Abstract: A model is provided for formants found in human speech. Under one aspect of the invention, the model is used in formant tracking by providing probabilities that describe the likelihood that a candidate formant is actually a formant in the speech signal. Other aspects of the invention use this formant tracking to improve the model by regenerating the model based on the formants detected by the formant tracker. Still other aspects of the invention use the formant tracking to compress a speech signal by removing some of the formants from the speech signal. A further aspect of the invention uses the formant model to synthesize speech. Under this aspect of the invention, the formant model is used to identify a most likely formant track for the synthesized speech. Based on this track, a series of resonators are used to introduce the formants into the speech signal.

Abstract: A method and apparatus is provided for two-tier noise rejection in speech recognition. The method and apparatus convert an analog speech signal into a digital signal and extract features from the digital signal. A hypothesis speech word and a hypothesis noise word are identified from respective extracted features. The features associated with the hypothesis speech word are examined in a second tier of noise rejection to determine if the features are more likely to represent noise than speech. The hypothesis speech word is replaced by a noise marker if the features are more likely to represent noise than speech.

Abstract: A method and apparatus are provided for positioning a head over a disc in a disc drive while maintaining servo loop stability. The apparatus includes an actuator-head assembly having a large scale actuator and a micro-actuator that are both able to move the head over the disc. An anti-windup compensation component detects when a micro-actuator controller is producing a micro-actuator control value that will saturate the micro-actuator. Using the micro-actuator control value, the anti-windup compensation component generates a saturation tracking error signal. A summing component then combines a position value that the head generates based on its position over the disc with the saturation tracking error signal to form a position error value that masks the saturation of the micro-actuator.

Abstract: A method and apparatus are provided for generating an adaptive feedforward cancellation signal for a next sector of a disc in a disc drive. The cancellation signal includes at least one tap weight multiplied by at least one trigonometric function. To form the cancellation signal, the method and apparatus first determine a servo loop transfer function relative to a feedforward cancellation component in the disc drive. The transfer function is then inverted to form filter parameters. A position error value is then measured for a current signal and is passed through a filter formed from the filter parameters. This creates a filtered position error value that is used with the tap weights of a current sector to determine the tap weights for the cancellation signal of the next sector.

Abstract: Methods of encoding and decoding as well as an encoder and decoder are provided for encoding data words into codewords and decoding codewords into data words. The data words are encoded by mapping each data word into a number of data segments. Each data segment is then encoded to form a codeword segment that has the same number of bits as the data word segment. The codeword segments are concatenated to form the codeword. The codewords are decoded by decoding individual codeword segments into data word segments that are the same size as the codeword segments. The data word segments are then mapped into the data word, which has fewer bits than the total number of bits across all data word segments.

Abstract: A mechanism for providing a multi point bolting action to secure a door or other leaf movable within a frame is disclosed. The mechanism comprises a drive gear and an axle gear mounted to rotate about substantially the same axis. The idle gear is arranged to rotate in one sense when the drive gear rotates in the other sense. The mechanism also comprises at least one first member arranged to move between a thrown position and a retracted position when the drive gear rotates, and at least one second member arranged to move between a thrown position and a retracted position when the idle gear rotates. The first and second members are each provided with a rack held in intermeshing engagement with the appropriate gear.

Abstract: A method and apparatus are provided for identifying errors in a detected sequence of values generates the detected sequence of values from channel samples using a detector designed for a channel with a first channel response characteristic. The detected sequence of values is filtered using a filter that emulates a second channel response characteristic to produce a sequence of simulated second channel samples. The channel samples provided to the detector are also provided to a second filter which emulates a response characteristic equal to the second channel response characteristic divided by the first channel response characteristic. This produces a sequence of actual second channel samples. A sequence of difference values is then generated by subtracting each of the simulated second channel samples from a corresponding actual second channel sample. Using the sequence of difference values, at least one error is identified in the detected sequence of values.

Abstract: A method of generating a position error estimate generates a phase field signal and a position error field signal. A set of operations is performed on the phase field signal and the same set of operations is performed on the position error field signal. The result obtained by performing the set of operations on the position error field signal is divided by the results obtained by performing the set of operations on the phase field signal. The result of the division is the position error estimate. In addition, a demodulation circuit is provided that utilizes field ratioing.

Abstract: A computer-readable medium has computer-executable instructions for retrieving a search query provided by a user and locating an ambiguity in the search query that affects the optimization of the search query. The instructions also provide for requesting additional information from the user to clarify the text of the search query. Additional aspects of the invention allow the user to see and select past search queries. Further aspects of the invention automatically construct a boolean search query from a non-boolean query. Other aspects of the invention modify a search query based on its scope. In one embodiment of the invention a search query is modified based on natural language parse data produced by performing a natural language parse of the search query.

