| Author | Message | ||
     syb Rating: N/A |
OK, enuff demanding/whining of 'them' to "Give us a 3D HMD with a 60+ deg FOV for <$1500!". Now that I've got your attention, I'd like to appeal to both your logic and obvious love/interest in this field of stereoscopic technology. I want everyone to ask themselves this: "If everyone wants such a good HMD, with the You may have heard the saying 'If you want something done right, do it yourself'. I don't know why it is so hard to make a wide FOV HMD affordable, and maybe you don't have a clue either but what we both have in common is enough 'spare time' to be reading [and posting] on this webboard. Now consider that it isn't just you and I participating in this webboard. There is perhaps 30+ ppl who actually post here and surely more that just lurk [possibly posting annonymously]. I propose that together - as a community - we can:
Let's brainstorm here in this thread why is it so very difficult to do? Like all complex problems, we need to initially have aggreement on what the problem is then break the problem down into smaller problems. For starters, to help define what we'd be aiming for, is there a definitive list of high FOV HMDs that is any more current and/or detailed than on this site? OK, I know I'm talking in the most abstract [non?]sense possible but if you have any better ideas how you too can [soon] have a high-res 3D HMD with 60+ deg FOV [for <$1500]... please share!!! Perhaps for us the saying should be more like: 'If you want something that no-one else has, design and make it yourself!'  | ||
| Anonymous Rating: N/A |
?????? syb ?????? | ||
| Anonymous Rating: N/A |
Its simple, the main problem is getting to small screens. HMD mfgs buy them from big players that invest millions of dollars in R&D, something a bunch of amateurs cant compete with (unless one gets a brilliant idea he/she can patent). When there is a good small screen for little money, the HMD mfgs would get it instantly a make a good HMD, but there isn't one... | ||
| Ray Price Rating: N/A |
There is however a lot of technology being used in portable LCD and DLP projectors that are becoming increasingly small. Can none of this technology be used to put together a prototype hi-resolution HMD? I think even a headtracker could be a secondary consideration considering there are already companies out there (iSense for example) selling reasonably priced head trackers that will integrate with most current entertainment software. | ||
| Marvio Rating: N/A |
SYB, This has been tried before, and it wasn't succeful, and I'l tell you why. The main reason that HMD's are so expensive is because the tech. required to build them is expensive! It's really quite that simple! The worst being the cost of small screen displays, specially one that wold give you 60deg. fov! Quite simply there's no high quality 60deg. fov LCD that is affordable right now, period. So you see, you can build your own HMD, and it might very well be "better" then comercially produced ones, but it will never be cheaper, at least not at this point... Sorry to busrt your bubble man....  | ||
| tj Rating: N/A |
First of all, I'm all for Syb's zeal. That kind of "why the hell not?" attitude is just the way we need to keep looking at this issue. I also agree with Ray, and I think the answer is in newer technologies like LCoS (not necessarily DLP). Lastly, I agree with Marvio to a certain point, but not as far as bursting anyone's bubble! First item: the microdisplay. This is key, easily the most expensive component in the HMD. I heard a few estimates (I don't know how accurate) that Sony's 1999 line of SVGA (800x600) HMDs contained *transmissive* microdisplays that cost over a thousand US dollars apiece. Apparently their yields were so low that each tiny piece of glass that *did* make it off their line not only cost more than it's weight in diamond, but netted Sony a loss even on their cheaper HMD models. I thought that Liquid Crystal on Silicon would be the answer. Transmissive LCDs involve the liquid being sandwiched between two layers of glass. However, LCoS is a *reflective* technology whereby bottom layer is a silicon chip. Since silicon etching processes are much more refined (easier) than applying films to a tiny silica substrate, yields are higher and costs are much cheaper. [Better example: ever bought a laptop LCD screen that had "stuck" pixels? That screen was, say, 14" diagonally, and still expensive. Yet if they can't even make a screen that big without flaws, imagine trying to do