After making such adjustment that an a main push-pull
signal MPP offset can be made zero when a main beam is positioned at the center of a main photoreceptor, using electric signals E and F photoelectrically converted by one of two two-divided side beam photoreceptors and electric signals H and G photo-electrically converted by the other thereof, a side push-pull
signal SPP including a constant α is determined for allowing respective (F−αE) and (αH−G) offsets to be made zero and is calculated according to the formula, SPP−(F−αE)+(αH−G). From this SPP
signal and the main push-pull signal MPP, a differential push-pull signal DPP is then calculated, thereby allowing stable tracking control to be carried out, even during recording.