I have to agree with blakbird. they probably can detect the different vibrations and feel different to them. And they associate the different vibrations with different people. Thats probably how they detect the different voices of different people.
They learn who you are, and they learn your vibrations. If you kept just saying "Fishy" and looking right at him, he knew you were talking to him. He probably learned the vibrations of his name, and can now pick them up.