it on a 1/2" diagonal piece of glass, but maintain the same resolution! You'd probably hafta throw away a LOT of scrap for each good one you'd make, seriously driving up your expenses.] The latest generation of HMDs (Daeyang and i-Glasses SVGA, I believe) use LCoS. However, there's been a certain amount of consolidation in the LCoS business, with once-strong players like Colorado MicroDisplay going belly-up... probably, though, due to the economy... and also, I believe, for lack of foresight (like targeting units for web browsing on cell phones, not stereo HMD PC apps). So I don't know, we'll see what happens. I just checked www.personaldisplays.com (formerly a Daeyang/i-ODS joint effort) and it looks like the i-Glasses SVGA are now selling for $650 (probably non 3D)... I'll hafta look at that more. So looking at current technology, your best bet is to start with LCoS. I know CMD was offering evaluation kits in their day. Second item: the optics. This is key also, because it's what interfaces the screen to your eye. Best optics are glass (vs. plastic), but which are more expensive (and heavier). Take it from a guy with a drawer full of torn up Casio mini TV LCD screens, well-frayed copies of "VR Construction Kit"/"Garage VR"/similar mid-'90s books, and lenses ordered from Edmund Scientific... Optics are the hard part. The issue is with how close the HMD is to your eyes. A tiny flaw in the optics or display is magnified many times. A tiny flaw in a TV or computer screen might not be noticed because you sit many times farther away. That, along with the many other ergonomic issues, is yet another reason why HMDs are a challenge to produce. Not impossible, just more challenging. As far as FOV goes, correct optics can give you any FOV you want. The issue there is to decide what you want to do about aspect ratio. A computer screen, and practically every microdisplay, has a 4x3 aspect ratio. You could get your optics to magnify horizontal more than vertical, if you want, but then your image would appear stretched horizontally. If you're displaying a Windows desktop, this probably wouldn't be desirable; things would look too wide. You could magnify both horizontally and vertically at the same rate until you get a 60 degree horizontal FOV, and then your vertical FOV would be pretty big too. However, if you wanted to run, say, Microsoft Word, it might be like having your face 3 inches from your computer's CRT screen -- your eyeballs would have to scan between their full left/right extents just to read a document! That is another reason why I think so many current HMD manufacturers provide such narrow FOVs --> they're taking the easy way out on these issues. Oh yeah, and such magnifications magnify your pixels too, so 800x600 looks like a lot less. SVGA might suffice for now, but I think we'll need higher resolutions first, to do these higher FOVs. Then I'm wondering if we could do some warping of the image, either in optics or simultaneously in software (in the driver). That is, the optics magnification could differ across the FOV, with pixels being more "stretched" out towards the edges, where resolution is not as important. To go with this, and to keep things looking more "normal," the display driver (or ideally, graphics hardware) could warp the image algorithmically to match that done by the optics. That way, we could fill the entire FOV, but still keep the greatest resolution toward the center, where we need it most (and where most people view their Word documents ;-). I believe this issue is the most interesting one, and that some cool ray-casting software could be written to help design the optics... and when I say optics, those can also include reflective optics, which can perform magnification too. (More on this later, like maybe in another post...) Third item: supporting electronics. Actually, not that big of a deal. LCoS manufacturers sell the chips to drive their displays, and you can buy off-the-shelf chips that do practically any video-related function you want to do. Fourth item: patents. Everyone is patent-crazy these days. If you design any kind of product that's not completely "out of the blue," you'll probably be violating someone's patent (even if you didn't derive your design from theirs, which is why the patent system is so miserably flawed). Then you'll need a law degree to go with your engineering degree. Whew! Guess that's it. How's that for a start? P.S. - Oh yeah, and another reason why early HMDs have failed: too many players in too small of a market with too high of prices and too similar of products. Case in point: 1995, Forte vs. Virtual i-O vs. VictorMaxx. ...and it was bad enough before they started suing each other. | ||
