although I agree with the statement below, I would say that X-ray
crystallography also only generates a molecular model (not "a
structure"). Rather than a structure, X-ray crystallography generates
an electron density map. The quality of the model based in it depends
on the quality of this electron density map (esp. resolution) and the
quality of the researchers interpretation of it.
Mark
As a physical principle, spectroscopic
methods do not deliver atomic resolution structures, but a set of
inferences that may or may no be compatible with a molecular model.