[simpits-tech] Falcon views 360? (Long)
Marv De Beque
mdebeque at woh.rr.com
Fri Oct 17 19:04:51 PDT 2003
Depends on the surface shape you are projecting on and the angle.
On 10/17/03 12:23 PM, "Ido Dekkers" <idekkers at clicksoftware.com> wrote:
> don't you need a view designed for a fisheye lens so it won't look distorted ?
>
> Ido
> -----simpits-tech-bounces at simpits.org wrote: -----
>
> To: Simulator Cockpit tech list <simpits-tech at simpits.org>
> From: Marv De Beque <mdebeque at woh.rr.com>
> Sent by: simpits-tech-bounces at simpits.org
> Date: 10/17/2003 01:27AM
> Subject: Re: [simpits-tech] Falcon views 360? (Long)
>
> 1: Version of Falcon (or other flight sim software) to support multiple
> monitors (either directly or through some sort of network). The narrow FOV
> will need a way to be switched to any desired panned position rapidly.
>
> 2: Displayable dome (Can be partial or full depending on your needs).
> Probably 15 feet in diameter is the minimum size.
>
> 3: Two projectors. You will need some optics to narrow the field of view for
> one and a fish eye optic lens for the wide view. Edmund Scientific, here we
> come!
>
> 4: 2 Optical grade flat mirrors. One is stationary, the second would be
> mounted to a two-axis gimbal controlled by high-speed servos. Basically, the
> pair form a periscope so that the narrow view projector image can be aimed at
> the pilots field of view. This can be done without the mirrors, but that
> would require the projector be mounted on the gimbal. The problem with that
> approach is mass. Rapid movements would require overcoming larger amounts of
> inertia and can result in over shoot or ringing as the system stops. The
> ringing could be computer dampened, but the mechanical stress would most
> likely be detrimental to the projector and the bulb.
>
> 5: Head Tracking hardware and a PC to control the servo actuated mirror.
> Software to drive the servo system.
>
> Obviously, the dome would require a lot of space, but you could reduce the
> size by limiting the view area to whatever works well.
>
> A slight modification of the above plan would be to reduce the wide angle
> projector to something closer to 180 degrees or less. The second projector
> would also be slaved to the narrow view projector’s image so it would move in
> tandem.
>
> A low tech solution would be to use one projector and project the image where
> the pilot is gazing. You loose peripheral images, but you save buying two
> projectors. This may be a good plan for getting started. Getting the head
> tracking and image projection working would be the lion’s share of the
> technical hurdles out of the way. Adding a second projector would be
> relatively easy.
>
> However, all of those ideas require a large room for a dome of some sort. I
> don’t have a theater in my home!
>
> Another thought would be to project the image from above onto a cockpit canopy
> that is translucent. This would reduce the amount of room size required and
> reduce the amount of swept area for the second projector. One problem with
> this approach would be a loss in depth perception. Your focal distance would
> be very close. I don’t know how that would be interpreted, but I would not
> think that it would be too bad. Still, I’ll bet that it will look like
> someone has projected an image onto your canopy, rather than an out the
> window view. Might be an interesting experiment.
>
> Projection onto a canopy has another interesting challenge. The image would
> need to be projected from at least two positions and the focal length would
> need to be a function of angular position. When you project onto a concaved
> curved surface from the inside and the radius of the curved surface coincides
> with the optic lens, the focal length is the same at every point on the
> sphere’s surface. Think of a very large ping pong ball with a projector at
> the very center. Now, if you project from the outside of that ball onto a
> convex surface, the focal length changes over the surface of the sphere you
> project onto.
>
> I think this can be compensated for optically. You will need to do this
> anyway since the image will also be grossly distorted. Imagine trying to
> project a uniform grid (like the Earth’s latitude and longitude lines) from
> inside a ball. If it was projected from the inside radius, as I cited in the
> first example, the grid would look perfectly uniform. When projected from the
> outside onto a sphere the grid would be highly distorted.
>
> Since a canopy is not a perfect sphere, but an irregular shaped object, the
> distortion would be even more complex. Again, if you can model the surface to
> be projected onto mathematically, then you could, in theory, grind a special
> lens or mirror to compensate for this. Well, to a point.
>
> Additionally, you need to also add either additional projectors to cover the
> outside surface or beam splitters to split off the projector’s output and
> project from multiple angles onto the canopy. Why do we need more projectors?
> Well the best coverage you can get for illumination of a sphere from the
> outside is no more than half a sphere. Even at that, the edges have a high
> degree of distortion. Since a full cockpit view is technically larger than
> 1/2 a sphere, multiple sources are the only answer.
