Abstract: In a speech recognition arrangement, a speech pattern is recognized as one of a plurality of reference patterns for which acoustic feature signal templates are stored. Each template includes a time frame (e.g., 10 millisecond) sequence of spectral features e.g., LPC and nonspectral e.g., acoustic energy (E) normalized to the peak energy over an utterance interval. LPC and normalized energy feature signal sequences are produced for an unknown speech pattern. For each time frame, the correspondence between the LPC features of the speech pattern and each reference is measured as well as the correspodence between the energy (E) features. In comparing the unknown speech features to those of the reference templates, the dynamic time warp distance DT=D.sub.LPC +.alpha.D.sub.E is used where .alpha. is a weighting factor selected to minimize the probability of erroneous recognition.

Abstract: Speech recognition using self-correlation (auto-correlation) parameters and time normalization is improved by first making a preliminary decision using lower-order (zeroth to third) self-correlation coefficients derived from the self-correlation function of order m: ##EQU1## where .phi.(0) is the power, .phi.(1) is the linear or first order coefficient, and so on.

Abstract: In a speech analysis-synthesis system for the narrow band transmission of a speech signal, speech is separated into a plurality of spectrum information (K.sub.i), average (V.sub.0) of linear prediction error signal, pitch period (L), voiced/unvoiced decision signal (V/UV), a pseudo exciting signal (I) which is a part of a linear prediction error signal or an impulse response of the same, and a frame information which determines the interval and the duration of the spectrum analyzation, then, in a receive side, the product of said pseudo excitation signal (I) and said pitch period (L), a white noise is switched according to said voiced/unvoiced decision signal (V/UV), and the output of the switch is applied to a synthesis filter which attaches correlation components to an input signal according to spectrum information (K.sub.i) to provide synthesized speech.

Abstract: Adaptive bit allocation to a set of spectral transform coefficients is optimized by setting a predetermined first (maximum) threshold, then allocating one bit to those coefficients above the threshold, no bits to those below. If the total number of bits has not been allocated, the threshold is lowered an "integer" step to a second threshold, and again one bit allocated to those coefficients above the second threshold. The procedure is iterated until all the total bits are allocated or, if too many bits are allocated after the last integer step, a "fractional" step threshold is used.

Abstract: An electronic talking learning aid is disclosed. The learning aid includes a speech synthesis circuit which, in the embodiment disclosed, includes a digital lattice filter circuit, a voiced/unvoiced speech excitation circuit, a speech parameter interpolator, an input parameter decoder, a digital-to-analog converter circuit and associated timing circuits. The learing aid is also provided with a controller and at least one memory for storing digitized speech parameters as well as other digital data. This other digital data may represent, for instance, correct answers to questions posed by the talking learning aid. A keyboard or other response insertion device is provided to permit an operator to input his or her answers to the questions posed by the learning aid. In the disclosed embodiment, the talking learning aid asks the operator to input the correct spelling of a spoken word and informs the operator whether or not the response was correct.

Abstract: Speech recognition with time warp is simplified by finding a certain portion of a word whose time duration is the same for all speakers. In comparing an unknown speech with a reference speech, the time duration of an unknown speech is coincided with the time length of a reference speech with the two processes. According to the invention, an element vector of a speech is classified to the first portion and the second portion. The former is a consonant and co-articulation which couples the two sounds, and the latter is a vowel. The length of the first portion is almost independent from a speaker, and the length of the second portion depends upon a speaker. Therefore, the present invention matches the first portion of an unknown speech with that of the reference speech directly without changing the time length. Next, the sample elements in the second portion of the unknown speech is linearly matched with that of a reference speech. Thus, excellent recognition is obtained using a simple calculation.

Abstract: A system for recognizing a continuously spoken word sequence with reference to preselected reference words with the problem of coarticulation removed, comprises a pattern memory for memorizing demi-word pair reference patterns consisting of a former and a latter reference pattern segment for each reference word and a word pair reference pattern segment for each permutation with repetition of two words selected from the preselected reference words. A recognition unit is operable as a finite-state automaton on concatenating the demi-word pair reference patterns so that no contradiction occurs at each interface of the reference patterns in every concatenation. It is possible to use the automaton in restricting the number of reference patterns in each concatenation either to an odd or an even positive integer.

Abstract: A point of sale cash register which displays the price of an item as read by a bar code reader, also announces in simplified form a sound representing the price-range of the item such that an error so detected by the operator may be corrected by keyboard.

Abstract: Speech recognition is improved by adaptive update of reference word codes. The apparatus can be used for controlling a piece of equipment and comprises memories which in coded form contains the references of the vocabulary of the machine; a microphone, a coding circuit and a memory for coding and memorizing the word pronounced by the user; a device for displaying the word recognized by the apparatus; and a control circuit for comparing the pronounced word to the different memorized references and for displaying at each repetition of the word the words corresponding to the references in their order of resemblance to the pronounced word. The user is able to train the apparatus easily by repeating a word until the displayed word agrees with what he has said, the coded form of this spoken word then being stored as the reference for the corresponding word of the vocabulary.

