It's not an either or situation, technology marches on and patents finally expire ... Oculus Rift can get a device for cheap which uses the fused results of accelerometers, gyros and magnetic field detectors. As I said, all this problem needs to be solved is a little volume.
MS Research: IllumiRoom *spin*
- Thread starter MarkoIt
- Start date
(((interference)))
Veteran
Illumiroom couldn't be a feature supported by pretty much every Durango game (besides its basic operating modes, starfield, grid etc) since it requires the game to render more pixels than just the 1080p/720p needed for the main display - which is something devs will have to budget for (like 3D rendering)
Right?
Right?
Last edited by a moderator:
Yes IMO it would require specific support by a title, and I'm also not sure how cheap they can make it which IMO leads to low adoptions rates.
I'm a huge fan of fill peripheral vision, it's way more impressive than you expect, I have a 3D surround setup on my PC, I IMO the surround part offers more immediate tangible benefits than the 3D part.
One of the reasons I think the Occulus rift is getting good press is it's been a long time since people have seen a wide FOV HMD. All HMD's used to be wide FOV, but over the last few years the push to smaller form factors has prevented that.
Silent_Buddha
Legend
It's basically an extension of the research that Microsoft has done in Foveated rendering (http://research.microsoft.com/apps/pubs/default.aspx?id=176610 ), only without the requirement to track where the player is focusing their vision, thus removing the most expensive part of implementation.
In the case of Illumiroom, it is assumed that anything on screen is within or potentially within the narrow field of view of maximum image resolution for the eye. Anything outside of the actual TV screen is considered to be in the area of rapidly declining visual acuity.
Hence, there is very little detail required for anything projected outside the screen as it is assumed that when playing a game it will always be outside of a person's ability to distinguish fine detail.
In other words, instead of using foveated rendering to reduce 3D rendering requirements (as that particular research was focused on), they are using the things learned from it to extend the visual "landscape" by taking advantage of the fact that it doesn't require much detail.
Perhaps, perhaps not. If a system can be devised that analysis patterns and can extrapolate a low detail approximation of an extension of the scene that it may be possible to do it for most games.
But for best effect you likely would want to have in game support for it, but due to the low detail required, I can't imagine it would require much effort or much resources.
Regards,
SB
No it's not. and the problems can be reduced over time. How much? not sure. I suspect (and don't really care if I'm right or wrong, I'll take whichever approach works best) that the projection methods will outperform HMDs as a quality option. I also suspect HMDs will be tech of choice for those on a budget or non-specialised play spaces.
Who knows, and since I've waited 25 years for this area to progress AT ALL, I'm happy to see some movement.
Maybe also the requirement for resolution? Sony's HMZ is 720p, and that's in each eye offering equivalent to a huge screen. 720p stretched over a huge screen isn't great resolution. Oculus rift increases the FOV but reduces the resolution to half, so almost an SD TV occupying an enormous FOV. Quality must suffer considerably.
Sure, but we're talking consumer living room here, and not custom-built-imaging-room.
But they use to be incredibly expensive. Give me 5 Pico projectors with a webcam each and ImmersaView tech (http://www.immersaview.com/) and you've got something that used to cost 100K now for an order of magnitude less. HMD's aren't showing the same level of cost reduction by a long long shot.
But I'd need five projectors placed in the centre of the room with no chairs, and I'd need to stand on top of or underneath the projectors otherwise my body will shadow the scene. It's totally infeasible to have a multi projector setup as a consumer level experience. It could be released as a niche product but it'll never be mainstream. Full 360 degree viewing in a headset only needs a headset. Everything else can be solved with technology, whereas the full-room projection can't because it can't solve the issue of furniture and cramped spaces.
As long as you are perfectly positioned in the calibrated reference position, perspective distortions can be address, but removing all colour and light influences is probably impossible. You could correct to some degree but it won't be perfect. A HMD solves that in a far more convenient and robust package. It'll also support multiple users far better - you multi-projector setup would be casting multiple people shadows. The price is still high as the tech is cutting edge (tiny OLED HD screens in the case of Sony's HMZ), but that'll get there. So from TV to environmental imaging to full 360 degree immersion, Illumiroom is an interim idea. It's not good enough to extend to full 360 degree immersion in the consumer space. MS haven't even tried to position it as such, experimenting with just one projector and camera positioned above and behind the player. Only in the patent with the magical hemispherical projector was the idea of filling a room hinted at, and patents can ignore the realities of science when trying to secure an idea
That was my original point. HMD for the low end, Projection Space for the high end. In projector based home theatres, the projectors are placed out of shadowing range. In M&S situations it's the same. Webcam calibration systems account for the perspective distortion due to elevation quite nicely and cheaply.
