Abstract: Memory chips with data memory (202), embedded logic (206) and broadcast memory (204) for two modes of operation are disclosed. A first mode of operation is the usual memory mode expected of a data RAM. The second mode of operation allows localized computation and/or processing of the data in data memory (202) by the embedded logic (206) with minimal handshaking with a remote CPU. In a functioning system, the memory chips are organized in a hierarchical manner and include address-associative memory systems.

Abstract: Speaker verification is performed by computing principal components of a fixed text statement comprising a speaker identification code and a two-word phrase, and principal spectral components of a random word phrase. A multi-phrase strategy is utilized in access control to allow successive verification attempts in a single session, if the speaker fails initial attempts. Based upon a verification attempt, the system produces a verification score which is compared with a threshold value. On successive attempts, the criterion for acceptance is changed, and one of a number of criteria must be satisfied for acceptance in subsequent attempts. A speaker normalization function can also be invoked to modify the verification score of persons enrolled with the system who inherently produce scores which result in denial of access. Accuracy of the verification system is enhanced by updating the reference template which then more accurately symbolizes the person's speech signature.

Abstract: An active memory 14 is provided which includes a data memory 20 including rows and columns of storage locations for holding data and computational results. A broadcast memory 22 includes rows and columns of storage locations for holding control instructions. Computing circuitry 26 is provided which is operable to perform a first computational operation using first and second words of data retrieved from the data memory 20 and perform a second computational operation using a result from the first operation and a result from a previous operation. Control circuitry 24 is operable in response to control instructions received from broadcast memory 22 to control the transfer of the first and second words of data from the data memory 20 to said computing circuitry 26 and the performance of the first and second operations.

Abstract: A memory system 10 is provided including a processor 12 and an active memory device 14 coupled to a processor 12. Active memory 14 includes a first memory 20 for storing a plurality of possible addresses and a second memory 22 for storing an actual address received from processor 12. Circuitry 26 is provided for identifying at least one active address from ones of the possible addresses stored in first memory 20 as a function of the actual address stored in second memory 22.

Abstract: System (100) receives a speech signal at an input (102) which is measured and transformed by speech feature measuring device (104). The output feature vector from speech feature measuring device (104) is then compared to a reference model in a statistical classification manner. Acoustic similarity measuring device (106) performs statistical measurements while temporal speech model constraints block (108) imposes transitional probabilities to the probability measurements generated by measuring device (106). Acoustic similarity measuring device (106) performs a weighted analysis of the error vector defined between the speech feature vector and reference vector utilized during the analysis.

Abstract: A method and system are provided for time aligning speech. Speech data is input representing speech signals from a speaker. An orthographic transcription is input including a plurality of words transcribed from the speech signals. A sentence model is generated indicating a selected order of the words in response to the orthographic transcription. In response to the orthographic transcription, word models are generated associated with respective ones of the words. The orthographic transcription is aligned with the speech data in response to the sentence model, to the word models and to the speech data.

Abstract: A speech system recognizes words from a spoken phrase that conform to checksum constraints. Grammar rules are applied to hypothesize words according to the checksum constraints. The checksum associated with the phrase is thus inherent in the grammar. Sentences which do not meet a predetermined checksum constraint are not valid under the grammar rules and are therefore inherently rejected. The checksum constraints result in increased recognition accuracy.

Abstract: A method and apparatus are provided for identifying one or more boundaries of a speech pattern within an input utterance. One or more anchor patterns are defined, and an input utterance is received. An anchor section of the input utterance is identified as corresponding to at least one of the anchor patterns. A boundary of the speech pattern is defined based upon the anchor section. Also provided are a method and apparatus for identifying a speech pattern within an input utterance. One or more segment patterns are defined, and an input utterance is received. Portions of the input utterance which correspond to the segment patterns are identified. One or more of the segments of the input utterance are defined responsive to the identified portions.

Abstract: An apparatus and method for verifying the identity of an unknown user of a telephone calling card service is disclosed which receives identification information corresponding to a valid user of the service, requests the unknown user to speak an authorization phrase without disclosing the authorization phrase to the unknown user, receives the unknown user's speech signal in response to the request, and verifies the identity of the unknown user based on the received speech signal. Other apparatus, systems, and methods are also disclosed.

Abstract: A speaker voice verification system uses temporal decorrelation linear transformation and includes a collector for receiving speech inputs from an unknown speaker claiming a specific identity, a word-level speech features calculator operable to use a temporal decorrelation linear transformation for generating word-level speech feature vectors from such speech inputs, word-level speech feature storage for storing word-level speech feature vectors known to belong to a speaker with the specific identity, a word-level speech feature vectors received from the unknown speaker with those received from the word-level speech feature storage, and speaker verification decision circuitry for determining, based on the similarity score, whether the unknown speaker's identity is the same as that claimed. The word-level vector scorer further includes concatenation circuitry as well as a word-specific orthogonalizing linear transformer. Other systems and methods are also disclosed.

