I don't think this is always the case.
I think I'm certified to be rather unreasonably fond of high resolutions, and I can still deal even with DK2 resolution in VR. You just have to calibrate your expectations for a new medium compared to a well-established, mature one.
Sleeping Lesson
Member
Totally. The benefits of VR for games far outweigh the loss in resolution.
Yeah, I started to think about that with the cost. My decision is becoming even more clouded, lol.
Cranking up the internal resolution also helps tremendously for the games that support it - for example, it made the screens legible in Elite.
Dreamwriter
Member
That's actually one of the advantages to the Oculus SDK over Valve's OpenVR - the developer can choose to render different "layers" at different resolutions. Like, rendering text at a much higher resolution than the rest, so text is easier to read without killing performance by rendering everything at a higher-res than needed. I'm hoping Elite Dangerous' official Oculus support makes use of that.
The Elite
BOSS
The Rifts were delivered yesterday and we are already in VR waiting for our shipment confirmations.
I heard about this before, what I really don't see is why you need SDK support for something like that, it would seem to be a few dozen lines of code.
Also, "a higher res than needed" is subjective, supersampling surely doesn't hurt non-text either, especially in VR
Sleeping Lesson
Member
I did not know this! That's fantastic. You're right about Elite; super sampling does wonders for text readability.
BrettWeir
Member
I thought we'd see a lot more people (kickstarter backers) posting about notifications today after they started yesterday. It's been dead silent.
welcometooculus.gif
viveks86
Member
I heard about this before, what I really don't see is why you need SDK support for something like that, it would seem to be a few dozen lines of code.
Also, "a higher res than needed" is subjective, supersampling surely doesn't hurt non-text either, especially in VR
Speaking of supersampling, is it possible to downsample games in vr, if we had extra juice left?
Not trivially in the driver, no. You'd need something like a VR-API-specific wrapper to externally inject it.
However, UE4 at least has a built-in supersampling setting which works with VR and which I'd hope all developers would keep accessible.
For Source 2, it seems like Valve will be proactive about that and scale IQ up as far as it can possibly go on your hardware anyway. Which should really be the ultimate goal for all engines.
That is something I want to understand. Adding some controllers and room scale, how is it related ?
For me, it seems to be damage control.
Speaking of APIs, I had a question about an idea I was kicking around - if you were playing a game on virtual cinema on a headset with eye tracking, could that eye tracking status be passed through to apply foveated rendering on the source game via a similar wrapper?
If we're only talking resolution, the the CPU has nothing to do with it.
Depending on the complexity of a game, the current weak link is either the GPU or the HMD.
If by the time the second generation rolls around we already have sufficiently fast and reliable eye tracking, that will decrease GPU requirements substantially and might increase CPU requirements tremendously too.
Foveated rendering is so different from a normal rendering process that I don't believe it could ever be injected with a wrapper in a way that makes sense (i.e. increases performance).
It's related in the sense that they will technically have the pieces to make some form of room scale (not the same size of Vive space due to FOV of cameras), but with Oculus recent downplaying of roomscale and every Touch game shown using a front facing setup who the hell knows what games will actually target. You shouldn't buy a Rift if your primary interest is room scale. Don't buy something for a possible future, but for the now. Seems more likely that CV2 would be room scale. Don't play their game.
And Unity. Valve is releasing a plugin that will do the adaptive quality stuff.
Zalusithix
Member
Ideally this will extend to every VR game within this generation, and then spill out to general PC gaming. Set a target FPS and perhaps give weights to what you as a user prefer to focus on. Then have the engine start adaptive scaling to maintain the framerate. In general gaming give the user a GPU reserve percentage setting so they can control how close to the ceiling it pushes the settings as it isn't as sensitive to a few late frames from outlier cases.
I won't pretend this is easy with all the settings in modern engines that can be tweaked. It'll probably require a pre-run benchmark to get a feel for how the given internal settings affect a user's setup to know how to scale things. It's something that's really overdue though.
