The software could first solve for one pair of tones then another, then add or subtract the offset and choose the pair that has the correct fundamental tones.