Speech processor

Claim 2

A signal processing system for converting a speech signal into a data signal for controlling a hearing prosthesis including an implanted electrode array adapted to stimulate the auditory nerve fibres of a patient by the application of electrical currents to selected electrodes in said array, said processing system comprising means (1,2;20,21) for generating an input signal corresponding to a received speech signal, means (3-6;30-36) for estimating the amplitude and frequency of the fundamental voicing component of said speech signal, means (8-10; 22-26, 27, 28) for estimating the amplitude and frequency of the second formant component of said speech signal, means (14;37) for determining whether said speech signal is voiced or unvoiced, programmable means (14;40) for producing instruction data which in use causes the application of said electrical currents to selected electrodes in said array, said programmable means being programmed with data (FIG. 2) defining a predetermined relationship between each electrode and a selected range of said second formant frequencies based on psychophysical testing of the patient and causing selection of said electrodes based on the estimated frequency of said second formant component such as to produce the desired percepts in the auditory-like sensations generated in the patient, means (14;38) for causing stimulation of said selected electrode at a frequency dependent on the estimated frequency of said fundamental voicing component for voiced speech signals and at a lower substantially constant frequency for unvoiced speech signals, said programmable means (14;40) further being programmed to produce data (FIG. 4) which determines the level of stimulation of each selected electrode dependent on said estimated amplitude of said second formant component of said speech signal as well as on predetermined data relating to the sensitivity of each electrode implanted in the patient.

Claim 3

The system of claim 1, wherein said programmable means (14;40) is further programmed with data (FIG. 3) defining a pre-determined relationship between the estimated frequency of said fundamental voicing component of said speech signal and the rate at which each said selected electrode is stimulated.

Claim 4

The system of claim 2, wherein said stimulation rate has a lower limit and an upper limit, said lower limit being substantially higher than said substantially constant lower frequency and said upper limit being lower than the upper limit of said estimated frequency of said fundamental voicing component.

Claim 5

The system of claim 2 or 3, wherein said programmable means (14;40) is further programmed with alternative data defining a different relationship between said electrode stimulation rate and said estimated frequency of said fundamental voicing component which is more suitable for auditory-like sensations resembling music or certain types of speech signal.

Claim 6

The system of claim 1, 2 or 3, wherein said programmable means (14;40) is programmed to compare the instantaneous values of the estimated frequency F2 of the second formant component of said speech signal and the estimated amplitude A0 of the fundamental voicing component of said speech signal to determine a voiced speech signal when the instantaneous value of F2 is low and the instantaneous value of A0 is high and to determine an unvoiced speech signal when the instantaneous value of F2 is high and the instantaneous value of A0 is low.

Claim 7

The system of claim 1, 2 or 3, wherein means (14,37) are provided to compare the instantaneous high frequency energy of said speech signal with the instantaneous low frequency energy of said speech signal and for causing said selector electrode to be stimulated at said lower substantially constant when said instantaneous high frequency energy is high compared to said instantaneous low frequency energy.

Claim 8

The system of claim 1, 2 or 3, wherein the frequency of said second formant component of said speech signal is estimated by means of a single high pass filter (8;22) having characteristics which produce an output signal in which the second formant frequency component of said speech signal dominates the first and third formant frequency components of said speech signal.

Claim 9

The system of claim 7, wherein high pass filter (8;22) is a two pole filter having a cutoff point of about of 1500 Hz and a Q of 2, said filter being used in conjunction with a zero crossings counter (9;23) which estimates the said frequency of second formant component over a short period.

Claim 10

The system of claim 8, further including a circuit (25) for converting said estimated frequency to a proportional voltage signal and an analogue to digital converter for converting said voltage signal into digital data form.

Claim 13

A signal processing system for converting a speech signal into a data signal for controlling a hearing prosthesis including an implanted electrode array adapted to stimulate the auditory nerve fibres of a patient by the application of electrical currents to selected electrodes in said array, said processing system comprising: means (1,2;20,21) for generating an input signal corresponding to a received speech signal, means (3-6;30-36) for estimating the amplitude and frequency of the fundamental voicing component of said speech signal, means (8-10; 22-26, 27, 28) for estimating the amplitude and frequency of at least one formant component of said speech signal, means (14;37) for determining whether said speech signal is voiced or unvoiced, programmable means (14;40) for producing instruction data which in use causes the application of said electrical currents to selected electrodes in said array, said programmable means being programmed with data (FIG. 2) defining a predetermined relationship between each electrode and a selected range of said at least one formant frequencies based on psychophysical testing of hearing impaired persons and causing selection of said electrodes based on the estimated frequency of said at least one formant component such as to produce the desired percepts in the auditory-like sensations generated in the patient, and means (14;38) for causing stimulation of at least one selected electrode at a frequency dependent on the estimated frequency of said fundamental voicing component for voiced speech signals and at a lower substantially constant frequency for unvoiced speech signals, said programmable means (14;40) further being programmed to produce data (FIG. 4) which determines the level of stimulation of each selected electrode dependent on said estimated amplitude of said at least one formant component speech signal as well as on predetermined data relating to the sensitivity of each electrode implanted in the patient.

