Phonetic sound recognition apparatus for all voices

Claim 3

The system as set forth in claim 1, wherein said means for dividing said sound wave into a said plurality of detected signals comprise a plurality of passband filters through which are passed said sound wave for obtaining said subbands at their outputs; a plurality of network arrangements each network comprising rectifying diodes, a series-shunt switch, and a capacitor, at the outputs of said plurality of filters, respectively, each of the series-shunt switches having a control means for either connecting the rectified output of the filter to the capacitor associated therewith for storage in proportional signal quantity, or shunting said capacitor for discharge of storage; means for producing pulse signals at some repeating time intervals during switch said frequency variations of said sound wave may occur; and means for applying said pulse signals to the control means of said plurality of switches simultaneously for effecting discharge of said capacitors after said matching signals have been derived.

Claim 4

In a system for analyzing speech sound wave in which the quality of sound and phonetic information are contained within varying one-wavelength time periods of the pitch glottal vocal cord and fundamental frequencies, the latter of which are at least equal to or higher than the glottal pitch frequencies, and in which the phonetic information is contained in a group of resonance peaks lying in a frequency band above the fundamental frequency, as identified by the frequency and amplitude ratios one with respect to another in a group, the system for cancelling out the glottal pitch and vocal cord fundamental frequency components from the sound wave by separating the glottal pitch and vocal cord fundamental components from the components representing the phonetic information comprising means for dividing said sound wave into a plurality of subbands and means for deriving therefrom a plurality of detected signals in proportional amplitudes respectively; means for producing signal quantities as functions of one-wavelength time periods of said vocal cord fundamental frequencies; a plurality of channels, each channel having an input and an output; a plurality of normal ''''on'''' switching means; means for coupling said detected signals in series with said plurality of switching means to the inputs of said channels for admittance to their outputs, respectively; coupling means from said signal amplitudes to said plurality of switching means in an arrangement that, those switching means that pass the detected signals derived from subbands of the said vocal cord fundamental and glottal pitch frequencies are rendered in ''''off'''' operating states by removing from the outputs of said channels those detected signals that are derived from the glottal pitch and vocal cord fundamental subbands; and means for selecting a group of signals from the remaining signals at the outputs of said channels, having the desired mutually related frequency ratio separations in the said subbands from which they had been derived, as representations of the phonetic information.

Claim 5

In a system for analyzing speech sound wave in which the quality of sound and phonetic information are contained within varying one-wavelength time periods of the pitch glottal vocal cord and fundamental frequencies, the latter of which are at least equal to or higher than the glottal pitch frequencies, and in which the phonetic information is contained in a group of resonance peaks lying in a frequency band above the fundamental frequency, as identified by the frequency and amplitude ratios one with respect to another in a group, the system for cancelling out the glottal pitch and vocal cord fundamental frequency components from the sound wave by separating the glottal pitch and vocal cord fundamental components from the components representing the phonetic information comprising means for dividing said sound wave into a plurality of subbands and means for deriving therefrom a plurality of detected signals in proportional amplitudes respectively; means for producing signal quantities as functions of one-wavelength time periods of said vocal cord fundamental frequencies; a plurality of channels, each channel having an input and an output; a plurality of normal ''''on'''' switching means; means for coupling said detected signals in series with said plurality of switching means to the inputs of said channels for admittance to their outputs, respectively; coupling means from said signal amplitudes to said plurality of switching means in an arrangement that, those switching means that pass the detected signals derived from subbands of the said vocal cord fundamental and glottal pitch frequencies are rendered in ''''off'''' operating states by removing from the outputs of said channels those detected signals that are derived from the glottal pitch and vocal cord fundamental subbands; and means for selecting a group of signals from the remaining signals at the outputs of said channels, having the desired mutually related frequency ratio separations in the said subbands from which they had been derived, as representations of the phonetic information.

Claim 6

In speech sound waves containing phonetic information waves, varying glottal pitch frequency waves, and varying vocal cord fundamental frequency waves, and wherein the glottal pitch and vocal cord fundamental frequency waves are distinguishable and wherein also the phonetic information is determined by waves having approximately specific mutually related frequency ratio separations in the sound spectrum the method of determining in the sound waves those waves that represent the phonetic information comprising the steps of determining the glottal pitch and vocal cord fundamental frequency waves in the speech sound waves by first differentiating between the glottal pitch and vocal bord fundamental frequency waves from the rest of the waves; and by determining from said rest of the waves those waves having approximately specific mutually related frequency ratio separation in the sound spectrum for representing the phonetic information aforesaid.

Claim 7

In speech sound waves containing phonetic information waves, varying glottal pitch frequency waves, and varying vocal cord fundamental frequency waves, and wherein the glottal pitch and vocal cord fundamental frequency waves are distinguishable and wherein also the phonetic information is determined by waves having approximately specific mutually related frequency ratio separations in the sound spectrum the method of determining in the sound waves those waves that represent the phonetic information comprising the steps of determining the glottal pitch and vocal cord fundamental frequency waves in the speech sound waves by first differentiating between the glottal pitch and vocal bord fundamental frequency waves from the rest of the waves; and by determining from said rest of the waves those waves having approximately specific mutually related frequency ratio separation in the sound spectrum for representing the phonetic information aforesaid.

