Towards the Simulation of Pathological Voice Qualities