| Marvio Rating: N/A |
Tj, I agree with you. But the initial suggestion was to produce this HMD for less then 1600 dollars. I think that's a rather optimistic price.... That's all. If the post was simply to build an HMD with those specs, I would probably have given a similar answer.  | ||
| Steve@IIS Rating: N/A |
You are on the right track TJ. Steve | ||
| syb Rating: N/A |
Thanks ppls for your great input into this thread and I'd like to add some more myself but have been [and will continue to be for some weeks] very busy working hard to ensure I aquire the funds required to purchase a Cy-Visor. I hope that this thread continues to be a place were people can comment on the (im)possibility of a 'better' consumer level HMD. Cheers, --Syb. "No fine work can be done without concentration and self-sacrifice and toil and doubt." --Max Beerbohm | ||
| Bricklayer Rating: N/A |
Hey Syb, I like your idea to form a "community" to brainstorm a HMD with decent specs for an attainable price. I'm game. I've been working on this for a while. I'm trying an idea to optically link the HMD, or Head Coupled Display, to a standard PC monitor via a jointed periscope. While my approach may not work, there must be other ways to get a 60+ deg. FOV image to your eyes without putting the displays on your head. Fiber optics is another approach that could be investigated. I've posted some stuff on my work here: http://pages.zdnet.com/johncraigrochester Later, Craig | ||
| Anonymous Rating: N/A |
How about dual 17" CRTs in a backpack with fiberoptics to the eyes? Its affordable and portable. Combined with a GeForce2 powered laptop you can run any game in stereo. | ||
| Ryan Rating: N/A |
Actually, just a thought here, anonymous makes a good joke, but could there be something there? I know fiber optic cable is pricey, but could moving the display off the head (allowing for larger/cheaper LCDs?) recover enough cost to justify that approach? | ||
| Bricklayer Rating: N/A |
I don't see the need for portability for most apps. To circumvent the expense of micro-displays and the associated hard/software I think linking the HMD eyepieces a PC monitor(s)is the way to go. How can you beat the price of established, mass marketed PC video hard/software? This will require being tethered to heavy monitor(s), but movement can still be achieved with 2D treadmills. The lightness of a Sony PD won't be achieved, but a couple of pounds of lenses hooked to a well balance linkage isn't so bad. Not bad at all, IMO, if 60+ deg FOV and 1600x1200(2.25 minutes) res. is delivered for an attainable price. Head tracking becomes trivial(and extreemly accurate) by using pots on the linkage. I have my design pretty far along. I need some help with the optics design...if anyone has expertice in this area? Regards, Craig | ||
| caused Rating: N/A |
A fiberscope costs around $580 to $800... http://www.spy.th.com/opticsb.html http://www.microcameras.com/fiber_vision/fiber_scope.htm http://www.ozspy.com.au/optical/fiber3500.asp but there is also resolution problems with fiber optics and I don't know what the quality is like for those scopes. On top of that there is probably the need for additional lens in order to get the 60 deg FOV and any focal issues, and head tracking would need to be handled differently, although I like the idea of high precision head tracking, poor head tracking can cause dizzyness. | ||
| StormRunner Rating: N/A |
Actually, A couple of 1024x768 DLP units, such as those used in projectors, could be placed in a backpack or shoulder bag and then have fibre carry the image to the eyes. The optics on the fibre would be the critical pieces. The rest is just off-the-shelf canabalized parts. Might be worth while to check into the current state of medical scopes for a good starting place. | ||
| Bricklayer Rating: N/A |
I always find the US Patent Office web site to be an excellent place to see what the state of the art is for devices. I found a very interesting patent that uses FO and "micro-lenses": United States Patent 6,215,593 Go the the site: http://www.uspto.gov/patft/index.html and search for the patent number. | ||
| Anonymous Rating: N/A |
Why does everyone always "suggest" a solution but never actually attempt it. | ||
| Bricklayer Rating: N/A |
I am attempting it. You want to help? http://pages.zdnet.com/johncraigrochester |