Claim 8

The method as set forth in claim 4, wherein is included the step for standardizing said variations in frequency regions of the frequency waves in the sound spectrum, in an order that a reference increment of some parameter represents the lowest frequency of said determined waves are represented by further increments of said parameter whose frequencies differ from the reference increment by the same factor of multiplication as the other frequencies differ from the waves having the lowest frequency in said rest of the waves, whereby standardizing the frequency variations of those waves representing the phonetic information.

Claim 9

In speech sound waves containing phonetic information waves, varying glottal pitch frequency glottal waves, and varying vocal cord fundamental frequency waves, and wherein the glottal pitch and vocal cord fundamental freqUency waves are distinguishable, and wherein also the phonetic information is determined by waves having approximately specific mutually related frequency ratio separations in the sound spectrum, and further yet by approximate specific magnitude ratio relations one with another, the method of determining in the sound wave those wave that represent the phonetic information comprising the steps of determining the glottal pitch and vocal cord fundamental frequency waves by way of their distinguishability by first differentiating between the glottal pitch and vocal cord fundamental frequency waves from the rest of the complex of waves; determining from said rest of the waves those waves having approximately specific mutually related frequency ratio separations in the sound spectrum; and comparing approximately the mutual magnitude ratio relations of said determined waves with that of specific quantity ratio standards, for determining if said determined waves are the correct ones for representing the specific phonetic information.

Claim 10

In speech sound waves containing phonetic information waves, varying glottal pitch frequency glottal waves, and varying vocal cord fundamental frequency waves, and wherein the glottal pitch and vocal cord fundamental freqUency waves are distinguishable, and wherein also the phonetic information is determined by waves having approximately specific mutually related frequency ratio separations in the sound spectrum, and further yet by approximate specific magnitude ratio relations one with another, the method of determining in the sound wave those wave that represent the phonetic information comprising the steps of determining the glottal pitch and vocal cord fundamental frequency waves by way of their distinguishability by first differentiating between the glottal pitch and vocal cord fundamental frequency waves from the rest of the complex of waves; determining from said rest of the waves those waves having approximately specific mutually related frequency ratio separations in the sound spectrum; and comparing approximately the mutual magnitude ratio relations of said determined waves with that of specific quantity ratio standards, for determining if said determined waves are the correct ones for representing the specific phonetic information.

Claim 11

The method as set forth in claim 6, wherein is included the step of standardizing said variations in frequency regions of the frequency waves in the sound spectrum, in an order that a reference increment of some parameter represents the lowest frequency of said determined rest of the waves and the other frequencies of last said determined waves are represented by further increments of said parameter whose frequencies differ from the reference increment by the same factor of multiplication as the other frequencies differ from the waves having the lowest frequency in said rest of the waves, whereby standardizing the frequency variations of those waves representing the phonetic information. 8. In a system for analyzing speech sound waves in which the quality of sound and phonetic information are contained within one wavelength time periods of the varying glottal pitch and vocal cord fundamental frequencies, the latter of which are at least equal to or higher than the glottal pitch frequencies, and in which the phonetic information is contained in a group of resonance peaks lying in a frequency band stove above the vocal cord fundamental frequency, the system comprising means for dividing the resonant components of the sound wave into a plurality of subbands; means for deriving from each of the divided subbands a plurality of detected signals; a bank of channels in a numerical order, starting from channel-1; a plurality of normally ''''off'''' switching means; coupling means from the plurality of detected signals to the said channels via said plurality of switching means in an arrangement that, in each combination the detected signal derived from the lowest frequency component from said subbands is coupled to the channel-1, while the other detected signals are coupled to the other channels by the same factor of multiplication from channel-1 as the other subband frequencies differ from the lowest subband frequency; means for producing a plurality of control signals, each signal representing a specific one wavelength time period of said varying vocal cord fundamental frequencies; coupling means from said control signals to said switching means for ''''on'''' states, each control signal assigned for operation of a specific combination of said signal couplings to said channels in a order that, the signal representing a higher frequency than the vocal cord fundamental frequency in said subbands is switched to the number 1 channel; and means for selecting a group of signals from said channels adaptable to be converted into some parameter as a representation of said phonetic information.

Claim 12

The system as set forth in claim 8, wherein is included a plurality of groups of signal-amplitude-ratio matching means; for coupling said selected group of signals to said plurality of matching means, and means for deriving a closest matching signal from one of said matching means as a representation of said phonetic information

Claim 13

The system as set forth in claim 8, wherein is included apparatus for producing said control signal, comprising means for producing first pulses representative in repetition rate of said vocal cord fundamental frequency; an alternately operating switching means; means for applying the first pulses to said switching means; for producing alternate steady state signals in opposite polarities; means for deriving alternately poled second pulses from said steady state signals, respectively; first and second capacitors; first and second discharger means associated with the first and second capacitors, respectively; means for applying said second pulses to the first and second discharger means alternately for discharging in said first and second capacitors previously stored signal quantities; first and second charging means for charging the first and second capacitors, respectively, at a predetermined charging rate; means for applying said steady state signals to said first and second charging means alternately for charging in signal quantities representative of the time periods between said steady state signals; means for dividing the stored quantities of the first and second capacitors into equally defined control signals; means or mixing signals of said equally defined control signals; and coupling means from said mixed control signals to said plurality of switching means.