System and method for real-time multimedia reporting
AssigneeStenotran Services, Inc.
InventorLynda Ruth Johansson
Life of this patent
Estimating the term…
Abstract
The present invention provides a real-time multimedia event reporting system and method that enable reporters to generate accurate reports or contents simultaneously in multiple languages accessible by users from anywhere in any form in real-time as the live event proceeds. In addition, the present invention enables the generation of a multi-language report in which words uttered during the event are represented in the languages that they were spoken. The disclosed system also enhances the real-time performance of the reporting process by enabling dynamic adjustment to the speech transcription operating parameters and by providing real-time editing of transcribed text using configurable event-specific text representations.
Claims7
A method for reporting audio in a plurality of languages, comprising the steps of:
- (a) receiving a sequence of audio;
- (b) substantially simultaneous with receiving said sequence of audio, generating a corresponding audio stream in each of said plurality of languages;
- and
- (c) converting each of said generated audio streams into a text stream in said respective language.
A method for reporting audio in a plurality of languages, comprising the steps of:
- (a) receiving a sequence of audio;
- (b) substantially simultaneous with receiving said sequence of audio, generating a corresponding audio stream in each of said plurality of languages;
- (c) converting each of said generated audio streams into a text stream in said respective language;
- and
- (d) substantially simultaneous with converting said generated audio streams, selectively combining said converted text streams into a combined multi-language text stream representing said sequence of audio in said plurality of languages.
A system for reporting a sequence of audio in a plurality of languages, the system comprising:
- means for simultaneously generating, in each of said plurality of languages, an audio stream corresponding to said sequence of audio;
- a converter for converting each of said generated audio stream into a text stream in said respective language;
- and
- a combiner for selectively combining said converted text streams into a combined multi-language text stream representing said sequence of audio in said plurality of languages.
A computer readable memory having recorded thereon statements and instructions for execution by a computer for reporting a sequence of audio in a plurality of languages, said statements and instructions comprising:
- means for simultaneously generating, in each of said plurality of languages, an audio stream corresponding to said sequence of audio;
- means for converting each of said generated audio stream into a text stream;
- and
- means for selectively combining said converted text streams into a combined multi-language text stream representing said sequence of audio in said plurality of languages.
A method of transcribing audio into text using a speech recognition system, comprising the steps of:
- (a) monitoring an input speed of a sequence of audio for transcription, at least one operating parameter of said speech recognition system, and an output speed of a transcribed text, said at least one operating parameter affecting said output speed of said transcribed text;
- and
- (b) adjusting said at least one operating parameter of said speech recognition system so that said transcribed text is synchronized with said sequence of audio within a threshold delay defined in a number of seconds or a number of words.
A method of incorporating source information about a sequence of audio into a text transcribed therefrom, the method comprising the steps of:
- (a) storing a location representation of a source of said sequence of audio and an identity representation of said source in a non-transitory computer accessible storage medium;
- (b) associating said location representation with said identity representation;
- and
- (c) upon detecting said audio from said source, accessing said location representation of said source thereby causing a processor to incorporate at least a portion of said associated identity representation of said source into said transcribed text.
Dependencies are inferred from the claim text.