Abstract: A word breaker utilizing a lexicon module and a processing module to identify word breaks in a stream of Asian (e.g. Japanese, Chinese, or Korean) language text. The lexicon module is a dictionary or database containing words native to the language of the input text. The processing module includes a plurality of analysis modules which operate on the input text. In particular, the processing module can include modules that analyze the input text using heuristic rules and statistical analysis to identify a first set of work breaks, thereby reducing the size of segments with undefined word breaks. The processing module also includes a database analysis module that identifies the remaining undefined word breaks in those smaller segments that have undergone heuristic or statistical analysis.

Abstract: The system disclosed herein employs a method for facilitating access to supplemental information relating to text items entered into an editing application program by a speech recognition system, e.g. alternate translations of utterances stored in a dictation event database (DEDB). A speech event database (SEDB) associated with the editing program is maintained so as to contain text items entered by said speech recognition system and context information identifying preceding and following text items. In response to a user initiated change in the editing point, a context list is created characterizing a region of text around the new editing point. The speech SEDB event database is searched for text items matching items in said context list thereby to create a mapping of the list to the SEDB. The speech event database is then used as an index into said dictation event database, e.g. for correcting previously misrecognized utterances.

Abstract: A method is given of encoding a speech signal using analysis-by-synthesis to perform a flexible selection of the excitation waveforms in combination with an efficient bit allocation. This approach yields improved speech quality compared to other methods at similar bit rates.

Abstract: A speech signal distribution system includes a transmitting subsystem and one or more receiving subsystems. The transmitting subsystem has a text to speech converter for converting text into a data stream of formant parameters. A supplemental parameter generator inserts into the data stream supplemental data, including linguistic boundary data indicating which parameters in the stream of formant parameters are associated with predefined linguistic boundaries in the text. In one preferred embodiment, the boundary data indicates which formant parameters in the data stream are associated with sentence boundaries. In addition, the supplemental parameter generator optionally inserts the text, lip position data corresponding to phonemes in the text, and voice setting data into the data stream. The resulting data stream is compressed and transmitted to the receiving subsystems.

Abstract: This invention improves information retrieval by providing a tokenizing apparatus and method that parses natural language text in a manner that increases the throughput of an information retrieval or natural language analysis system. The tokenizer includes a parser that extracts characters from the stream of text, an identifying element for identifying a token formed of characters in the stream of text that include lexical matter, and a filter for assigning tags to those tokens requiring further linguistic analysis. The tokenizer, in a single pass through the stream of text, determines the further linguistic processing suitable to each particular token contained in the stream of text.

Abstract: An improved apparatus for generating a spoken message is of the type formed by (i) first recording speech and (ii) then utilizing the recording so as to obtain at least one carrier, each carrier having at least one fixed part and at least one open slot, and then (iii) inserting an argument into each open slot. The improvement provides a phonetico-prosodic parameter generator for characterizing the message in terms of a sequence of phonetico-prosodic parameters for each carrier. An electronic memory stores the phonetico-prosodic parameters corresponding to each carrier and a controller constructs sequences of phonetico-prosodic parameters corresponding to the argument of each open slot. From the phonetico-prosodic parameters, a phonetics-to-speech converter generates a digital sound wave pattern which is converted, by a D/A converter, into an analog sound wave pattern. An output unit provides audible sound waves corresponding to the analog sound wave pattern.