I wonder if this will work as well as the Source 2 implementation apparently does, and how much effort it will require from devs.
I do hope it will be smart enough to effectively distinguish frametimes due to GPU load from frametimes due to CPU, and not downgrade my GPU scaling every time Unity does one of its trademark CPU load bubbles
I agree. (I also feel slightly smug given that DSfix could automatically disable/enable some of its postprocessing settings based on a minimum framerate target in 2011
)ashecitism
Member
I wonder if this will work as well as the Source 2 implementation apparently does, and how much effort it will require from devs.
I do hope it will be smart enough to effectively distinguish frametimes due to GPU load from frametimes due to CPU, and not downgrade my GPU scaling every time Unity does one of its trademark CPU load bubbles
Valve already implemented it for the Lab, so it should be ok? However I read some had performance issues with the desktop theater mode, which is also in Unity
The Elite
BOSS
You can go either way. If you're willing to spend $200 more now you get Vive's touch controls right now, which are an important part of the VR experience.
You can also use a controller right now with Rift and pick up their controllers at some point later this year.
If that's what you are going to be using it for with absolute certainty, then likely the Rift.
However, if you haven't used VR much before and are jumping in with these consumer headsets, and you only think you are going to want to use it mostly for seated games (and aren't restricted to that by completely unchangeable e.g. health reasons), then you might want to reconsider.
I wonder if this will work as well as the Source 2 implementation apparently does, and how much effort it will require from devs.
I do hope it will be smart enough to effectively distinguish frametimes due to GPU load from frametimes due to CPU, and not downgrade my GPU scaling every time Unity does one of its trademark CPU load bubbles
If a cpu issue is intruding on your frametimes, the solution to prevent dropping below 90fps would still have to be dialing back the graphics temporarily. Doesn't really matter what the cause is.
Some of valve's Lab stuff is built in Unity so I would assume their plugin works pretty well. A lot of the scaling is in the resolution and AA which shouldn't require much effort from devs. Valve knows that unity is the big VR player right now so it's in their interest to make games built with that run as good as possible.
The majority of non motion control games right now are exclusive to the Rift. It's launching with an Xbox controller and all of the launch software targets it.
Hardware wise not really.
I would ask if you have played a lot of VR games? Motion becomes something people tend to want, but I'm sure not everyone.
This makes no sense. The primary scaling parameters Valve pointed out are AA level and resolution. Changing either of those won't change your CPU time at all.
If anything, you'd require a separate set of CPU-specific and GPU-specifc scaling parameters. (And the former seem far harder to identify and implement at first glance)
I don't think that's really the case. The majority of announced VR games with traditional seated controls are multi-platform. The majority of exclusive ones (or probably all of them) are exclusive to Rift, certainly (though maybe not forever).
Chairmanchuck
Member
The thing is, that some of the games announced for Oculus, besides the Oculus developed ones, are not exclusive that long.
Adam Orth said e.g. Adr1ft is still coming to VR-headsets, they havent announced yet.
I really hope someone will code a wrapper even though I preordered the Rift now so that people with the Vive can play smaller games/projects like the New Retro Arcade on the Vive.
Adam Orth said e.g. Adr1ft is still coming to VR-headsets, they havent announced yet.
I really hope someone will code a wrapper even though I preordered the Rift now so that people with the Vive can play smaller games/projects like the New Retro Arcade on the Vive.
OmarTheHippo
Member
Rift is consensus better for sitting, due to higher number of sitting experience games (from exclusives)
Hardware wise, they're equals at the sitting experience. So the above advantage is only artificial and could change
But buying only for what's for sure and not the "it's possible", Rift is a better fit if you'll only play sitting in a chair
OmarTheHippo
Member
Do you take it as a good sign that the SteamVR test specifically measures "frames CPU bound"? Or is that more artificial and not indicative of progress being able to measure it on-the-fly
Probably phrased it incorrectly but there are a healthy number of launch window exclusive rift titles and a few Oculus published titles.
kind of how g-sync should have been addressed in the first place. Don't adapt the framerate, adapt the content