| Author | Message | ||
     Anonymous Rating: N/A |
In my understanding there are two possible ways of reaching a high FOV. Either making the display of the VR Glasses big or moving a relatively small display very close(1mm) to the eye. Why do all existing VR Glasses use the first(expensive) method? | ||
| Andreas Schulz Rating: N/A |
a) your eyes cannot focus to 1mm, so you will need some optics in between; the closer the display, the stronger the optics (see the contact lenses thread), Stronger optics usually also means more distortion. b) the probably most dominant cost factor in this range would not be size (whether it's .5 in or .1 in ), but resolution - and that would have to be still the same for a smaller display. | ||
| Ray Price Rating: N/A |
I am pondering the same issues atm. Obviously big displays are going to be too heavy and therefore pretty much out of the question. The thing about optics is that they are lenses and almost immediately start to obscure your FOV. Which is why I have start posing the question, how about replacing the optics with contact lenses. I want to try a small static-based experiment using two projectors projecting onto two pieces of curved (half-sphere or less) back projection material (one for each eye) which can then be viewed at close range using magnifying contact lenses. Don't know much about optics, and trying to get some more info on another thread. | ||
| Bricklayer Rating: N/A |
Hey Ray, Cool idea to use a contact lens as part of the HMD optics. I'm working on something similar and may be a little ahead of you on the learning curve(maybe not).  http://pages.zdnet.com/johncraigrochester The eye is a lens and a screen(retina}. The lens has an adjustable focal length from infinity to about 10"(until you get to my age). The eye does this by changing the lens' shape(closer = rounder, farther = flatter). Put simply(and incompletely), light rays diverge from a point and the eye bends the rays back to rejoin(focus) at a point on the retina. The rays from a point very far way are almost parallel and the eye only has to bend them a moderate amount to focus. As the point gets closer the rays fan out and must be bent more. Eyepiece lenses allows you to see stuff closer by pre-bending the light before it hits the eye's lens. The stuff can be a pencil, a small lcd display, or the image transmitted by another lens. There are limits on how much lenses can bend light. This gets complicated and expensive. One good 80 degree FOV eyepiece might cost in the $500 range and be the size of a camera lens. As stated above, the resolution of the display is crucial to a wide FOV image. The eye can discern a resolution of about 1 minute of arc(1/60th degree). What this means is that two different points in space can separated by a distance that forms a 1/60th of a degree angle to the eye. At a distance of 10" you could discern a 0.003" object(fine hair width). A 20" monitor viewed at 24" distance would have to have a resolution of better than 1800x1350(0.18 dot pitch) to look completely smooth. It is very difficult and very expensive to make a 1" display with even a 1024x760 resolution. HMDs with 60+ degree FOV and xVGA resolution are in the $100,000 ball park. If you can get a display 1mm away in focus run(don't walk) to a patent attorney. ;) Later, Craig | ||
| tj Rating: N/A |
I still think that wide FOV (say 60 deg) is well within the realm of being "doable." For example, fisheye lenses are commonplace -- they convert a very wide FOV to a narrow FOV. Wide FOV optics for a HMD would do the same, but in reverse (although of course it's not quite that simple! requiring dual lenses or combo reflective/transmissive/prismatic optics) Sure there are limits on how much refraction lenses can do to light, but a fisheye lens give a good example of what's possible. (See link below to article featuring a Nikon fisheye with 183 degree FOV!) Granted, high quality lenses are expensive, but if you're grinding a lens for a HMD already, why not simply make it wider and shaped to simulate wider FOV? A larger lens uses more glass, so will be more expensive, but only slightly more (not proportionately). The lens has to be wide enough -- span enough physical space in front of the eye -- to occupy more of the eye itself's FOV. However, if the lens is closer to the eye (just out of eyelash reach), it can be physically smaller in size... and doesn't have to be THAT big. Look at some of the eyeglasses that people wear. Many glasses styles these days incorporate rather small lenses, worn relatively close to the eye. This means that a significant portion of peoples' FOV is not corrected. But do they complain about it? No! If we can get to around 60 deg FOV, I think it will be good enough. http://www.ephotozine.com/equipment/tests/testdetail.cfm?test_id=105 | ||
