Abstract: In a speech recognition system which time-normalizes (i.e. aligns by time-warping) each pre-stored reference pattern of feature vectors b.sub.j before comparison with the input signal pattern vectors a.sub.i, an improved time-alignment technique requiring less calculation, by selecting "standard" vectors v.sub.m(j) which approximate the reference vectors thereby simplifying the derivation of the time-warp mapping function j=j(i).

Abstract: The bandwidth of human speech signals, 125 to 3000 Hz, is reduced to approximately 125 to 1500 Hz by folding or shifting the high-frequency consonant-band portion, 1000 to 3000 Hz, into the low-frequency vowel-band portion, 125 to 1000 Hz. After transmission, the reduced speech signal is expanded to reconstruct an approximation of the original speech signal. One embodiment features folding (modulator carrier frequency is 3000 Hz) and modulator feedthrough. Another embodiment features shifting (carrier is 1800 Hz) and no modulator feedthrough. Both embodiments boost lower vowel frequencies prior to modulation, and use no switching circuitry.

Abstract: A digital intercept recorder/announcer system for use in conjunction with a telephone central office. The system performs the function of recorded message announcement for messages such as wrong number, time or weather announcements. The system samples and digitizes audio input signals from local or remote audio sources. Sampled data is stored in sequential fashion in memory so that it may be read out of memory exactly as it was stored. Upon demand, these stored signals are retrieved from memory, reconverted to an analog form and coupled to a subscriber's line for audio transmission to the subscriber. Delta modulation techniques are employed in the system.

Abstract: A portable, lightweight, self-contained stethoscope includes:(a) an enclosure containing a speaker, and having an acoustic wave outlet connectible to stethoscope flexible tubing,(b) an upright housing having upper and lower portions, and an intermediate portion therebetween, to provide a controllable and interdigitally graspable unit,(c) a transducer carried by the housing, the transducer including a microphone located to be placed in proximity to a patient's body,(d) circuitry in the housing and having an input terminal coupled with the transducer, said circuitry including amplifier means and filter means, and a battery power supply coupled with said circuitry, said circuitry having a first output terminal, that terminal and circuit controls protectively mounted between the upper and lower housing portions, and(e) flexible connection means coupling the first output terminal with said speaker,(f) the circuitry including an on-off switch and an indicator light on the housing to indicate when the battery power

Abstract: Human voice sounds are modified to produce the effect of a different person speaking. A signal representative of the original voice sounds is separated into a plurality of voice signal components each having a different frequency band. The frequency of at least one voice signal component is shifted and the voice signal components are recombined to produce a modified voice signal representative of the modified but intelligible voice sounds.

Abstract: A speech recognition system adaptable to noisy environments is disclosed. The system includes a recognition unit for recognizing input speech signals and a noise measuring unit for measuring the intensity of ambient noises. The system also includes a rejection unit responsive to a rejection standard controlled by the intensity of the measured noise for rejecting the rejection results given from the recognition unit when the rejection standard is exceeded.

Abstract: A processor for audio program material which synthesizes simulated reflection and reverberation signals for use in creating the illusion of a specific modeled listening environment. A preferred embodiment of the processor accepts stereophonic audio signals and simulates the multiple reflections of an acoustically modeled environment by recirculating the signals through a pair of digital signal delays which have different delay times. The analog input signals are sampled and converted to digital form by first attenuating the signals with a digitally controlled step attenuator and then encoding the amplitude of the attenuated signal using a ramp comparison technique. The encoded amplitude signals, and a code representative of the attenuator position, are stored in a random access memory and are retrieved after the desired delay. The retrieved digital data are converted back to analog signals by reversing the encoding process, using the same ramp voltage generator.

Abstract: A method of compressing or expanding the frequency of signals while keeping the signals' original gross temporal relationship relies upon an electrooptical processor. An apertured mask is interposed between an area-array charge coupled device (CCD) and a light emitting diode (LED). Optical signals are emitted from the LED at a clock rate which is the same as the vertical shift rate of charge packets in the CCD. Sampling the CCD's horizontal shift register output at varying rates allows a changing of the frequencies of the optical signals or a reoccurring reversal of sequential portions. Weighting the apertured mask to define at least one Gaussian curve or triangular waveform smooths and eliminates breaks in the compressed or expanded frequencies. Sampling the CCD output at a slower rate than the vertical clock rate compresses the frequency of the representative optical signals and if the CCD output is sampled at a faster rate then the signals will be expanded or rearranged in a sequentially reoccurring order.

Abstract: The output signals of a phonograph and of an auxiliary audio signal source, such as a high impedance microphone, electric guitar, or the like, are combined before being fed to the amplifier of the phonograph, whereby the combined signals are amplified without distortion, and are fed to one or more speakers, thus audibly reproducing the phonograph recording with the auxiliary sound signal superimposed thereover. The device which combines the signals includes means for suppressing the base frequencies in the auxiliary signal, and for selectively adjusting the amplitudes of the two signal components.

Abstract: A moving-average notch filter with a well-defined phase characteristic for use in eliminating oscillation frequencies in a sound amplification system, wherein this moving-average filter is designed to have an output Y.sub.K in accordance with the system equation ##EQU1## where b.sub.1 are the weighting coefficients, X.sub.K-i are input samples, and K is a constant. This filter may be designed to have a linear phase characteristic. In one embodiment, the location of the notch in the frequency response of the filter may be made to automatically track the drift of an acoustic oscillation frequency.

