Phase-based TOF systems operate with reduced depth error due to motion blur, and/or spatial blur, and/or pixel offset by intelligently determining how best to combine pixel values, and how best to compensate for individual pixel offsets. Such determination(s) may be carried out on a per pixel basis, dynamically, in real-time during TOF operation, or on archived TOF data. Offsets for individual pixels may be dynamically calculated and subtracted from the values acquired by those pixels Individual pixel offsets may be calculated for example by combining data acquired by the same pixel at two acquisitions, 180° out of phase with respect to each other. Calculated offsets may be averaged, or on a per pixel basis, and if target object motion is detected, one or more offset calculations can be discarded rather than averaged to reduce motion blur. Offsets acquired a priori during a TOF system calibration procedure may be used.