Example code for VizCave

[stefs]

Hi everybody,
I intend to create dynamic off-axis projections on a table-top 3D display and understand that the vizcave object provides all needed functionality for this. Unfortunately I do not get it to work. In particular when using manual update of the head position, I get unexpected results. I did not find some example code on the web or in this forum. Below is my simple beginner code to explore the vizcave functionality. Any pointers?
thanks
stefs

import viz
import viztracker
import vizcave

viz.go()

c0 = -2,3,0
c1 = 2,3,0
c2 = 2,0,0
c3 = -2,0,0
vp = -1, 1.8, -6

track = viztracker.add()
track.setPosition(vp)

cave = vizcave.Cave()
FrontWall = vizcave.Wall(upperLeft=c0,upperRight=c1,lowerLeft= c3,lowerRight=c2,name='Front Wall' )
cave.addWall(FrontWall, mask =None, window = viz.MainWindow)

def mytimer(num):
global cave
tposition = track.getPosition()
print tposition
cave.update(pos = tposition)

def drawWall():
viz.startlayer(viz.LINE_STRIP)
viz.vertex(c0)
viz.vertex(c1)
viz.vertex(c2)
viz.vertex(c3)
viz.vertex(c0)
viz.vertexcolor(1,1,0.1)
viz.startlayer(viz.LINES)
viz.vertex(c0)
viz.vertex(vp)
viz.vertex(c1)
viz.vertex(vp)
viz.vertex(c2)
viz.vertex(vp)
viz.vertex(c3)
viz.vertex(vp)
wall_frame = viz.endlayer()

viz.add('tut_ground.wrl')
mini = viz.add("mini.osgx")
mini.translate(0,1,-1)
mini.setScale(0.5, 0.5, 0.5)
mini.add(vizact.spin(0,1,0,15))
drawWall()

viz.callback(viz.TIMER_EVENT,mytimer)
viz.starttimer(0,0.01,viz.FOREVER)

[farshizzo]

Are the wall corner position values in the same coordinate frame as your track object?

[stefs]

The printout of tposition shows the coordinates are in the same system.
I also tried to cave.update(pos=vp), with constant vp as defined along with the corner points. Should give me a slightly skewed projection - yet it doesn't.
I am currently using the trial/evaluation version of Vizard - can this explain things?

[farshizzo]

It seems to work fine for me. I created this small test that uses the arrow keys to move the view left/right/up/down. As you move to the edges of the cave wall, the image starts to skew.

import viz
import vizcave

viz.go()

c0 = -2,3,0
c1 = 2,3,0
c2 = 2,0,0
c3 = -2,0,0
vp = [0, 1.8, -2]

cave = vizcave.Cave()
FrontWall = vizcave.Wall(upperLeft=c0,upperRight=c1,lowerLeft= c3,lowerRight=c2,name='Front Wall' )
cave.addWall(FrontWall, mask =None, window = viz.MainWindow)

def UpdateCave():
	if viz.key.isDown(viz.KEY_LEFT):
		vp[0] -= 1*viz.elapsed()
	if viz.key.isDown(viz.KEY_RIGHT):
		vp[0] += 1*viz.elapsed()
	if viz.key.isDown(viz.KEY_UP):
		vp[1] += 1*viz.elapsed()
	if viz.key.isDown(viz.KEY_DOWN):
		vp[1] -= 1*viz.elapsed()
	cave.update(pos = vp)
	viz.MainView.translate(vp)
vizact.ontimer(0,UpdateCave)

viz.add('gallery.ive')

[stefs]

Thank you very much for providing this code, which helped me to resolve the problem. Apparently one has to update the current head position i two different places to make this work. BOTH in the cave object (cave.update()) AND in the main view (viz.MainView.translate()). This appears a little counterintuitive - but once you know it you can live with it...
Thanks again,
stefs

[farshizzo]

Yes, this seems to confuse a lot of people. The reason is that the the cave.update() function only needs to know the physical head position. The virtual head position can be something completely different, in this case it happens to be the same as the physical head position.

[jpriede]

Originally Posted by farshizzo

Yes, this seems to confuse a lot of people. The reason is that the the cave.update() function only needs to know the physical head position. The virtual head position can be something completely different, in this case it happens to be the same as the physical head position.

I am not understanding this. Any CAVE environment should have the virtual camera in the same position of the physical head position (relative to the CAVE frame of reference), right? If the virtual and physical head positions differ (again, relative to the CAVE frame of reference), than the perspective will not work out.... it would be like sitting in the front row at a movie theatre. In order for it to look right, the camera must be where your viewpoint is.

The CAVE can be moved around anywhere within the environment, but that doesn't change the virtual or physical head locations relative to the CAVE frame of reference, right?