Abstract: An input device for a computer system includes a housing and an output circuit located in the housing. The output circuit is capable of providing an output signal to the computer system that is indicative of movement of at least a portion of the input device. A physical feedback device is also located in the housing and is capable of stimulating a localized area of the housing in response to a signal from the computer system.

Abstract: A system and method for synthesizing a facial image, compares a speech frame from an incoming speech signal with acoustic features stored within visually similar entries in an audio-visual codebook to produce a set of weights. The audio-visual codebook also stores visual features corresponding to the acoustic features. A composite visual feature is generated as a weighted sum of the corresponding visual features, from which the facial image is synthesized. The audio-visual codebook may include multiple samples of the acoustic and visual features for each entry, which corresponds to a sequence of one or more phonemes.

Abstract: A method is implemented in a computerized system that provides access to a search tool capable of searching at least one stored record. The method includes determining whether a search query is a logical query by comparing each search term to a set of logical operators. If a search term is in the set of logical operators it is removed from the search query to produce a modified search query. The modified search query is then passed to a search engine.

Abstract: An asynchronous demodulator and method is provided which determines a position error of a read head relative to a position on a medium in a storage device. The read head generates a-read signal as the read head passes over a servo area on the medium. The demodulator generates a normal demodulating signal that is asynchronous with the read signal and a quadrature demodulating signal that is ninety degrees out of phase with the normal demodulating signal. The read signal is multiplied by the normal demodulating signal and the quadrature demodulating signal to produce a normal position signal and a quadrature position signal. The demodulator produces a position error magnitude and a position error direction based on the normal position signal and the quadrature position signal.

Abstract: A detector is provided to detect data values within a data signal that is sampled to provide temporally separated data samples. A first detector portion is configured to determine the location of a first sample vector in a first signal space. A second detector portion is configured to determine the location of a second sample vector in a second signal space. The second detector portion determines the location by using a logic statement to combine a plurality of location indicators. Each location indicator provides the location of the second sample vector relative to a respective boundary surface. The form of the logic statement is independent of the values of the location indicators. In addition, each location indicator is independent of all other location indicators.

Abstract: An encoder is provided that converts data words into code words. The code words, when concatenated together, are such that sub-strings of consecutive first symbols in the code words have no more symbols than a maximum number; sub-strings having the maximum number of consecutive first symbols and beginning at selected locations within the code words appear without restriction, and sub-strings having the maximum number of consecutive first symbols and beginning at locations other than the selected locations appear only where the sub-string is concatenated with an acceptable secondary sub-string. A method is also provided for generating a set of such code words for an encoder.

Abstract: The present invention provides a method and storage device for reducing the occurrence of data miscompare errors. The method includes identifying a destination track as either an even track or an odd track and selecting a pseudo-random set of data based on this identification. The pseudo-random set of data is combined with write data to produce randomized data that is used as the basis for a write signal to be written to the destination track.

Abstract: A prosody modification system and methodology calculates synchronization marks in an original, quasi-periodic signal to a finer precision than the sampling rate of the original signal. Synthetic synchronization marks are generated according to the desired prosody modification also to a finer precision than the sampling rate of the original signal. Waveforms are extracted from the original signal and are fine-shifted to the exact location on the synthetic time axis by a resampling technique. The fine-shifted waveforms are windowed by an asymmetric filtering window, overlapped, and summed together to produce a synthetic signal.

Abstract: The present invention relates to a method of processing speech, in which input speech is processed to determine an input speech vector (or) representing a sample of the speech. A number of possible output states are defined with each output state (j) being represented by a number of state mixture components (m). Each state mixture component is then approximated by a weighted sum of a number of predetermined generic components (x), allowing the likelihoods of each output states (j) corresponding to the input speech vector (or) to be determined.

Abstract: A method of computerized searching receives parameters of a search query from a user and adds a restriction to the parameters to require that at least two of the search terms of the search query appear in a same sentence in a document. A representation of a set of documents is then searched based on the parameters of the search query and the added restriction. Documents that meet the search parameters and the added restriction are thus identified.

Abstract: A method and apparatus are provided that identify confused characters in a text written in a language having a large number of distinct characters. To identify the confused characters, a set of characters from the text are segmented into individual characters. A confusable character for at least one of the segmented characters is then retrieved. Lexical information is identified for both the segmented characters and the retrieved confusable characters and is used to parse the segmented characters and the confusable characters. Based on the parse, a segmented character is identified that has been confused with a confusable character.