Abstract: The present disclosure provides a speech editor for creating a session file having text segments synchronized with an audio file. The editor receives a text segment and a first file that has tag information lines associated with the text segment. Each tag information line has a portion of original text transcribed from the audio file and a corresponding audio length value. The editor stores the audio length of a current line in a new tag information line for the second file, determines whether the original text portion of the current line is found in the text segment, and if the original text portion of the current line is not found in the text segment, identifies a corrected text of the text segment adjusting the audio length of the new tag information line of the session file to correspond to the corrected text.

Abstract: An apparatus comprising a session file, session file editor, annotation window, concatenation software and training software. The session file includes one or more audio files and text associated with each audio file segment. The session file editor displays text and provides text selection capability and plays back audio. The annotation window operably associated with the session file editor supports user modification of the selected text, the annotation window saves modified text corresponding to the selected text from the session file editor and audio associated with the modified text. The concatenation software concatenates modified text and audio associated therewith for two or more instances of the selected text. The training software trains a speech user profile using a concatenated file formed by the concatenating software.

Abstract: An apparatus for transforming data input by dividing the data input into a uniform dataset with one or more data divisions, processing the uniform dataset to produce a first processed dataset with one or more data divisions, processing the uniform dataset to produce a second processed dataset with one or more data divisions, wherein the first and second processed datasets have the same number of data divisions, and editing data selectively within each one of the one or more divisions of the first and second processed dataset. This apparatus has particular utility in error-spotting in processed datasets, and toward training a pattern recognition application, such as speech recognition, to produce more accurate processed datasets.

Abstract: A system and method for creating a final text from an audio file. This has particular utility in completing forms with speech-to-text conversion. The system and method includes transcribing the audio file into a transcribed text file using a speech recognition program. They further include comparing the transcribed text file with a previously created text file to determine differences between the transcribed text file and the previously created text file. Finally, the system and method includes correcting one of the transcribed text file or the previously created text file based upon the differences to create the final text.

Abstract: An apparatus comprising a session file and session file editor with main window and one or more document windows and annotation window and divide/merge and scramble/unscramble features. The session file may include text, audio, image, and other bounded divisions with source data divided into segments or other bounded divisions and other bounded divisions associated to original data. The session file may be derived from processing third-party application output. The session file editor displays text and other content, provides text selection capability and plays back audio of session files with audio-linked text as embedded content, and supports entry of text and password-protected document lock/unlock. The session file editor supports selection of a parent session file and divide, scramble, or merge of bounded divisions to create one or more child session files that may be processed at one or more nodes to create one or more processed child session files.

Abstract: A method for comparing text in a first file to text in a second file. The method includes segmenting text in the first and second files to one word per line; comparing the segmented versions of the versions of the first and second files on a line by line basis; creating a result file using the segmented version of the first file; and augmenting the result file with indication of error using a sandwiching technique. This sandwiching technique includes identifying correct segments that are immediately adjacent any differences identified by comparing the segmented versions of the first and second files on a line by line basis toward sandwiching the erroneous segments between correct segments. Said method incorporates video monitor (26), keyboard (24), and mouse (23), along with microphone (25) and digital recorder (14) for implementing the invention.

Abstract: A system for substantially automating transcription services for multiple users (10, 11, 12) including a manual transcription station (50), speech recognition program (40) and a routing program (200). A uniquely identified voice dictation file is generated from a user and—based on the training status—routes the voice dictation file to a manual transcription station and speech recognition program. A human transcriptionist creates transcribed files for each voice dictation file. The speech recognition program creates written text for each dictation file if the training status is training or automated. If the training status of the current user is enrollment or training, a verbatim file is manually established and the speech recognition program is trained with an acoustic model using the verbatim and voice dictation files. The transcribed file is returned to the user if the training status is enrollment or training or written text is returned if the status is automated.

Abstract: A system for automating transcription services for one or more users. This system receives a voice dictation file from a current user, which is automatically converted into a first written text based on a set of conversion variables. The same voice dictation file is automatically converted into a second written text based on a second set of conversion variables. The first and second sets of conversion variables have at least one difference, such as different speech recognition programs, different vocabularies, and the like. The system further includes a program for manually editing a copy of the first and second written text to create a verbatim text of the voice dictation file. This verbatim text can be delivered to the current user as transcribed text. The verbatim text can also be fed back into each speech recognition instance to improve the accuracy of each instance with respect to the human voice in the file.

Abstract: A system and method for improving the accuracy of a speech recognition program. The system is based on a speech recognition program that automatically converts a pre-recorded audio file into a written text. The system parses the written text into segments, each of which can be corrected by the system and saved in a retrievable manner in association with the computer. The standard speech files are saved towards improving accuracy in speech-to-text conversion by the speech recognition program. The system further includes facilities to repetitively establish an independent instance of the written text from the pre-recorded audio file using the speech recognition program. This independent instance can then be broken into segments and each erroneous segment in said independent instance replaced with the corrected segment associated with that segment. In this manner, repetitive instruction of a speech recognition program can be facilitated.

Abstract: A system and method for quickly improving the accuracy of a speech recognition program. The system is based on a speech recognition program that automatically converts a pre-recorded audio file into a written text. The system parses the written text into segments, each of which is corrected by the system and saved in an individually retrievable manner in association with the computer. The standard speech files are saved towards improving accuracy in speech-to-text conversion by the speech recognition program. The system further includes facilities to repetitively establish an independent instance of the written text from the prerecorded audio file using the speech recognition program. This independent instance can then be broken into segments and each segment in said independent instance replaced with an individually retrievable saved corrected segment associated with that segment. In this manner, repetitive instruction of a speech recognition program can be facilitated.

Abstract: The present invention relates to a method and apparatus for directing a pre-recorded audio file to a speech recognition program that does not normally accept such files, such as IBM Corporation's Via Voice™ speech recognition program. The method includes: (a) launching the speech recognition program to accept speech as if the speech recognition program were receiving live audio from a microphone; (b) finding a mixer utility associated with the sound card; (c) opening the mixer utility, the mixer utility having settings that determine an input source and an output path; (d) changing the settings of the mixer utility to specify a line-in input source and a wave-out output path; (e) activating a microphone input of the speech recognition software; and (f) initiating a media player associated with the computer to play the pre-recorded audio file into the line-in input source. The method may additionally save and restore the original configuration settings of the mixer utility.

Abstract: A system for substantially automating transcription services for multiple voice users including a manual transcription station, a speech recognition program and a routing program. The system establishes a profile for each of the voice users containing a training status which is selected from the group of enrollment, training, automated and stop automation. When the system receives a voice dictation file from a current voice user based on the training status the system routes the voice dictation file to a manual transcription station and the speech recognition program. A human transcriptionist creates transcribed files for each received voice dictation files. The speech recognition program automatically creates a written text for each received voice dictation file if the training status of the current user is training or automated.