Many previous studies were focused on the influence of anatomical, physical, and detection-system parameters on recorded surface EMG signals. Most of them were conducted by simulations. Previous EMG models have been limited by simplifications which did not allow simulation of several aspects of the EMG generation and detection systems. We recently proposed a model for fast and accurate simulation of the surface EMG. It characterizes the volume conductor as a non-homogeneous and anisotropic medium, and allows simulation of EMG signals generated by finite-length fibers without approximation of the current-density source. The influence of thickness of the subcutaneous tissue layers, fiber inclination, fiber depth, electrode size and shape, spatial filter transfer function, interelectrode distance, length of the fibers on surface, single-fiber action-potential amplitude, frequency content, and estimated conduction velocity are investigated in this paper. Implications of the results on electrode positioning procedures, spatial filter design, and EMG signal interpretation are discussed.