>
> A compromise would be to construct a canopy that is faceted and project
> separate images onto each flat facet. Maybe 4 to 6 projectors could cover a
> very large field of view. At some point with enough facets, the projector
> resolution for a given area or facet becomes good enough that the head
> tracking is no longer needed. Flat surfaces provide a wonderful way to
> relieve the problem of complex optics, but you still have seams between the
> facets.
>
> Another problem with projecting onto a surface that is in close proximity to
> the pilot is the parallax effect. When you view some object out the window
> and the observer moves slightly side to side, the object moves relative to the
> window. You loose this with an image that is close to your field of view.
> Head tracking can be used to apply an image shift to compensate for the
> movement and would give back that illusion of the object’s intended distance.
> I think that would greatly reduce the painted on my canopy feeling.
>
> Personally, I think that there may be some advantages to this system. While
> more projection systems are required, the space needed to do it would be much
> less than a full or even a partial dome! You also could eliminate some of the
> complexity of having head tracking.
>
> Further down the road would be the use of VR goggles. As we mentioned before,
> the current resolution is poor and you don’t have the ability to see real
> objects such as your real cockpit. However, that problem is not really a
> technological barrier that hasn’t already been solved. There has been work
> already for specialized application (such as surgery) where both real world
> and virtual images are mixed. This can be done via small cameras on the
> glasses for the real world view or a complex shuttering mechanism that allows
> external light to pass through selectively. Both systems would require a
> computer to have a programmed field of view that is predetermined (such as the
> cockpit rails where the canopy intersects). Either pattern recognition or
> head tracking system would determine where the virtual and real world
> intersect.
>
> Obviously, there are cost issues with this since the technology is new and
> high resolution images for VR goggles are very expensive. If the market for
> this type of technology expands, the cost may someday drop to a point where we
> could afford it.
>
> However, today we are seeing a significant drop in the price of projectors and
> a new family of projectors that use DLP. DLP essentially is a high density
> array of miniature mirrors that are electronically steered by a silicon chip.
> Rather than pass light through a LCD, light is reflected off of the DLP
> mirrors which are precisely positioned electronically to produce an image.
> The result is a much, much improved image brightness. And, no smoke is
> required. :-)
>
> Marv
>
>
> On 10/14/03 9:56 PM, "JL" <Johnlimd at bellsouth.net> wrote:
>
>>> What I proposed is not a perfect solution (you would have to fly a real
>>> jet), but think of the possibilities. First, you only need two views. You
>>> could easily build this system today for under $3,000. Probably under
>>> $2,000. The projectors are 80% to 90% the total cost. What are the
>>> alternatives? Placing a half dozen monitors around your desk? What about
>>> all those seems? How much area can 6 monitors really cover? VR goggles?
>>> Good field of view, but lousy resolution. You get 60 degrees or so of FOV
>>> at 1024 by 768 pixels on a really good pair. Most goggles have mush less
>>> resolution and still cost some bucks. Of course if you have $100,000 you
>>> buy what the military uses. Add to that you can’t see your cockpit (which
>>> you just spent years putting all that detail into). Would you rather flip a
>>> virtual switch or a real one you can feel?
>>>
>>> Marv
>>>
>>>
>>>
>>> Alright Marv!
>>> I dare say that many of us would love to see your idea work for $2000.
>>> I'm in, even at a cost of $3000. Who else is in?
>>>
>>> Design and implementation plan? -- All smoke and mirrors??
>>> Doesn't Falcon 5 need to be written to support the low rez 360 view?
>>>
>>> H ere's another idea.... With projectors getting so lightweight (< 3 lb)
>>> and relatively less expensive, how about mounting one on your helmet so it
>>> paints hi rez where you're looking. Gives new meaning to "Hot headed pilot"
>>>
>>> B ut seriously, can you enumerate the steps that need to happen to make
>>> this reality? (not that you don't have enough to do... ;- )
>>>
>>> John
>>
>>
>> _______________________________________________
>> Simpits-tech mailing list
>> Simpits-tech at simpits.org
>> http://www.simpits.org/mailman/listinfo/simpits-tech
>> To unsubscribe, please see the instructions at the bottom of the above page..
>> Thanks!
>>
>
> _______________________________________________Simpits-tech mailing
> listSimpits-tech at simpits.orghttp://www.simpits.org/mailman/listinfo/simpits-te
> chTo unsubscribe, please see the instructions at the bottom of the above page.
> Thanks!
>
>
> _______________________________________________
> Simpits-tech mailing list
> Simpits-tech at simpits.org
> http://www.simpits.org/mailman/listinfo/simpits-tech
> To unsubscribe, please see the instructions at the bottom of the above page.
> Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.simpits.org/pipermail/simpits-tech/attachments/20031017/f9a8fb52/attachment-0001.html
More information about the Simpits-tech
mailing list