Moho Forum

http://www.slowtiger.de/examples/farbkreisel_0.mov

Have a look at this. I made a setup (in 9.5) which needs to rotate a circle in perspective. Version 1 has the circle layer (circles and lines all of same width) put into a group layer. The circle rotates, and the group layer was squashed. All lines retain the same width, which is nice, but the whole thing is only isometric. The loop is 4 sec long and invisible.

Version 2 rotates the circle layer as well but then tilts it 75°. The line width changes with distance, as it should be. But the 'spokes" now change width unexpectedly so the loop becomes visible. And if you look closely you'll see the circles are not perfect as well, they move like a vinyl record which laid in the sun too long (for the younger ones: Vinyl record is like a DVD, but for music and bigger and analogue). I could get away with this in a short scene, but not in a longer one.

Any thoughts about this?

camera settings and scale compensation??

warped vinyl:
as the spinning disk is meant to be horizontal (?), lift the camera in y and tilt it down; and put the disk at 90 degrees around the x axis?
you might try shifting the camera back to (say) z=5 and changing the zoom to maintain visual size.

strangeness in line widths:
it looks to me as though scale compensation is just not working properly - I first thought that adding extra points would help it (I imagined that AS would look at where the point was relative to the camera and adjust width accordingly, but it seems to be more complicated than that ... maybe also looking at the difference in distance-to-camera between adjacent points?? I'm really not at all sure here ...

((what I did: set up a 4-point circle; origin at 0,0; did the camera set-up as above and observed the line width changes - which seemed to follow the points (fat when there's a point near the camera; thinnest when the points are 45 deg rotated) and NOT the curve (which, of course, should show no change, but would not be feasible for AS to do) -- so I expected adding extra points - a 24-sided polygon centred on 0,0 - would sort things, but it didn't!?))

The result is even weirder when done in 8.2: all lines become lighter during rotation!

The angle of the spinning disk plane shouldn't matter at all.

I tested in v11 as well: The older files showed the same behaviour. Another file, completetly created in 11, shows the same phenomenon albeit in a much less visible fashion.

Did some testing as well, in V11, this is strange indeed!

The one on the left as scale compensation turned off, the one on the right, well, You guessed it!