Derivation of the impulse response: comments on the method of Roufs and Blommaert