The best method is the one you described, technically it is called extrusion. It gives the perfect control over size and lamellarity, since you squeeze the vesicles through a filter with pores of a determined size.
Other methods don't give you the same degree of control, but it is nevertheless possible to achieve decent size distribution... but not a definite size. I am a fan of sonication, in case I don't have access to extrusion, which works particularly well for slightly "harder" lipids. Many colleagues of mine are happy with ether/ethanol injection. But neither can replace, IMHO, extrusion.