Abstract: A signal synthesizer apparatus having a parameter dependent multiplier which includes charge-coupled devices. The multiplier has a gating input to store a charge corresponding to an input signal applied thereto; a transmission circuit for selectively transferring portions of the stored input charge, the selectivity of the transmission circuit being dependent on the prescribed parameter; and an output for receiving and combining all of the charges transferred from the transmission circuit. The transmission circuit comprises a first circuit for selectively dividing the charge corresponding to the input signal into given fragments and for transferring the divided charge, and a second circuit for determining the selectivity of the dividing of the charge according to the prescribed parameter. The rate of the dividing of charge corresponds to the multiplicative coefficient of the multiplier.

Abstract: Speech recognition by first log-compressing the speech signal, then filtering into subband signals which are subtracted to form an indicator of voiced, unvoiced and silence intervals, whereby sequence and duration patterns are coded in +1, -1, and zero.

Abstract: To avoid pinched fingers, a verbal "Stop Window" command causes a stop and momentary reversal of window motion. A spoken-instruction controlled system for an automotive vehicle which can stop a moving device such as a door window closing device after having driven it reversely for a fixed period when a spoken instruction is inputted thereto through a microphone. In the system according to the present invention, even if there exists a time delay from when the driver utters the spoken instruction to when the device stops perfectly, it is possible to stop the device at or before a position at which the driver utters the spoken instruction. The system according to the present invention comprises at least one motor reversing controller and a timer unit, in addition to the conventional spoken-instruction controller including at least one switch, at least one motor stop controller and a speech recognizer.

Abstract: Audible entertainment signals are also filtered to detect speech presence whereby only frequencies in the band 0-150 Hz, are applied to vibrate a listener's chair for special effects.

Abstract: A spoken-instruction controlled system for an automotive vehicle by which a single actuator can be activated repeatedly or a plurality of actuators can be activated sequentially by repeatedly depressing a recognition switch within a predetermined time period after a spoken instruction corresponding to the single actuator or actuators has been recognized by a speech recognizer. Therefore, it is possible to activate an actuator repeatedly or actuators sequentially without uttering the same spoken instruction repeatedly. The spoken-instruction controlled system according to the present invention comprises a two-circuit analog switch, at least one analog switch, at least one retriggerable timer unit, an AND gate, an OR gate, a shift register, etc., in addition to a speech recognizer.

Abstract: A speech recognition system for an automotive vehicle which can selectively activate actuators according to vehicle-travelling conditions in such a way that a spoken instruction such as "Door open" or "Trunk lid open" can be recognized only when vehicle speed is zero, "Headlight high/low beam" can be recognized only when vehicle speed reaches a predetermined value, and "Air conditioner on/off" or "Radio on/off" can be recognized regardless of vehicle speed. The speech recognition system for an automotive vehicle according to the present invention comprises a speed sensor or the like and a speed-dependent speech selection unit including AND gates or analog switches, in addition to the conventional speech recognizer.

Abstract: Disclosed is an electronic copying machine comprising detecting means for detecting an operating status of a copying machine main body, audio information storage means for storing predetermined audio information, audio information synthesizing means for reading out the audio information from said audio information storage means and for synthesizing the audio information, voice output means for converting the audio information synthesized by said audio information synthesizing means to a voice and for outputting the voice, and control means for detecting by said detecting means an operation made by an operator when trouble occurs in said copying machine main body and controlling said detecting means, said audio information storage means, said audio information synthesizing means and said audio output means so as to signal, with a voice, an operator an operation to be performed subsequently.

Abstract: In this pcm digital-to-analog converter, a probability value (soft decision), rather than the conventional one/zero value (hard decision), is given each sampled input of signal-plus-noise as a function of signal-to-noise ratio, thereby reducing the mean squared error in the output analog signal.

Abstract: A vocal indicating device for a copying machine includes a speech synthesizer which says "Please wait a minute" if logic-control circuitry detects the machine in Wait-Mode (e.g., warmup) for more than 5 seconds.

Abstract: A data compressor includes circuit elements for counting the number of bytes transmitted in a binary bit stream and further includes storage circuit elements for temporarily storing each byte transmitted in the stream and the corresponding byte count of each byte. The data compressor also includes combinational logic control circuitry for detecting when one of the bits in a temporarily stored byte has a predetermined logic value and for generating a retain data control signal for signalling that the byte contains data desired to be retained.

Abstract: Voice commands of two types are recognized by an air conditioner: (1) a first command to jump to a different air delivery (below-out) temperature, after which a gradual automatic return to the previously set temperature occurs, or (2) a second command which will stop the automatic return at some arbitrary temperature.