| Ray Price Rating: N/A |
Would this be a speciality lense? How much would this cost to manufacturer? Are there any lenses available now that would give this FOV? | ||
| tj Rating: N/A |
Hi Ray. I'm thinking it'd have to be custom-made, since it's so special-purpose. Basically what you're trying to do is this (please excuse my attempt at ASCII art, ignore the dots (for spacing), and my inability to get a backslash to display): microdisplay: .. ____ ............... / .. ` light rays .. / ...... ` emited: ... / .......... ` ......... / .............. ` lens: .. ==================== ......... ` .............. / refracted . ` .......... / light rays: . ` ...... / ............... ` .. / thru pupil: ..... `/ Of course this diagram sucks. The lens would be wider and the angles of the rays steeper. Also, lenses aren't flat (except for low-quality fresnel), so obviously lens thickness is a big consideration. However, this makes my point. The virtual image is magnified and appears at a greater distance. Taking this same concept and applying it to reflective optics seems to be the way to go. That's what Sony and Olympus have used, in particular, reflecting the image off the inside of a curved prism. That way, you can fold the light path, making the HMD smaller. I'm thinking a person could come up with a parabolically curved piece of glass that would reflect/magnify as well as allowing transmission, perhaps for augmented reality (think of the image you see reflected in your eyeglasses of what's behind you, while you can still see out of them). I don't think the wide FOV is too much to ask. In fact, it's already been done, even in 4:3 formats. The Virtual Research HMDs had pretty wide FOVs, around 60 degrees horizontally. Sony tried it in '99 with a dual-display-per-eye 8:3 format! I still think the reason HMD manufacturers these days are avoiding wide FOV isn't necessarily due to price, but because they're dedicating themselves to PC applications (which are narrow FOV) rather than VR. As far as cost to manufacture, I don't know. Yet consider that every day, "one hour optical" eyeglasses shops are grinding/polishing lenses to a wide range of specifications. While that still might not be fully applicable, it makes me want to surmise that prices for custom optics could be reasonable, for constructing prototypes. I must admit that I can't claim to be an expert on optics. It's not my profession (software graphics is). Although I did have an optics course as part of my BSEE and I study it and experiment with it. Right away, I don't know of any off-the-shelf sources for such specialized (non-spherical)optics as what you might need. My main experimental source is Edmund Scientific, and their stuff is pretty general... but perhaps could suit prototyping. | ||
| Ray Price Rating: N/A |
It frustrates me these guys insist of making products for a market that is not interested. How many people have you seen on a plane checking emails using a Sony Glasstron or cy-Visor? Instead I know tonnes of gamers that would purchase a wide FOV headset for $1,000 or more tomorrow if it became available. Nearly all NVidia cards now support dual-output which can be spanned horizontally, so a dual-display-per-eye with a wide FOV magnifying prism does not sound too technologically challenging for the manufacturers. Anyway, now I am off my soapbox, thanks for the info TJ. I wish I knew enough about optics to prototype something like this. Sounds like most of the components are already available. Maybe I will run the idea past Daeyeng since they seem to be the most active and responsive of companies out there atm trying to fill this void. All the best Ray | ||
| Bricklayer Rating: N/A |
Here's a good link describing the design of an eyepiece: http://www.users.bigpond.com/PJIFL/index.html I'm using an erfle design. My test rig is: Eyepiece- Bought from surplus place. "5 element erfle design. Coated optics. 60 degree field of view, 16mm eye relief, and 35mm clear exit aperture." Webpage(search 60): http://www.surplusshed.com/pages/Eyepieces_Sets.html Objective- Used 50mm, 1.2 Nikon camera lens. The objective produces a 1.2" diameter image at 14mm in front of the eyepiece. The eyepiece is 1.75" long by 1.25" diameter. My next step is to relay this image (like a periscope) a distance of about 2' from objective to eyepiece. I've found stuying optics a mumbling experience. |