Abstract: A speech recognizer which utilizes hypothesis testing to determine formant frequencies for use in speech recognition. A pre-processor (36) receives speech signal frames and utilizes linear predictive coding to generate all formant frequency candidates. An optimum formant selector (38) operates with a comparator (40) to select from the formant candidates those formants which best match stored reference formants. A dynamic time warper (42) and high level recognition logic (44) operate to determine whether or not to declare a recognized word.

Abstract: A speaker verification system receives input speech from a speaker of unknown identity. The speech undergoes linear predictive coding (LPC) analysis and transformation to maximize separability between true speakers and impostors when compared to reference speech parameters which have been similarly transformed. The transformation incorporated a "inter-class" covariance matrix of successful impostors within a database.

Abstract: A cost-effective word recognizer. Each frame of spoken input is compared to a set of reference frames. The comparison is equivalent to embodying the reference frame as an LPC inverse filter, and is preferably done in the autocorrelation domain. To avoid the instability and computational difficulties which can be caused by a high-gain LPC inverse filter, a noise floor is introduced into each reference frame sample. Thus, for each input speech frame, a scalar measures its similarity to each of the vocabulary of reference frames.To achieve connected word recognition based on this similarity measurement, a dynamic programming algorithm is used in which time warping to match a sample to a reference is in effect permitted, and in which matching is performed with unconstrained endpoints.

Abstract: An efficient pruning method reduces central processing unit (CPU) loading during real time speech recognition by instructing the CPU to compare a current state's previously calculated probability score against a predetermined threshold value and to discard hypothesis containing states with probability scores below such threshold. After determining that the current state should be kept, the CPU is directed to locate an available slot in the scoring buffer where information about the current state is then stored. The CPU locates an available slot by comparing the current time-index with the time-index associated with each scoring buffer slot. When they are equal, the slot is considered not available; when the current time-index is greater, the slot is considered available. After the information about the current state is stored, the CPU then sets the current state's backpointer to point at the start state of the current best path if the current states represents a completed model.

Abstract: Method of encoding speech signals which is based upon determining the roots of the linear prediction polynomial describing the spectrum of an analog speech signal, wherein the roots are candidates in determining the formants of the speech signal. The method involves the analysis of respective frames of sampled digital speech data using a linear predictive technique to determine a set of reflection coefficients or K-parameters which are then converted into the equivalent predictor coefficients or A-parameters describing a prediction polynomial having a plurality of roots corresponding to the poles of an all-pole filter characterizing the vocal tract. A modified Bairstow technique is then empolyed for factoring out quadratic factors which are then sorted in an ordered arrangement in terms of ascending bandwidths.

Abstract: A low cost speech recognition system generates frames of received speech having binary feature components. The received speech frames are compared with reference templates, and error values representing the difference between the received speech and the reference templates are generated. At the end of an utterance, if one template resulted in a sufficiently small error value, the word represented by that template is selected as the recognized word.

Abstract: Recognition of sound units is improved by comparing frame-pair feature vectors which helps compensate for context variations in the pronunciation of sound units. A plurality of reference frames are stored of reference feature vectors representing reference words. A linear predictive coder (10) generates a plurality of spectral feature vectors for each frame of the speech signals. A filter bank system (12) transforms the spectral feature vectors to filter bank representations. A principal feature vector transformer (14) transforms the filter bank representations to an identity matrix of transformed input feature vectors. A concatenate frame system (16) concatenates the input feature vectors of adjacent frames to form the feature vector of a frame-pair. A transformer (18) and a comparator (20) compute the likelihood that each input feature vector for a frame-pair was produced by each reference frame. This computation is performed individually and independently for each reference frame-pairs.

Abstract: Speech recognition is improved using reference pattern templates which have an added noise signal (noise floor) to avoid LPC high-gain synthesizer instability at low signal levels. Also, input signal frames have a length one-half that of reference frames whereby dynamic time warp computation steps are cut almost in half.

Abstract: A speech encoder is disclosed quantizing speech information with respect to energy, voicing and pitch parameters to provide a fixed number of bits per block of frames. Coding of the parameters takes place for each N frames, which comprise a block, irrespective of phonemic boundaries. Certain frames of speech information are discarded during transmission, if such information is substantially duplicated in an adjacent frame. A very low data rate transmission system is thus provided which exhibits a high degree of fidelity and throughput.

Abstract: A method for generating connected word templates begins with generating isolated word templates of selected words. The isolated word templates are used to extract a continuous word template from a segment of continuous speech containing the selectd words. Both the isolated word templates and the connected word templates can be used to generate speech to determine the quality of the generated templates through aural judgment.