Understand that we are talking about a price in the the thousands range here. Not 10k's or 100k's.
HMD's of any quality are only going to be an order of magnitude less in the foreseeable future. These systems are available now and will continue to decrease in cost.
I'm not saying it will be for everyone, but the price rate comparison is heavily in the projection camp's favour.
No, but arguably HMD displays are even further from perfect.
Based on current and foreseeable tech, highly dependent on location for convenience and not even close on robustness.
Depends on positioning of the projectors. I'm not suggesting this is a simple installation, HMDs have this as a big advantage, no doubt.
I disagree. I believe HMD is the interim idea.
No possible not. Both options at this stage are only baby steps forward.
MS haven't even tried to position it as such, experimenting with just one projector and camera positioned above and behind the player. Only in the patent with the magical hemispherical projector was the idea of filling a room hinted at, and patents can ignore the realities of science when trying to secure an idea
In the current iteration, true. There are however actually implemented versions way above microsoft's. can't say the same for HMDs.
For 360 degree projection it is a matter of cost reduction and simplification of installation. For HMDs there are a heap of technical challenges:
Visual resolution:
Multiple projection has the advantage here. Projectors are decreasing in cost at a higher rate than small custom LCD's are increasing in resolution.
Tracking:
Tracking adds latency. Even optimised down to zero it can only approach the projection system in this regard.
From a quality perspective, projection rooms were always the preferred solution 30 years ago, 20 years ago, and 10 years ago and today. From a cost perspective The HMD was the poor mans option. That restraint is becoming less by the day. 20 years ago the price difference was 3 orders of magnitude. Today it is only 1.
1) How do you achieve 3D in your projected space?
2) What is the relevance to Microsoft's consoles, or are you talking specifically about replacing high-end tech (in which case what is Illumiroom doing that a blank room with four projectors can't already do?) and not a living room experience? The debate about HMD versus virtual spaces in corporate modelling etc. is a very different one with very different requirements. Illumiroom is talking specifically about added immersions in a cluttered living space, and there's no obvious fix to that, whereas HMD, even if not ideal for commercial apps (as long as you don't want 3D?), ideally fit the consumer space.
As for projector prices coming down, you might want to state listing prices at resolutions and brightnesses. It may be laser tech at 1080p is pretty cheap these days at reasonable brightness, although every 1080p projector I've ever checked out is very expensive, especially short-throw projectors. Considering the HMZ T2 is £1000 with dual 720p OLEDs, and the Sony NEX FDA-EV1S viewfinder is XGA for £210, getting a £500 1080p headset (ignoring completely alternative imagine techs for headsets) within a few years seems credible to me. Multiple 1080p projectors for £500 in the same time period I find doubtful, but there may be some super breakthrough laser tech.
2) What is the relevance to Microsoft's consoles, or are you talking specifically about replacing high-end tech (in which case what is Illumiroom doing that a blank room with four projectors can't already do?) and not a living room experience? The debate about HMD versus virtual spaces in corporate modelling etc. is a very different one with very different requirements. Illumiroom is talking specifically about added immersions in a cluttered living space, and there's no obvious fix to that, whereas HMD, even if not ideal for commercial apps (as long as you don't want 3D?), ideally fit the consumer space.
As for projector prices coming down, you might want to state listing prices at resolutions and brightnesses. It may be laser tech at 1080p is pretty cheap these days at reasonable brightness, although every 1080p projector I've ever checked out is very expensive, especially short-throw projectors. Considering the HMZ T2 is £1000 with dual 720p OLEDs, and the Sony NEX FDA-EV1S viewfinder is XGA for £210, getting a £500 1080p headset (ignoring completely alternative imagine techs for headsets) within a few years seems credible to me. Multiple 1080p projectors for £500 in the same time period I find doubtful, but there may be some super breakthrough laser tech.
LCD shutter lenses if you required. Most don't.
Consumer tech from a GPU / Game Engine / Computational perspective is starting to flow back towards commercial / defence sectors.
