On the interpretation of 1/f noise