I've seen this in operation and am acquainted with the researchers. I may be able to attempt to answer some questions about it.
I would guess that the relatively long wavelength of ultrasound could potentially cause resolution issues.
I doubt ultrasound can provide the same fidelity as a highly-tuned laser high-frequency laser beam when it comes to keeping something in one place.
(Thanks @iliis for posting about this research)
That page says the femtosecond plasma display is pretty safe. This display was - and I believe still is, though I could be wrong - using a UV laser for confinement, which is sorta scary.
Can you levitate multiple particles at the same time? How fast are they moving? How sensitivie is it to disturbances in the air?
Yes, but their device currently is only capable of levitating a single particle. With the current approach, the device would need one laser for confining each particle. The lasers used for RGB could potentially be multiplexed between particles, but this is less likely to work for the confinement due to the instability of the trap.
> How fast are they moving?
Pretty slow. All the "big" images (Leia, grad student, etc) are long-exposure. You can see the particle moving in real-time at about 0:50. In the videos, you can see the device can almost do real-time persistence of vision for volumes substantially smaller than a fingertip.
> How sensitive is it to disturbances in the air?
As I mentioned above, the confinement is weak and the trap is pretty unstable. Its possible that they have improved the longevity of the trap in the last few months, but when doing long-exposure the device is surrounded by a heavy cloth barrier to block both external light and air movement.
What kind of safety equipment is required to view it?