Gaming tech has done a great job in the past of taking high end M&S tech and producing it at a consumer level, e.g. 3dfx vs SGI: that was still IMHO the greatest breakthrough of all in the early 90's. It still amazes me when retelling my experience with developing on SGI Onyx RE2s to using the original Voodoo cards. Yeah we lost color precision, but the speed was just unbelievable for the cost difference. Think 2 orders of magnitude overnight.
20 years ago, You could get a HMD with 250ms latency at 320x240 resolution and 0.5degree accuracy for about $5000.00.
A Cave environment with 4 projectors of 640x480 or less in a 3m x 3m space would cost around 100,000K.
Today, You could create the same cave environment with 1080P projectors for around 2-3K.
Occulus Rift provides (supposedly) 1080P per eye visuals for $300.00
Accuracy and latency are not specified so no idea on their improvement.
The interest in IllumiRoom is that it may provide 90% of a caves capability without requiring multiple monitors and a sparse environment.
The interesting thing for me was the obvious awareness that peripheral vision is important but only at a lower fidelity.
I don't think ImmersaView will license you that stuff cheap enough for it to make sense to go cheap on projectors, unless you mean that it can in theory be cheap if you had the tech ... but then I could make the argument that the theoretical component costs of HMDs have seen similar drops in prices. The optics and the display/SLM are the most expensive components ... the latter obviously shares the cost reductions in CAVE's, the former have also seen huge reductions because of things like plastic aspheres. As far as realising the full potential of those savings goes CAVEs have the advantage because it's all COTS except for the software and there are less patents ... but still.
PS. with your price examples HMDs didn't even do that much worse ... a factor 2x, meh ... who cares.
french toast
Veteran
I think illumiroom will prove to be ground breaking I really do, as long as it priced like kinect and works like advertised, it ill completely change gaming in a way that no gpu upgrade could.
The key to this is its promise of simplicity to set up, room arrangement tolerance and hopefully accessible pricing.
The key to this is its promise of simplicity to set up, room arrangement tolerance and hopefully accessible pricing.
Maybe the Display Planes hardware in Durango could help minimize the (performance) impact and ease the implementation? (i.e. a very low res "illumiroom" layer separate from the main game layer, rather than the console needing to render at a native res that it isn't equipped to handle...)
Last edited by a moderator:
Wouldn't make a lot of difference, as it's been demonstrated you have to render the outside of the view at lower resolution, that in and of itself might not be a big deal performance wise, but to do it effectively you'd probably end up with a very low LOD that you might otherwise forgoe, and possibly simpler shaders specifically for the peripheral view.
I don't think the support/rendering cost is dramatic, but I think games would need to support it.
There are high end HMD's that just extend the edge colors into the periphery, and that actually works fairly well, something like that might be doable at the OS level, but there are other considerations like how the final frame is composited in that case.
Either way you'd need a second video out to driver the projector.
So you'll need to factor in higher framerate 3D projectors. For gamers and immersion, 3D is going a be a huge plus.
Oculus Rift provides only 640x800 resolution per eye, and a very wide FOV. Sony's HMZ is a far better quality. but the important point here is that an HMD provides stereoscopy without any faults (no crosstalk) and better clarity than projections (better blacks).
That makes no sense to me. You're talking about cave as a commercial venture, for a professional userbase. Why will they want to use four-wall projection in a cluttered room instead of having a room for the job? That can be done without any need for Illumiroom. Illumiroom is attempting to filter out the imposition of furniture, or use it to artistic effect, and in that environment four wall projection makes no sense.
Which isn't what cave projections are about. They splat the same resolution everywhere and waste all that rendering behind you and to the side of you. An HMD only concerns itself with where you are looking. I even expect eye-tracking and foveal rendering to be easy to incorporate into an HMD, which'll be much harder to manage in a multi-projector setup.
So in summary, in the consumer space HMDs are clearly the way forward for full 360 VR immersion. In the commercial space, there'll be different pros and cons for different systems, but Illumiroom isn't going to help really. It could be used for some niche purposes, but it provides nothing for 360 degree imaging where you'd clearly set aside a blank room for that. Illumiroom is something else that may have it's time in the sun but is destined to be phased out, unless people decide the fashion with video interacting with furniture is the future.
Similar threads
- Poll
