In principle, yes. In practice, it's painful. I built a vector laser projector back in the 80s when lasers were huge and my XY galvanometer probably cost $1000, but my college was paying for it, so why not? I drove the thing with a Fairlight CMI, onto which I painstakingly drew the parametric equations of X and Y with a light pen. (Nowadays I'd write a program to do that bit for me.) The biggest problem -- other than the cost -- was the fact that the galvos are not perfect: They exhibit lag and resonance at certain frequencies, so you have to cancel all that out and overdrive them in particular ways. This is probably easier today than then, and it can be done more easily with closed-loop feedback than the open-loop way I did it, but it looked amazing.