Should we use sophisticated, state-of-the-art Speech Feature & Recognition Technology?

A Quote from the Literature:

• The patient's CNS is likely to be far more capable than the "not-appropriately-humble" engineers' designs!  The following excerpt from pp. 72 of White, 1978" reflects a similar view.  This view was gained largely from a wonderful, interdisciplinary education at Berkeley and UCSF.  White was trained in engineering and neurophysiology.

• There is significant redundancy amongst speech cues (i.e., features) in the speech signal, but there is less redundancy in a channel vocoders' output signal (see excerpt 1 from pp. 75 of White, 1978) and even less in a Formant Vocoder's output (see excerpt 2 from pp. 80-81 of White, 1978). Clearly, excerpt 2 strongly emphasizes the importance of NOT ELIMINATING "redundant cues," when in fact the patient's CNS may be more than capable of using such cues in listening environments where these cues are no longer "redundant" at all!