This paper describes a new method to be used in photoreflectance experiments in order to reduce the influence of parasitic effects in the useful signal. Although theoretically the proposed method eliminates all the parasitic effects, the non-ideal behavior of some system components like the modulator finite switching time and the lock-in phase error, reduce the predicted performance and have been simulated using the programming language LabVIEW. Results show that the method has clear advantages over others previously described in the literature. Depending on the experimental facilities, two possible working regimes are discussed.