A new frequency transposition device for the deaf; A simulation and a validation study

Abstract

To increase the high frequency speech information available to the sensory-neural deaf, with low-frequency residual hearing only, a frequency "recoding" device was constructed which "shifts" a selected band of high frequency speech information and superimposes it on the low frequency range, in a manner designed to maintain the 'speechlike' nature of the "recoded" input signal (patent applied for).

The design and evaluation of the "recoding" device are considered in the context of factors likely to be involved in the acquisition of "recoded" speech, e.g. the separation of sounds that are ' speechlike' from those that are not, by the ear-brain system, the interaction in speech processing of auditory, visual and kinaesthetic cues, and the influence of already established strategies for processing "non-recoded" speech on the acquisition of altered strategies for "recoded" speech.

The 'speechlike' nature and the utility of the "recoding" were assessed (a) in a simulation study involving normal hearing subjects under simulated deafness conditions, and (b) in a validation study with sensory-neural deaf children.

In the simulation study significant improvements in the ability to imitate CVC nonsense syllables were brought about both by "recoding" and by visual cues (from articulatory movements) without formal discrimination or3imitation training, the "recoded" high frequency information contributing in particular to imitation of "manner" and "place" of articulation of phonemes with major energy components in the "recoded" High Frequency region (HF phonemes).

Further, in the validation study, "recoding" produced a significant improvement in the articulation learning of HF phonemes, indicating (together with the simulation study findings) that the "recoded" signals were sufficiently 'speechlike' to be of use to the ear-brain system in speech processing.

It was concluded therefore that the generality of utility (to the hearing impaired) of the "recoding" mode proposed, merits serious further investigation.