Abstract: The memory and data management requirements for text independent speaker dependent recognition are drastically reduced by using a novel approach that eliminates the need for separate acoustic recognition models for each speaker. This is achieved by using speaker independent recognition models at the acoustic level. The speaker dependent data stored for each item to be recognized consists only of information needed to determine the speaker independent recognition model sequence for that item.

Abstract: An electronic interpreting machine comprising vocal input device for vocal input of a language, dictionary device incorporating an input language dictionary and an output language dictionary, language setting device for setting the input language as the first language and the output language as the second language, voice recognition device for recognizing and storing the first language, translating device for translating the first language which has been recognized into the selected second language, voice information generating device for generating voice information representing the translated second language, and voice output device for giving output of the voice information, wherein the maximum amount of language information which can be vocally input is set beforehand, while information volume computing device that computes the ratio of the amount of information which has been input to the vocal input device to the maximum amount of information which can be input in real time, and computed information not

Abstract: An improved method and an apparatus for reducing an undesirable characteristic of a spectral estimate of a noise signal (203) between occurrences of voice signals (202) in an input signal (117). A spectrum of the input signal (117) is estimated to produce a spectral estimate (119) of the input signal (117) including the undesirable characteristic of the noise signal (203). The spectrum of the input signal (117) is smoothed over a first bandwidth (f3-f4) during the occurrences of the voice signals (203) and over a second bandwidth (f2-f5), substantially greater than the first bandwidth (f3-f4), between the occurrences of the voice signals (203). Alternatively, the spectral estimate (119) of the input signal (117) is filtered between the occurrences of the voice signals (203).

Abstract: A long-term analysis of an input speech signal is carried out to adaptively select parameters of a pitch synthesis filter in respective variation ranges. Successively selected values of said parameters are processed to estimate maximum magnitudes of an error component of the output signal of the pitch synthesis filter. The variation range of at least one of said parameters is determined on the basis of the estimated maximum magnitudes.

Abstract: In a multipulse processing device for achieving a high encoding efficiency without using a high sampling frequency for an input signal and with a great degree of freedom given to positions of multipulses, an input speech signal is subjected by an LPC analyzer/processor 3 to LPC analysis of each analysis frame for extraction of LPC coefficients after sampled by an A/D converter 2. Multipulses are retrieved as a result of decision by a multipulse analyzer 20 with a degree of freedom given relative to sampling points of a sampled speech signal supplied through an auditorily weighting filter 4. Encoded by an encoder 41 and together with k parameters used as an example of the LPC coefficients, retrieved multipulses are multiplexed by a multiplexer 42 for delivery to a synthesis side. A multipulse waveform synthesizer 45 synthesizes a waveform by using decoded multipulse data and the LPC coefficients.

Abstract: A computer system having a control unit, a storage, a first register for storing a scale value transmitted from the storage, a second register for storing ADPCM sound data transmitted from the storage, and an ADPCM decoder for reproducing sound in accordance with calculating the ADPCM sound data and the scale value and holding the ADPCM sound data and the scale value therein until next ADPCM sound data are transmitted thereto normally. The ADPCM decoder reproduces the sound from the middle of the ADPCM sound data by using the ADPCM sound data and scale value held therein.

Abstract: The present invention provides a device, method and system of noise injection to maximize compressed audio quality while enabling bitrate scalability. It includes at least one of an encoder and a decoder. The encoder includes a zero detection unit, coupled to receive a frequency domain quantized signal, for determining a control signal that indicates whether noise injection is implemented and a normalization computation unit, coupled to receive at least unquantized signal values and the control signal, for determining a normalization term in accordance with the control signal. The decoder includes a zero detection unit, coupled to receive a frequency domain quantized signal, for determining a control signal that indicates when noise injection is active and a noise generation and normalization unit, coupled to receive a normalization term and the control signal, for generating, normalizing, and injecting a predetermined noise signal where indicated by the control signal.

Abstract: A computer system having a storage, calculating portion, an ADPCM decoder, and recovery portion. The recovery portion calculates a recovery value based on control data and an initial scale value and replaces a reproduced value with the recovery value if a reproducing operation stops halfway through a sound data. Thus, the ADPCM decoder reproduces sound from the middle of the sound data.

Abstract: A language identification and verification system is described whereby language identification is determined by finding the closest match of a speech utterance to multiple speaker sets. The language identification and verification system is implemented through use of a speaker identification/verification system as a baseline to find a set of well matched speakers in each of a plurality of languages. A comparison of unknown speech to speech features from such well-matched speakers is then made and a language decision is arrived on based on a closest match between the unknown speech features and speech features for such well matched reference speakers in a particular language.

Abstract: A method and apparatus is provided for a low bit rate speech transmission. Speech spectral parameter vectors are generated from a voice message and stored in a sequence of speech spectral parameter vectors within a speech spectral parameter matrix. A first index identifying a first speech parameter template corresponding to a first speech spectral parameter vector of the sequence of speech spectral parameter vectors is transmitted. A subsequent speech spectral parameter vector of the sequence is selected and a subsequent speech parameter template is determined having a subsequent index. One or more intervening interpolated speech parameter templates are interpolated between the first speech parameter template and the subsequent speech parameter template. The one or more intervening speech spectral parameter vectors are compared to the corresponding one or more intervening interpolated speech parameter templates to derive a distance.

Abstract: An individual's speech sample, obtained for the purpose of speaker verification, is used to create a "protected" model of the speech. The protected model, which is stored in a database in association with a personal identifier for that individual, is arranged so that the characteristics of the individual's speech cannot be ascertained from the protected model, without access to an encryption key or other secured information stored in the system.

