Some general questions

[rophat]

Hi,

We are a group of researchers who work on product design. We found Vizard in one of those tech fairs and liked what we saw from the demos. We are seriously considering buying it but not sure whether it is the solution that we are looking for. We'll appreciate that if you can let us know more about its capabilities. Our project has some specific requirements as follows:

1. Can we export the Vizard application as a Web-based program like Flash or some other similar output with special plug-ins?

2. Following the quesiton above, can AR stuff be incorporated and run on the internet?

3. We are also hoping to acheive a high level of visual realism so would like to know the Vizard's rendering capability of virual objects in AR-related applications.

4. For the tracking in AR, how small the marker can be? And this might be a dumb quesiton: Is it possible to use features found on, for example, human faces, instead of markers? Or we have to ask the user to stick a small pad on his/her forehead anyway?

I know it's probably too much to ask for Web-based AR applications, considering the varieties of task that need to be run in real time. So, we also welcome any suggestions of Vizard's alternative solutions to this problem.

Many thanks!

[farshizzo]

1,2) Vizard does not support publishing a script into a web application. You can only publish it into an executable.

3) Vizard has built-in support for the OpenGL shading language, which allows you to write your own shaders and rendering effects. We also have an exporter for 3dsmax which exports models into Vizards native file format. Having a good graphic artist that knows how to use 3dsmax is usually the best way to improve the realism of your virtual world.

4) The size of the actual marker is not very significant, it's how large the marker appears on the camera image. Having a small marker that is very near the camera lens will work better than a large marker that is very far from the camera lens. The ARToolkit plugin only supports marker based tracking. There are other tools out there for performing markerless face tracking, you should look into those if you don't want to use markers.

[rophat]

Thanks for the prompt reply! It helps a lot!

[rophat]

Hi,

Back to the marker/tracking question:

I know it's probably OK to render the virutal object in another location rather than stick it on the marker (i.e. apply some tranformations). However, if the marker is used as a tracking reference, i mean for matching the 2D projection of the marker and the virutal object, how do you then work out the projection angle of the virtual object when it is placed at another location?

For example, if we place the marker on the wall, how do you make the virtual object stand vertically on the table rather then sticking horizontally on the marker on the wall? Of course, this is a simple case because of they are roughly right-angled, we will appreciate a more general approach.

We will definitely wriite Vizard into our budget plan if this can be solved.

Many thanks!

[farshizzo]

You can use the ARCamera.setGlobalMarker() command to use a marker as a global reference frame. For example, if you have a marker that is 300mm wide and hanging on the wall 1 meter away and 0.5 meters above the table, then you would use the following code:import viz
viz.go()

#Add ARToolkit extension
ar = viz.add('artoolkit.dle')

#Create camera from texture
camera = ar.addWebCamera()

#Create marker from pattern file
marker = camera.addMarker('ar/patt.worldviz',width=300)

#Generate matrix of marker in global reference frame
m = viz.Matrix()
m.setPosition(0,0.5,1)
m.setEuler(0,-90,0)

#Use marker as global reference frame for camera
camera.setGlobalMarker(marker,m)

#Add object that should appear on table
viz.add('logo.ive')If you are not using a camera parameter file that has been calibrated for your camera then the offset positions won't exactly match the physical positions. So you might need to tweak them for the object to appear at the correct location.