Tails and Defects

[buffo]

Too be honest, if you've tuned your scanners on the ILDA pattern at a big scan angle, there's a good chance you're not actually driving them hard enough at 8 degrees.

Actually, if he tuned them at a wide angle, he could still be driving them just as hard at 8 degrees. "Speed" is relative to the difference between actual and desired position only up to the scanners maximum performance.

the impact of a significant change in the scanned angle as in 8 to 30 degrees I would think needs to be a part of the original tuning sequence. The substantial increases in velocity probably will change the balance between gain and damping.

Once the scanners are ballistic, it doesn't matter how much the difference between desired and actual position really is. The scanners are only going to move so fast, no matter how wide you scan. Tuning with the pattern set too large just makes it more difficult (or even impossible) to get them tuned properly.

Remember: the center circle is already designed to be well beyond the scanners ability to reach, even at very small scan angles. (You can tune your scanners just fine even at 3 degrees. You don't have to be at 8 degrees.) Making it twice as large doesn't change that fact, but it might make other parts of the pattern beyond the scanners ability to reach as well, which would introduce more distortion that you would then be trying to correct with tuning. (And failing at)

The velocity change between a small jump and a large jump are already accounted for in the ILDA test pattern. When scanners are properly tuned, there is no distortion at all on a small jump, and a very predictable distortion (the rounding off of the 12 corners of the dodecahedron into a circle) of large jumps. This predictable behavior on a large jump is what is used when artists "pull points" in a frame to make fast circles while keeping the total point count of a frame low to avoid flicker. An example of this is the mermaid sequence that comes with Pangolin. Her eyes consist of something like 4 or 5 points each, but they are stretched far enough apart that the scanners round them off to a pair of small ovals that look just like eyes.

And what, if as I do, the scan is made non-square and the rectangular pattern is what needs to be tuned?

The maximum angle will dictate the limit of the wide axis of your rectangle. There's nothing wrong with scanning below this point (as the Y axis in your setup would.) If you want to push your X scanner beyond this point to get a wider pattern, you will get distortion in that axis. If it's just a small amount, maybe you can live with it. But you will probably end up with tails on some frames, at least on the X axis.

Adam