Abstract: An object, such as a robot, is located at an initial state in a finite state space area and moves under the control of the unsupervised neural network model of the invention. The network instructs the object to move in one of several directions from the initial state. Upon reaching another state, the model again instructs the object to move in one of several directions. These instructions continue until either: a) the object has completed a cycle by ending up back at a state it has been to previously during this cycle, or b) the object has completed a cycle by reaching the goal state. Upon reaching a state, the neural network model calculates a level of satisfaction with its progress towards reaching the goal state. If the level of satisfaction is low, the neural network model is more likely to override what has been learned thus far and deviate from a path known to lead to the goal state to experiment with new and possibly better paths.

Abstract: Text may be converted to audible signals, such as speech, by first training a neural network 106 using recorded audio messages 204. To begin the training, the recorded audio messages are converted into a series of audio frames 205 having a fixed duration 213. Then, each audio frame is assigned a phonetic representation 203 and a target acoustic representation 208, where the phonetic representation 203 is a binary word that represents the phone and articulation characteristics of the audio frame, while the target acoustic representation 208 is a vector of audio information such as pitch and energy. After training, the neural network 106 is used in conversion of text into speech. First, text that is to be convened is translated to a series of phonetic frames 401 of the same form as the phonetic representations 208 and having the fixed duration 213. Then the neural network produces acoustic representations in response to context descriptions 207 that include some of the phonetic frames 401.

Abstract: A noise reducing method for speech signals is provided in which the probability of speech occurring is calculated by spectral subtraction of subtracting the estimated noise spectrum from the spectrum of the input signal, and the maximum likelihood filter is adaptively controlled based upon the calculated speech occurrence probability. Adjustment to an optimum suppression factor may be achieved depending on the SNR of the input speech signal, so that is it unnecessary for the user to effect adjustment prior to practical application. In addition, a method for detecting the noise domain is provided in which the value th employed for finding the threshold value Th.sub.1 for noise domain discrimination is calculated using the RMS value of the current frame or the value th of the previous frame multiplied by the coefficient .alpha., whichever is smaller, and the coefficient .alpha. is changed over depending on the RMS value of the current frame.

Abstract: To prevent a great variation of a detected wheel velocity for a period of time after a front wheel sensor or a rear wheel sensor for detection of a wheel velocity actually fails until a microcomputer detects the failure, it is determined whether or not an output pulse period of a front wheel sensor is equal to or shorter than a predetermined period. When the output pulse period exceeds the predetermined period, it is determined that there is the possibility that the front wheel sensor may be in failure, thus a front wheel velocity for a predetermined number of control cycles before, or a front wheel velocity before the failure, is adopted as a front wheel velocity of the present control cycle. Also, a rear wheel velocity is calculated in a similar manner. For a time after a failure actually occurs until the failure is detected by a microcomputer, driving force control is not started even if other starting conditions for driving force control are satisfied.

Abstract: High level synthesis of datapaths has traditionally concentrated on synthesizing a specific implementation for a given computational problem. Methods to compose a reconfigurable BISR (built-in-self-repair) implementation with a minimum amount of area overhead are disclosed. Previously the BISR scope has been restricted to the substitution of operation modules with only those of the same type. Novel resource allocation, assignment and scheduling and transformation methods, primarily for ASIC designs, are described. These methods are based on the exploration of the design solution space accomplished by use of high level synthesis processes to find designs where resources of several different types can be backed up with the same unit.

Abstract: A method of and an apparatus for placing blocks in a semiconductor integrated circuit determined the placement of a plurality of blocks having different sizes. An externally contacting frame is set for the plurality of blocks initially placed in such a manner as to eliminate overlapping between the blocks. The blocks are moved by compressing the externally contacting frame by a predetermined length in a first direction, and the blocks in which overlapping has consequently occurred are moved, or deformed without changing their areas. Further, the blocks are moved by compressing the externally contacting frame by a predetermined length in a second direction perpendicular to the first direction, and the blocks in which overlapping has consequently occurred are moved, or deformed without changing their areas. The compression is repeated in the two directions until a target size is obtained.

Abstract: A register circuit an arithmetic circuit a register circuit and a logic circuit form a bit slice cell corresponding to a path of propagation connecting the circuits in this order. Similarly, an arithmetic circuit register circuits and a logic circuit form a bit slice cell and an arithmetic circuit register circuits and a logic circuit form a bit slice cell. The bit slice cells are arranged generally in parallel to form a bit slice circuit which prevents redundant lines for connecting the functional circuits, whereby the bit slice circuit is developed in a short period without a decreased degree of integration and prolonged delay time.

Abstract: A control unit for vehicle is provided where an increase in a number of input/output and upgrading functions can easily be performed and, in addition to this, changing of a program in a control unit can easily be performed even when a single-chip micro-computer is used. A control unit is constructed such as to comprise an interface software memory storing an interface software program for connecting an application software program with an operating system in an internal ROM, a CPU for performing computation of the application software program and the interface software program, a RAM storing data such as the result of the computation, an I/O unit for extending the control unit, and an extending device for communicating memory data through a bus or a LAN. A single-chip micro-computer used in vehicle control can easily cope with an increase in the input/output ports or additional functions, and an application software program can continuously be used only by rewriting an interface software program.

Abstract: A Global Positioning System Synchronized Active Light Autonomous Docking System (GPSSALADS) for automatically docking a chase vehicle with a target vehicle comprising at least one active light emitting target which is operatively attached to the target vehicle. The target includes a three-dimensional array of concomitantly flashing lights which flash at a controlled common frequency. The GPSSALADS further comprises a visual tracking sensor operatively attached to the chase vehicle for detecting and tracking the target vehicle. Its performance is synchronized with the flash frequency of the lights by a synchronization means which is comprised of first and second internal clocks operatively connected to the active light target and visual tracking sensor, respectively, for providing timing control signals thereto, respectively.