Stream server that modifies a stream according to detected characteristics

Patent metadata

Field	Value
Publication number	US-10652396-B2
Application number	US-201816144217-A
Country	US
Kind code	B2
Filing date	Sep 27, 2018
Priority date	Sep 27, 2018
Publication date	May 12, 2020
Grant date	May 12, 2020

Abstract

A stream server manages connections by multiple clients. The stream server receives, sends and manages streams for each client. The stream server analyzes incoming streams and maintains a database of device signatures and a database of user profiles based on the analyzed data. The stream server receives a stream from a client, analyzes the stream to identify an applicable device signature, user profile, or both, and modifies the stream. Modification of the stream may include filtering noise, placing the user in receive-only mode, ending a session for a user, adjusting metadata, etc. In addition, modification of the stream may vary based on the role or authorization of the user, or based on commands from one or more of the users.

First claim

The invention claimed is:

1. An apparatus comprising:
at least one processor;
a memory coupled to the at least one processor;
a caller profile database residing in the memory that includes a plurality of entries, each of the plurality of entries comprising voice characterization data corresponding to a voice portion of a previously-received audio stream for an identified caller;
a device signature database residing in the memory that includes a plurality of entries, each of the plurality of entries comprising a device identifier and corresponding device signature data corresponding to non-voice portions of a previously-received audio stream for an identified device corresponding to the device identifier, wherein the device signature data comprises repetitive sounds, transient sounds, and volume issues; and
a voice stream server residing in the memory and executed by the at least one processor, the voice stream server receiving a plurality of audio streams from a plurality of callers, analyzing the plurality of audio streams for voice data and for non-voice data, identifying a first entry in the caller database that corresponds to a first audio stream of the plurality of audio streams, identifying a first entry in the device signature database that corresponds to the first audio stream, and modifying the first audio stream based on the analyzed voice data and the voice characterization data in the first entry in the caller database and based on the analyzed non-voice data and the device signature data in the first entry in the device signature database.

2. The apparatus of claim 1 further wherein the voice stream server comprises a voice to text transcription tool that transcribes the voice data in the first audio stream to text and compares the text to a key word list, and when specified criteria are satisfied with respect to the occurrence of words in the key word list in the text, the voice stream server modifies voice data in the first audio stream.

3. The apparatus of claim 1 wherein the voice stream server modifies the first audio stream by placing a first caller corresponding to the first audio stream in listen-only mode by not forwarding the voice data in the first audio stream received by the voice stream server.

4. The apparatus of claim 3 wherein the voice stream server provides a voice prompt to the first caller that states the first caller is in listen-only mode.

5. The apparatus of claim 1 wherein the voice stream server modifies the first audio stream by ending a session for a first caller corresponding to the first audio stream.

6. The apparatus of claim 5 wherein the voice stream server provides a voice prompt to the first caller before terminating the session for the first caller that states the session for the first caller is being terminated.

7. The apparatus of claim 1 wherein the voice stream server enables and disables modifying the first audio stream based on:
caller role;
caller authorization; and
verbal commands from at least one caller.

8. The apparatus of claim 1 wherein when the identified caller corresponding to one of the plurality of entries in the caller profile database uses a plurality of devices for a plurality of past calls, the device signature database comprises a plurality of entries corresponding to the plurality of devices used by the identified caller.

9. An apparatus comprising:
at least one processor;
a memory coupled to the at least one processor;
a caller profile database residing in the memory that includes a plurality of entries, each of the plurality of entries comprising voice characterization data corresponding to a voice portion of a previously-received audio stream for an identified caller;
a device signature database residing in the memory that includes a plurality of entries, each of the plurality of entries comprising a device identifier and corresponding device signature data corresponding to non-voice portions of a previously-received audio stream for an identified device corresponding to the device identifier, wherein the device signature data comprises repetitive sounds, transient sounds, and volume issues; and
a voice stream server residing in the memory and executed by the at least one processor, the voice stream server receiving a plurality of audio streams from a plurality of callers, analyzing the plurality of audio streams for voice data and for non-voice data, identifying a first entry in the caller database that corresponds to voice data in a first audio stream of the plurality of audio streams, modifying the first audio stream based on voice characterization data in the first entry in the caller database, identifying a first entry in the device signature database that corresponds to the first audio stream, and modifying the first audio stream based on the device signature data in the first entry in the device signature database, wherein the voice stream server comprises a voice to text transcription tool that transcribes the voice data in the first audio stream to text and compares the text to a key word list, and when specified criteria are satisfied with respect to the occurrence of words in the key word list in the text, the voice stream server modifies the voice data in the first audio stream by placing a first caller corresponding to the first audio stream in listen-only mode by not forwarding the voice data in the first audio stream received by the voice stream server, wherein the voice stream server provides a voice prompt to the first caller that states the first caller is in listen-only mode, wherein the voice stream server enables and disables modifying the first audio stream based on:
caller role;
caller authorization; and
verbal commands from at least one caller.

10. The apparatus of claim 9 wherein when the identified caller corresponding to one of the plurality of entries in the caller profile database uses a plurality of devices for a plurality of past calls, the device signature database comprises a plurality of entries corresponding to the plurality of devices used by the identified caller.

11. A computer-implemented method executed by at least one processor for processing a plurality of audio streams from a plurality of callers, the method comprising:
providing a caller profile database that includes a plurality of entries, each of the plurality of entries comprising voice characterization data corresponding to a voice portion of a previously-received audio stream for an identified caller;
providing a device signature database residing in the memory that includes a plurality of entries, each of the plurality of entries comprising a device identifier and corresponding device signature data corresponding to non-voice portions of a previously-received audio stream for an identified device corresponding to the device identifier, wherein the device signature data comprises repetitive sounds, transient sounds, and volume issues;
receiving the plurality of audio streams from the plurality of callers;
analyzing the plurality of audio streams for voice data and for non-voice data;
identifying a first entry in the caller database that corresponds to a first audio stream of the plurality of audio streams;
identifying a first entry in the device signature database that corresponds to the first audio stream; and
modifying the first audio stream based on the analyzed voice data and the voice characterization data in the first entry in the caller database and based on the analyzed non-voice data and the device signature data in the first entry in the device signature database.

12. The method of claim 11 further comprising:
transcribing the voice data in the first audio stream to t

Assignees

• IBM

Inventors

• Koster David M
• Thorstensen Andrew
• Reznechek Adam D
• Nikolai Jason A

Classifications

• H04M2201/40

  using speech recognition · CPC title

• H04M2250/74

  with voice recognition means · CPC title

• G10L15/07

  to the speaker · CPC title

• H04L65/403 Primary

  Arrangements for multi-party communication, e.g. for conferences (data switching systems for conference H04L12/18; arrangements for connecting several subscribers to a common circuit, i.e. affording conference facilities H04M3/56; television conferencing systems H04N7/15) · CPC title

• H04M3/568 Primary

  audio processing specific to telephonic conferencing, e.g. spatial distribution, mixing of participants (echo suppression in two-way loud-speaking telephone systems H04M9/02; sound field processing per se H04S7/30) · CPC title