On the Teager-Kaiser energy operator "low frequency error"