Claim 14

A signal processing system for converting a speech signal into a data signal for controlling a hearing prosthesis including an implanted electrode array adapted to stimulate the auditory nerve fibres of a patient by the application of electrical currents to selected electrodes in said array, said processing system comprising: means (1,2;20,21) for generating an input signal corresponding to a received speech signal, means (3-6;30-36) for estimating the amplitude and frequency of the fundamental voicing component of said speech signal, means (8-10; 22-26, 27, 28) for estimating the amplitude and frequency of at least one formant component of said speech signal, means (14;37) for determining whether said speech signal is voiced or unvoiced, programmable means (14;40) for producing instruction data which in use causes the application of said electrical currents to selected electrodes in said array, said programmable means being programmed with data (FIG. 2) defining a predetermined relationship between each electrode and a selected range of said at least one formant frequencies based on psychophysical testing of hearing impaired persons and causing selection of said electrodes based on the estimated frequency of said at least one formant component such as to produce the desired percepts in the auditory-like sensations generated in the patient, and means (14;38) for causing stimulation of at least one selected electrode at a frequency dependent on the estimated frequency of said fundamental voicing component for voiced speech signals and at a lower substantially constant frequency for unvoiced speech signals, said programmable means (14;40) further being programmed to produce data (FIG. 4) which determines the level of stimulation of each selected electrode dependent on said estimated amplitude of said at least one formant component speech signal as well as on predetermined data relating to the sensitivity of each electrode implanted in the patient.

Claim 15

The system of claim 11, wherein said programmable means (14;40) is further programmed with data (FIG. 3) defining a predetermined relationship between the estimated frequency of said fundamental voicing component of said speech signal and the rate at which each said selected electrode is stimulated.

Claim 16

The system of claim 12, wherein said stimulation rate has a lower limit and an upper limit, said lower limit being substantially higher than said substantially constant lower frequency and said upper limit being lower than the upper limit of said estimated frequency of said fundamental voicing component.

Claim 17

The system of claim 12 or 13, wherein said programmable means (14;40) is further programmed with alternative data defining a different relationship between said electrode stimulation rate and said estimated frequency of said fundamental voicing component which is more suitable for auditory-like sensations resembling music or certain types of speech signals.

Claim 18

The system of claim 11, 12 or 13, wherein said programmable means (14;40) is programmed to compare the instantaneous values of the estimated frequency F2 of the second formant component of said speech signal and the estimated amplitude A0 of the fundamental voicing component of said speech signal to determine a voiced speech signal when the instantaneous value of F2 is low and the instantaneous value of A0 is high and to determine the unvoiced speech signal when the instantaneous value of F2 is high and the instantaneous value of A0 is low.

Claim 19

The system of claim 11, 12 or 13, wherein means (14,37) are provided to compare the instantaneous high frequency energy of said speech signal with the instantaneous low frequency energy of said speech signal and for causing said selector electrode to be stimulated at said substantially lower constant when said instantaneous high frequency energy is high compared to said instantaneous low frequency energy.

Claim 20

The system of claim 15, wherein the frequency of said second formant component of said speech signal is estimated by means of a single high pass filter (8;22) having characteristics which produce an output signal in which the second formant frequency component of said speech signal dominates the first and third formant frequency components of said speech signal.

Claim 23

A signal processing system for converting a speech signal into a data signal for controlling a hearing prosthesis including an implanted electrode array adapted to stimulate the auditory nerve fibres of a patient by the application of electrical currents to selected electrodes in said array, said processing system comprising: means (1,2;20,21) for generating an input signal corresponding to a received speech signal, means (3-6;30-36) for estimating the amplitude and frequency of the fundamental voicing component of said speech signal, means (8-10; 22-26, 27, 28) for estimating the amplitude and frequency of at least one formant component of said speech signal, means (14;37) for determining whether said speech signal is voiced or unvoiced, programmable means (14;40) for producing instruction data which in use causes the application of said electrical currents to selected electrodes in said array, said programmable means being programmed with data (FIG. 2) defining a predetermined relationship between each electrode and a selected range of said at least one formant frequencies based on psychophysical testing of hearing impaired persons and causing selection of said electrodes based on the estimated frequency of said at least one formant component such as to produce the desired percepts in the auditory-like sensations generated in the patient, and means (14;38) for causing stimulation of at least one selected electrode at a frequency dependent on the estimated frequency of said fundamental voicing component for voiced speech signals and at a lower substantially constant frequency for unvoiced speech signals, said programmable means (14;40) further being programmed to produce data (FIG. 4) which determines the level of stimulation of each selected electrode dependent on said estimated amplitude of said at least one formant component and said estimated amplitude of said fundamental voicing component of said speech signal as well as on predetermined data relating to the sensitivity of each electrode implanted in the patient.

Claim 24

A signal processing system for converting a speech signal into a data signal for controlling a hearing prosthesis including an implanted electrode array adapted to stimulate the auditory nerve fibres of a patient by the application of electrical currents to selected electrodes in said array, said processing system comprising: means (1,2;20,21) for generating an input signal corresponding to a received speech signal, means (3-6;30-36) for estimating the amplitude and frequency of the fundamental voicing component of said speech signal, means (8-10; 22-26, 27, 28) for estimating the amplitude and frequency of at least one formant component of said speech signal, means (14;37) for determining whether said speech signal is voiced or unvoiced, programmable means (14;40) for producing instruction data which in use causes the application of said electrical currents to selected electrodes in said array, said programmable means being programmed with data (FIG. 2) defining a predetermined relationship between each electrode and a selected range of said at least one formant frequencies based on psychophysical testing of hearing impaired persons and causing selection of said electrodes based on the estimated frequency of said at least one formant component such as to produce the desired percepts in the auditory-like sensations generated in the patient, and means (14;38) for causing stimulation of at least one selected electrode at a frequency dependent on the estimated frequency of said fundamental voicing component for voiced speech signals and at a lower substantially constant frequency for unvoiced speech signals, said programmable means (14;40) further being programmed to produce data (FIG. 4) which determines the level of stimulation of each selected electrode dependent on said estimated amplitude of said at least one formant component and said estimated amplitude of said fundamental voicing component of said speech signal as well as on predetermined data relating to the sensitivity of each electrode implanted in the patient.