A local active noise control system is described which uses a virtual microphone arrangement. This arrangement is based on the assumption that the primary pressure at the physical and at the virtual microphone locations are similar. The implication of this assumption on the acoustic performance of a local system in a diffracted primary sound field is theoretically studied. The results show that the error at the virtual microphone position is lower when the virtual microphone arrangement is in the vicinity of a diffracting surface. A practical local active noise control system in a headrest has been built and used to measure the zone of quiet produced by a single and a dual channel system when the total pressure is canceled at one or two virtual microphone positions. It is shown that this type of arrangement is capable of projecting the zones of quiet further away from the secondary source than the position of the physical microphone. The measured zones of quiet produced by a single-channel system have been compared with the results produced with a theoretical model which includes the diffraction between two spheres, one representing the secondary source and the other the listener’s head. It has been found that, in the frequency range of applicability of a local active noise control system, the two-sphere model predicts the experimental results well. The effect that the diffracting head has on the performance of an adaptive feedforward controller using a physical and a virtual microphone is also explored and the results show that this effect is small provided the initial system identification is performed in the presence of the head.