Hi Stefan,

Here are answers to a few of your questions:

On Wed, 10 Oct 2001, Stefan Gruen wrote:

> Therefore I have some more questions:
> 1. Which of these solutions is also cappable of managing stereo video input
> (two cameras)?

ARToolKit can manage stereo camera input. If you have two video capture
cards in your computer you can capture stereo frames easily. However, the
current version of the ARToolKit library on our website does not provide
stereo camera support. We have a version that we are using internally with
stereo support. Hopefully we will be able to release this onto our website
soon.

We have also managed to get several firewire cameras running off a single
firewire port, meaning that we can do stereo tracking on a laptop
computer. On our Dell Inspiron 8100 ARToolKit captures, processes and
displays two 320x240 pixel video frames at about 25 frames/second.

> 2. Which camera type  should one choose for the different solutions.

In general you should choose a camera with low lens distortion and a field
of view close to that of the eye (perhaps 60 degrees). You can buy small
cameras for about $100-200 that are suitable.

> 3. Which solution is easier to calibrate?

ARToolKit camera calibration takes about 15 minutes - there are two
utility programs that are distributed with the software that you can use
to do the calibration.

I can't comment on camera calibration for scan convertor methods. I've
never done this.

> 4. In your perspective: which solution do you prefer?
>

I have never tried video compositing using external hardware so I'm not
sure how that compares to using ARToolKit. However ARToolKit is free, so I
would recommend trying it and seeing what you think..

	Mark

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Mark Billinghurst 		| Human Interface Technology Laboratory
grof@h ..................	| University of Washington, Box 352-142
fax: +1-206-543-5380		| Seattle, WA 98195

==^================================================================
EASY UNSUBSCRIBE click here: http://topica.com/u/?a84Ao5.a9zIQj
Or send an email To: arforum-unsubscribe@t .........
This email was sent to: webmaster@e ............

T O P I C A -- Register now to manage your mail!
http://www.topica.com/partner/tag02/register
==^================================================================

Hi Mark and Yohan

thank you for your quick answer for the two solutions of the problem. These
are exactly the same two ideas of mixing the CG with the signal of the
camera I had, before sending my mail: the first is to mix the signals
outside the computer with the scan converter, the other is to mix the
signals inside the computer ( I did not knew, that the ARToolkit can manage
this). I think, both methods have their advantages and disadvantages. The
first solution does not use the computer for mixing, thus the computer has
more free resources. It also seems to me, that this solution is good for
high frame rate and high resolution of the video signal.
The second solution has the advantage, that one can use the video signal to
detect real objects in the video stream, such that the computer graphics can
be covered by the real objects, if the cg is behind the real object in the
real space.

Therefore I have some more questions:
1. Which of these solutions is also cappable of managing stereo video input
(two cameras)?
2. Which camera type  should one choose for the different solutions.
3. Which solution is easier to calibrate?
4. In your perspective: which solution do you prefer?


Thank you a lot
  STefan


-----Ursprungliche Nachricht-----
Von: Mark Billinghurst [mailto:grof@h ..................]
Gesendet: Mittwoch, 10. Oktober 2001 02:02
An: arforum@t .........
Betreff: Re: [ARFORUM] Mixing Video an CG in realitme



Hi Stefan,

We have that working with the current version of the ARToolKit software we
are developing here at the HIT Lab. We have a laptop with firewire video
input that can capture video frames at 30fps. Then we render 3D computer
graphics on the ARToolKit tracking markers.

We have not released this into the public domain yet, but should be able
to do so by the end of October. It'll be put onto the ARToolKit website -
http://www.hitl.washington.edu/artoolkit/

With this software there is no programming needed appart from the 3D
graphics application code - ARToolKit takes care of the tracking and video
compositing for you.

The laptop we are using has a GeForce2 Go chip in it which is a pretty low
end GeForce card (17 million polygons/sec), so practically and GeForce
card would do the trick..

Please let me know if you have any questions about this..

	Mark

On Tue, 9 Oct 2001, Stefan Gruen wrote:

> Hello everybody,
>
> for building a video-see through AR-system I have the following problem.
> Suppose, that I want to  overlay the video stream of a fire wire camera in
> realtime with the computer graphics which my computer produces. One idea
> would be to mix video stream and the computer graphics in the computer and
> connect the HMD to the graphics card. What kind of graphic cards do you
> recommend for mixing (sorry: no SGI). And much more important: is there
any
> software with which I can mix the video stream and the computer graphics
in
> realtime? Or do I have to program it for my self?

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++
Mark Billinghurst               | Human Interface Technology Laboratory
grof@h ..................        | University of Washington, Box 352-142
fax: +1-206-543-5380            | Seattle, WA 98195

==^================================================================
EASY UNSUBSCRIBE click here: http://topica.com/u/?a84Ao5.a9zIQj
Or send an email To: arforum-unsubscribe@t .........
This email was sent to: webmaster@e ............

T O P I C A -- Register now to manage your mail!
http://www.topica.com/partner/tag02/register
==^================================================================