Computerimplementiertes Verfahren zur Transkription eines Audiostroms und Transkriptionsmechanismus

Details

Veröffentlichungs-Nr.

EP3929916

Aktenzeichen

EP20182036

Anmeldetag

24. Juni 2020

Veröffentlichung

3. April 2024

Erteilung

3. April 2024

Rechtsraum

EP

IPC

G10L15/32

Abstract

The present invention relates to a computer-implemented method of transcribing an audio stream (1), in particular, during an audio- and/or video-conference, using a transcription service, the method comprising the steps of transcribing the audio stream (1) using a first transcribing instance (2), the first transcribing instance (2) having a first predetermined transcription size (s1) that is smaller than the total length of the audio stream (1), the first transcribing instance (2) providing a plurality of consecutive first transcribed text data snippets (t1) of the audio stream (1), the size of the first transcribed text data snippets (t1, t2, t3, tn) respectively corresponding to the first predetermined transcription size (s1); transcribing the audio stream (1) using at least a second transcribing instance (3), the second transcribing instance (3) having a second predetermined transcription size (s2) that is smaller than the length of the audio stream (1), the second transcribing instance (3) providing a plurality of consecutive second transcribed text data snippets (t1', t2', t3', tn') of the audio stream (1), the size (s2) of the second transcribed text data snippets (t1') respectively corresponding to the second predetermined transcription size (s2); wherein the first transcribing instance (2) starts transcription of the audio stream (1) at a first point of time and the second transcribing instance (3) starts transcription of the audio stream (1) at a second point of time with a predetermined delay with respect to the first transcribing instance (2), wherein the predetermined delay is selected such that each one of the plurality of second text data snippets (tl', t2', t3', tn') respectively overlaps at least a portion at which a first transcribed text data snippet (t1) of the plurality of the first text data snippets ends and a consecutive first transcribed text data snippet (t2) of the plurality of the first text data snippets starts. Further, the present invention relates to a transcription mechanism for a communication system for carrying out a video and/or audio conference with at least two participants, wherein the transcription mechanism is adapted to carry out the method of transcribing an audio stream (1).

Anmelder

Firmen

Unify Beteiligungsverwaltung GmbH & Co. KG
Unify Patente GmbH & Co. KG

Land

🇩🇪 Deutschland

Fachgebiete