Abstract: An apparatus and method of voice encryption uses segment swapping. Features of the invention include weighting the input time-function segments by a Hamming or Hanning Window function, before converting to frequency domain segments.

Abstract: The addition of two digitized audio signals, each of which has been non-linearly compressed and PCM encoded, is accomplished by converting only one of the signal samples (preferably the smaller) to an exponential-format value which may be linearly added to the other.

Abstract: A speech digitizer is disclosed including an analyzer for generating power and filter coefficient parameters representative of an analog speech waveform. The digitizer also includes a pitch detector for generating a digital pitch parameter substantially representing the fundamental periodicity of the waveform and including range restrictor means for restricting the pitch signal to a range of pitches within a predetermined tolerance if the average pitch of the periodicity signal is below a predetermined level. The pitch detector also includes means for determining the number of extreme maximum and minimum points within a predetermined range of an absolute magnitude difference function thereby generating a structure number signal representing a voiced event. The digitizer includes a voicing detector for generating a three-level voicing/unvoicing parameter representing whether the speech waveform is voiced or unvoiced.

Abstract: An antisidetalk signal control circuit for preventing transmission of signals from a first source connected with a first input circuit to an output circuit while allowing the transmission of signals from a second source connected with a second input circuit to the output circuit. The signal control circuit comprises a first switching circuit having an input coupled to the first input circuit, a second switching circuit having an input coupled to the first output circuit through a phase inverter and a third switching circuit having an input coupled to the outputs of the first and second switching circuits. An oscillator drives the first switching circuit with signals having a given phase and the second and third switching circuits with signals of opposite phase. A low-pass filter couples the junction of the first and second switching circuits to the second input circuit and a feedback circuit couples the second input circuit to the junction of the first and second switching circuits.

Abstract: A multi way speaker system comprises an amplifier, frequency dividing network including a plurality of filters, a plurality of speakers respectively responsive to the output signals of the filters, and a summing circuit responsive to the output signals of the filters for producing a composite signal to be fed back to the input of the amplifier to establish a negative feedback loop. Because of the feedback loop undesirable influences by various elements included in the filters are diminished and thus the speakers reproduce audio frequencies with high fidelity. At least one equalizer may be employed in the feedback loop, while auxiliary filters may be used in such a manner that the output terminals of the auxiliary filters are connected to only the summing circuit so that the output signals of the auxiliary filters are added to an output signal of at least one filter of the frequency dividing network.

Abstract: To transmit a number of individual speech channels over a smaller number of transmission channels, the frequency range of each speech channel is broken into sub-channels and each of these is considered separately for operational activity. Composite speech signals are then formed from the active frequency sub-channels of the individual speech channels and these are transmitted with coding signals indicative of their composition.

Abstract: Disclosed is apparatus for and a method of decreasing the time duration of playback of recorded speech without distortion. Speech recorded at a normal rate is played back at a higher rate using two variable length buffers in alternating fashion to receive sampled speech segments. Threshold analysis logic circuitry is provided to continually compare the output voltage levels of both buffers and to switch the output between the buffers to a using device when the voltage level of both buffers is approximately zero and of an equal amplitude. In this way output to the using device consists of segments from both buffers joined with a minimum of generated noise and no associated dead zone without regard to any predetermined segment length.

Abstract: A pre-processing system for a speech recognition system nullifies the distortion in a received signal caused by the transmission line by forming a normalized received signal whereby the distortion is cancelled: the received signal is spectrally averaged to form a mean spectrum which is used to adjust a filter to a characteristic inverse to the distortion; this filter then filters the received signal to form the normalized received signal s.sub.j '(.omega.,t)=S.sub.j (.omega.,t)/S.sub.j (.omega.) which is equal to the normalized source signal =S.sub.i (.omega.,t)/S.sub.i (.omega.).

Abstract: An adaptor is disclosed for use with audiometers to enable the measurement of the masking level difference (MLD) in a person's hearing capability. The adaptor comprises a signal channel and a noise (masker) channel which are connected respectively with the audiometer outputs and the earphones. Means are provided to apply the signal voltage and the noise voltage to respective mixers in the two channels and for selectively inverting the phase of the signal voltage at one of the mixers and for selectively inverting the phase of noise voltage at one of the misers. Various MLD test conditions can be established including binaural and monaural conditions and adjustable intensity at one ear.

Abstract: A TDMA multiplexer-demultiplexer includes a plurality of input/output ports and common control equipment. Each input port is capable of accepting information asynchronous to the TDMA clock and at typical terrestrial clock rates. The input/output ports perform all the necessary signal processing for a TDMA terminal including pulse stuffing, changing continuous asynchronous low rate information to high bit rate burst form as well as forward acting error correcting encoding, if desirable. By incorporating all the signal processing in the multiplexer-demultiplexer, much equipment duplication is eliminated along with eliminating a requirement for multiple high frequency control and timing signal line drivers and buffers. More particularly, the transmit side of each port includes a separate elastic buffer to raise the input rate to a synchronous common higher rate. In addition, on the receive side, clocking on the low speed side of the expansion buffer is derived from the station's transmit clock.