Abstract: A useful, novel, and nonobvious speech verification apparatus, method and media incorporating the same, as well as a secure interactive voice response ("IVR") system. The speaker verification apparatus and method involves a determination whether one or more speech samples captured from a speaker substantially matches a speaker template in a registration database. If no substantial match is found, alternate access control is utilized to allow the speaker to verify herself, without the use of speaker verification techniques. If alternative access control is able to verify the speaker, the captured speech samples are used to modify the corresponding speaker template. The so-modified speaker template may account for acoustic idiosyncrasies or anomalies perceived in the speech samples to improve verification rates in subsequent acoustic attempts.

Abstract: A ring generator circuit provides both a ring signal and talk battery to a subscriber telephone line. The ring generator recognizes whether talk battery voltage or a ring signal voltage is required and supplies the right voltage without the use of external relays. A switching network is coupled to the primary of a transformer to generate AC power at the transformer secondary. The switches intermittently conduct current from a DC power source through the primary of the transformer at a frequency substantially greater than the ring signal. A rectifier circuit converts the generated AC power into DC power to be provided to the telephone line. A sampling circuit samples the AC power, and the sampled power is additively combined with the DC power from the rectifier to form a ring signal component whose frequency is equal to the difference between the sampling frequency and the AC power frequency.

Abstract: A method and apparatus for producing an audio tag in response to calling line identification information. The method includes the steps of, receiving calling line identification information from a telephone line, searching a directory, using at least some of the calling line identification information as a search parameter to locate a directory record corresponding to at least some of the calling line identification information, communicating information held in the directory record or in the calling line identification information to a text to speech converter to produce the audio tag.

Abstract: A ROM for storing constants used in both the MPEG-2 and AC-3 audio decoding algorithms. These constants include (i) matrixing constants for AC-3 and MPEG audio decoding algorithms and (ii) windowing coefficients for AC-3 and MPEG audio decoding algorithms. The ROM includes (a) a first partition for storing a first set of constants which are windowing coefficients for AC-3 decoding, (b) a second partition for storing a second set of constants which are constants used for pre-IFFT and post-IFFT steps of AC-3 decoding (Blkswflag=1), (c) a third partition for storing a third set of constants which are both IDCT coefficients for MPEG decoding and IFFT coefficients for AC-3 decoding, (d) a fourth partition for storing a fourth set of constants which are constants used for pre-IFFT and post-IFFT steps of AC-3 decoding (Blkswflag=0), and (e) a fifth partition for storing a fifth set of constants which are windowing coefficients for MPEG decoding.

Abstract: To accurately determine rate and voice activity in moderate-to-low signal-to-noise ratios (SNRs) to maximize voice quality, system capacity and/or battery life, parameters from a noise suppression system are used as inputs to the rate determination function. Using this method, more of the speech is extracted from the background noise and a lower number of false onsets during fluctuating noise conditions compared with conventional systems are detected. The method is beneficial for voice activity detection (VAD) as well as rate determination (RDA) and unlike other RDA/VAD implementations, is independent of the type of speech coder employed (IS-127, CDG-27, IS-96 and GSM).

Abstract: A system and method enabling a user to remotely access an electronic calendar using a telephone network is provided. The system includes the electronic calendar for storing a plurality of schedules and a scheduler for accessing the schedules as a function of input signals received over the telephone network. Each schedule contains future appointments and events for either daily, weekly or other data interval periods. The scheduler enables selective retrieval and display of the appointments and events for any selected time period. The system is coupled to interfaces allowing the schedule information to be transmitted as audio information, video information, or e-mail information. The scheduler can compare the schedules to one another to automatically generate common meeting times. The common meeting time can then be transmitted to the various attendees using e-mail, audible instructions or voice mail.

Abstract: An object of the present invention is to convert a voice to a character string and output attribute data associated with the character string. The central controller of the data processing apparatus converts a voice entered from the voice input unit through the voice/character code conversion processing to a character string. Then, any date and time part of the character string are detected through the date detection processing and the time detection processing and the detected date and time part are converted to predetermined date and time words through the date conversion processing and the time conversion processing. Finally, the converted date and time are associated to a voice and a character string as attribute data through the monitor controlling and stored in the schedule memory as schedule data. Consequently, the date and time are associated to a voice and a character string automatically.

Abstract: A signal processing arrangement for discriminating a time varying band-limited input signal from other signals using time encoded signals. A received input signal is encoded as a time encoded signal symbol stream from which a fixed size matrix is derived. A plurality of archetype matrices corresponding to a plurality of different input signals are stored, each having been generated by encoding a corresponding input signal into a respective time encoded signal stream from which a respective archetype matrix is derived. A plurality of features are selected and excluded from the archetype matrices to generate corresponding archetype exclusion matrices. An input signal exclusion matrix is generated from the input signal matrix and each of the archetype exclusion matrices. The input signal exclusion matrix is compared with each of the archetype exclusion matrices to generate an output identifying the input signal.

Abstract: An audio decoder makes use of various component sharing techniques and operates to efficiently prevent deadlock without introducing decoding errors or adding significant complexity to the audio decoder. In one embodiment, the audio decoder comprises a bitstreamer, a synchronization controller, a decode controller, a memory module, a data path, and an output buffer. The bitstreamer retrieves compressed data and provides token-aligned data to the synchronization controller and decode controller. The synchronization controller initially controls the bitstreamer to locate and parse audio frame headers. After each frame header is parsed, the decode controller controls the bitstreamer to parse the variable length code compressed transform coefficients. The coefficients are passed to the memory module and data path which operate under the control of the decode controller to inverse transform the coefficients and produce digital output audio data.

