Low cost Virtual Reality?

I originally posted to the "stereo3D Problem" topic, but decided to
move to a new thread now that I have some new info;

I've come across www.cymouse.com, which sells a 6dof head tracker
for under $100. (though for some reason one of the angles is
supressed) It has 1mm positional accuracy, and 1degree angle accuracy.
This headtracker is the final element, and should make virtual reality
very cheap to implement. However, their product is marketed as a mouse
alternative, not a VR product.

I've E-mailed the company, asking if they have or know of any VR or
CAD software which is compatable with their drivers. Here is his
reply,

---------------
The data output of Cymouse is rather basic. At the end, you have five
coordinates representing the position of the mobile sensor. The
coordinates are floating point precise, however, the precision can be
scaled down if not needed. I guess it would depend on how the
SolidWorks interface expects the incoming information. I know that
the current product does not attempt to interface with any other
hardware on the system, but that doesn't mean such a thing could not
be done.

-Dave
---------------

I am really puzzled that no company has jumped on this. It probably
wouldn't take much code to implement.

Joe Dunfee

 

The key to a tracker is latency - how quickly can it detect change in
position or velocity or acceleration, convert that to numbers and feed it
sensibly to the CPU. Any delay over 10ms is detectable by a decent
responsive human, and the lag can cause dizziness, loss of spatial
awareness, and so on (the puke factor). ie. You move your head, the image
takes a while to catch up. With S/W complex graphics, the delay would be
even longer. Stick to flat screens. HMDs don't cut it in CAD.

an old VR thespian

bp

 

[Your reply too over a day on Google. I tried posting this not via
another server, but ran into problems, so it may appear twice)

In my case, I am actually thinking about fishtank VR. I.e. just
using a desktop 19" monitor. Is this type of system as suceptible to
delay issues?

I am guessing that the result will be that the object seems to be
less fixed in place... perhaps "sloshing" around a bit in its
"fishtank"?

Ultimately, my goal is not for myself. When another person looks
over my shoulder at a complex model, all they see just a bunch of
colored lines criss-crossing the screen. But because I already know
what I am looking at, it is possible for me to see the 3d object.

Of course, the Shaded images, and just regular orbiting is a major
help. But there are still many times I will go up to the storage area
to see something created before I came to the company. The reason I
must do this, is that the real object is MUCH easier to understand
than a CAD 3d model.

I am speculating that stereo vision is not necessarily the most
important aspect of this. Just being able to bob your head around to
look at something seems to be an important part of seeing it. Is this
true in your experience?

Finally, is a "Fishtank VR" system as suceptible to delay issues and
HMD's?

Joe

 

Just an update. I tried the Cam2Pan software
http://www.mousevision.com (a free 30 day trial, and you use your own
camera).

The results were interesting. I used the rotate view command, with
the settings for its sensitivity way down. I also set the speed on
the Cam2Pan software very low (values like 3 to 6)

I can't seem to get the view angle to really coordinate with my
viewpoint, if the object were really inside the monitor. I am still
playing around with it. I haven't tried it with stereo glasses yet.

Just to give you something to play with over the weekend.

Joe D