Is it possible to distiguish mics when a given recording is listened?

It should be a very good exercise.