Abstract: The invention relates to a method and an apparatus for improving the responsiveness and accuracy of speech recognition systems with tightly integrated Automatic Speech Recognition (ASR) and voice resources, more particularly to a system and method for providing an improved speech recognition enhanced user interface through a reduction in the effective talk-over period. The speech recognition system comprises a host server, managing and arbitrating between a voice resource and an ASR resource, among other resources, through the exchange of run-time controls. The novel method provides for a daisy chain architecture within the speech recognition system, whereby the voice data, or prompt, output by the voice resource is passed directly to the ASR resource, where it is output to the system user. Upon detection of speech from the user, the ASR resource stops transmitting the prompt to the system user, transparent to both the voice resource and the server, followed by a "speech detected" message sent to the server.

Abstract: A method of signal compression. An input signal is divided into a plurality of segments. Each of the segments is decomposed by wavelet packet transform to a plurality of levels, so that each of the segments is in a form of a group of coefficients in each of a plurality of domains corresponding to each of the levels. Each of the domains is divided into a plurality of sub-domains to calculate a degree of characteristic concentration corresponding thereto. One of the domains having a highest degree of characteristic concentration is select to perform bit allocation and quantization, so that the group of coefficients in the selected domain is represented as an information. The information is formatted into a frame for transmission, and then the frame is output.

Abstract: A speech recognition system uses language independent acoustic models derived from speech data from multiple languages to represent speech units which are concatenated into words. In addition, the input speech signal which is compared to the language independent acoustic models may be vector quantized according to a codebook which is derived from speech data from multiple languages.

Abstract: For training a speech recognition to a multi-item repertoire, the following steps are executed: a speech item is presented by a user person, and the distinctivity thereof in the repertoire is asserted. Under control of a distinctivity found the speech item is inserted into the repertoire. These steps are repeated until reaching repertoire sufficiency. In particular, the asserting determines a likeness among the actually presented speech item and all items already in the repertoire, wherein undue likeness with one particular stored item creates a contingency procedure. This implies offering to the user a choice between ignoring the actually presented speech item and alternatively inserting the actually presented speech item at a price of deleting the particular stored item.

Abstract: A method and system that allows users, or maintainers, of a speech-based application to revise the phonetic transcription of words in a phonetic dictionary, or to add transcriptions for words not yet present in the dictionary. The application is assumed to communicate with the user or maintainer audibly by means of speech recognition and/or speech synthesis systems, both of which rely on a dictionary of phonetic transcriptions to accurately recognize speech and pronunciation of a given word. The method automatically determines the phonetic transcription based on the word's spelling and the recorded preferred pronunciation, and updates the dictionary accordingly. Moreover, both speech synthesis and recognition performance are improved through use of the updated dictionary.

Abstract: The invention provides a method for speech coding using Code-Excited Linear Prediction (CELP) producing toll-quality speech at data rates between 4 and 16 Kbit/s. The invention uses a series of baseline, implied and adaptive codebooks, comprised of pulse and random codebooks, with associated gain vectors, to characterize the speech. Improved quantization and search techniques to achieve real-time operation, based on the codebooks and gains, are also provided.

Abstract: A method for managing a speech application, comprising the steps of: categorizing text from a sequential list of playable elements recorded in a dictation session into segments of only dictated playable elements and segments of only non-dictated playable elements; and, playing back the list of playable elements audibly on a segment-by-segment basis, the segments of dictated playable elements being played back from previously recorded audio and the segments of non-dictated playable elements being played back with a text-to-speech engine. The list of playable elements can be played back without having to determine during the playing back, on a playable-element-by-playable-element basis, whether previously recorded audio is available. The list of playable elements can be simultaneously played back audibly and displayed whether the playable elements are dictated or non-dictated.

Abstract: A method, and corresponding program routine, for displaying text in a proofreader associated with a speech recognition application, comprising the steps of: retrieving initial text responsive to a user request; displaying the initial text in a centered position in a display window of a graphical user interface; retrieving further text responsive to a further user request; replacing the initial text with the further text in the centered position of the display window; and, repeating the retrieving further text step and the replacing step until no further text is requested. Previous text and following text can be selectively retrieved for use in the replacing step. The initial text and the further text can be contextual phrases including a target word and words immediately preceding and following the target word. The target words are displayed in the centered position even if the contextual phrase as a whole is not centered in the display window.

Abstract: A data processing apparatus having breath detecting function and an image display control method using breath detection in such a data processing apparatus where in a breathing sound inputted by input means such as a microphone is detected, a feature quantity such as a voice power is transformed into another physical amount such as a temperature and a moving speed, and a display state of an image on a display screen or a driving state of a movable object such as a robot is controlled, whereby a user can feel that the user's breath directly operates the image and robot so that a feeling of physical disorder is eliminated, and thus a difference between the user and a virtual world on the display screen or robot is eliminated.

Abstract: Method and system of determining an out-of-vocabulary score for speaker-independent recognition of user-defined phrases comprises enrolling a user-defined phrase with a set of speaker-independent (SI) recognition models using an enrollment grammar. An enrollment grammar score of the spoken phrase may be determined by comparing features of the spoken phrase to the SI recognition models using the enrollment grammar. The enrollment grammar score may be penalized to generate an out-of-vocabulary score.

Abstract: Adaptive speech coding includes receiving an original speech signal, performing on the original speech signal a current coding operation, and adapting the current coding operation in response to information used in the current coding operation. Adaptive speech decoding includes receiving coded information, performing a current decoding operation on the coded information, and adapting the current decoding operation in response to information used in the current decoding operation.