The digital storage of speech signals in a semiconducting memory is known in the art. Large quantities of storage space are required for storing speech signals and therefore, for long recordings for example dictations, a very large semiconductor memory is needed. This is advantageously implemented using a so-called flash memory. However, the contents of such a memory can only be erased in large sections at a time so that minor changes such as partial erasures and insertions are not possible. The total available memory is subdivided into clusters and a table of contents is created in which the stored files and the state occupation of each cluster is stored, at is known in the art concerning the DOS file manager. Moreover, information which indicates which cluster is the next cluster of a file is stored for each cluster in a file allocation table. This process in accordance with the invention allows for erasures or insertions to be carried out by altering the indication in the file allocation table to the following cluster. In this manner, clusters may be both inserted into and removed from a file. In addition, information on each cluster includes the number of speech frames stored in the cluster and from which address the first speech frame is stored. It is thereby possible to logically remove individual speech frames, forming the smallest subdivision of the stored speech signal, from the file at the beginning or end of a cluster without having to execute an erasing process.

A system and method for compressing a data set. The method comprises the steps of: (i) requesting a data set from a server to be sent to a user, (ii) determining the size of the data set to be sent, (iii) compressing the data set, (iv) assigning a compression flag to the compressed data set, (v) sending the compression flag and compressed data set to the user, (vi) inspecting the data set for compression, (vii) uncompressing the data set; and (viii) processing the data set. Alternatively, upon determining the size of the data set to be sent, assigning an uncompressed flag to the data set, sending the uncompressed flag and data set to the user, inspecting the data set for compression, and processing the data set. In another alternative embodiment, upon determining the size of the data set to be sent, sending the uncompressed data set to the user, inspecting the data set for compression, and processing the data set.

A method of automatic recognition of company names in speech utterances includes generating at least one word sequence hypothesis by a speech recognizer from a speech utterance consisting of one or more words, comparing the word sequence hypothesis with the entries representing company names stored in a database, and selecting, in dependence on the result of the comparison, one company name as a recognition result.

The systems and methods described herein allow dictation and associated routing and formatting information to be forwarded to a transcription system. The transcription system converts the information into a document. The additional information associated with the dictation is then applied to the document to ensure proper formatting, routing, or the like. The completed document is returned to the original dictator for review and proofing. Upon approval, the document is distributed via the transcription system in accordance with distribution information associated with the document.

Disclosed is an MPEG portable sound reproducing system and a method for reproducing sound data compressed using the MPEG method. The inventive system includes power supply means for supplying operational power to the system, the power supply means being realized through a secondary battery; power processing means for rectifying power supplied from the power supply means to stable voltage and current; information display means for displaying numbers and combinations of letters related to operational states of the system; control means for controlling all operations of converting and reproducing sound data compressed using the MPEG method; data storage means for storing MPEG-compressed sound data in a designated address according to signals output from the control means; information selecting means for selecting general operations to reproduce, download, and apply selected sound data stored in the data storage means; sound reproducing means for converting sound data stored in the data storage means into a format audible to users according to signals output from the control means; and transmitting/receiving means for transmitting and receiving sound data and program data from external devices.