System and process for realtime/neartime call analytics with speaker separation

• US 8,855,106 B1
• Filed: 10/05/2011
• Issued: 10/07/2014
• Est. Priority Date: 10/05/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Nachnani, Harkishin, Vasquez, Juan
• Primary Examiner(s): LI, GUANG W
• Application Number:
  US13/253,631
• Time in Patent Office: 1,098 Days
• Field of Search: 370/206, 370/352
• US Class Current: 370/352
• CPC Class Codes: H04L 65/1069 Session establishment or de-establishment
• CPC Class Codes: H04L 65/1083 In-session procedures
• CPC Class Codes: H04L 65/80 Responding to QoS

Abstract

A system, method, and program product for realtime/neartime call analytics, the method comprising: receiving an inbound call connection; capturing a destination network address; initiating placement of a call; capturing first data of the initiating party and converting the first data to first data packets with time stamps and a first designation; creating a first audio stream; sending the first audio stream of audio data packets to an audio analytics service; capturing second data including audio data of the destination party and converting the second data to second data packets with time stamps and a second designation; creating a second audio stream; and sending the second audio stream of audio data packets to the audio analytics service.

Claims

1. A method for realtime/neartime call analytics, comprising:

receiving, by one or more computers via a network, an inbound call connection comprising inbound call data from an initiating party;

capturing a destination network address based on the inbound call data;

connecting, by the one or more computers, the inbound call to the destination network address to permit audio data transmission;

capturing, by the one or more computers, first data of the initiating party including audio data from the inbound call connection and producing from the first data, by the one or more computers via a Voice Over Internet Protocol (VOIP) signaling protocol, first time-stamped data packets including a first designation;

creating, by the one or more computers, a first audio stream comprising audio data packets from the first data packets;

using the first data of the initiating party to identify the initiating party and to determine, based on the identity of the initiating party, to generate a first transcript of the first audio stream;

sending, by the one or more computers responsive to the determination to generate the first transcript, the first audio stream to an audio analytics service to transcribe the first audio stream;

capturing, by the one or more computers from the destination network address, second data including audio data of the destination party and producing, by the one or more computers via the Voice Over Internet Protocol (VOIP) signaling protocol, second time-stamped data packets including a second designation;

creating, by the one or more computers, a second audio stream comprising audio data packets from the second data packets;

using the second data of the destination party to identify the destination party and to determine, based on the identity of the destination party, to generate a second transcript of the second audio stream; and

sending, by the one or more computers responsive to the determination to generate the second transcript, the second audio stream to the audio analytics service to transcribe the second audio stream.

2. The method as defined in claim 1, wherein the inbound call is one selected from the group of an analog audio call, a Voice Over Internet Protocol (VOIP) call, and a video call.

3. The method as defined in claim 1, wherein the capturing a destination network address comprises accessing one or more databases to obtain the destination network address.

4. The method as defined in claim 1, wherein the capturing a destination network address comprises obtaining the destination network address from a data received with the inbound call.

5. The method as defined in claim 1, wherein the signaling protocol comprises one selected from the group of Session Initiation Protocol (SIP), H323 and IMF.

6. The method as defined in claim 1, comprising:

converting or having converted the first data including the audio data of the initiating party for transmission to the destination network address in a different form than was received on the inbound call connection.

7. The method as defined in claim 1, wherein at least one of the inbound call connection and the call connection to the destination network address is on a public switched telephone network (PSTN).

8. The method as defined in claim 1, wherein at least one of the creating a first audio stream step and the creating a second audio stream step comprises using a physical network tap to detect and create the audio stream of audio data packets.

9. The method as defined in claim 1, wherein the creating a first audio stream step comprises a server that received the inbound call connection detecting and creating the first audio stream of audio data packets.

10. The method as defined in claim 1, wherein the creating a second audio stream step comprises a server that initiated the connection of the inbound call to the destination network address detecting and creating the second audio stream of audio data packets.

11. The method as defined in claim 1, further comprising:

the audio analytics service providing a transcription of an audio conversation occurring between the initiating party and a party at the pre-configured destination network address, with accurate speaker separation in realtime/neartime; and

storing the transcription in one or more transcription databases.

12. The method as defined in claim 1, further comprising:

storing a plurality of transcriptions from the audio analytics service in one or more databases; and

performing data mining of one or more of the transcriptions stored in the one or more databases, based at least in part on one or more parameters provided by a requestor.

13. A system for realtime/neartime call analytics, comprising:

a network connection; and

one or more computers configured to;

receive an inbound call connection comprising inbound call data from an initiating party via the network connection;

capture a destination network address based on the inbound call data;

connect the inbound call to the destination network address to permit audio data transmission;

capture first data from the inbound call connection including audio data of the initiating party and producing from the first data, via a Voice Over Internet Protocol (VOIP) signaling protocol, first time-stamped data packets including a first designation;

create a first audio stream comprising audio data packets from the first data packets;

use the first data of the initiating party to identify the initiating party and to determine, based on the identity of the initiating party, to generate a first transcript of the first audio stream;

send, in response to the determination to generate the first transcript, the first audio stream to an audio analytics service to transcribe the first audio stream;

capture, from the destination network address, second data including audio data of the destination party and producing, via the Voice Over Internet Protocol (VOIP) signaling protocol, second time-stamped data packets including a second designation;

create a second audio stream comprising audio data packets from the second data packets;

use the second data of the destination party to identify the destination party and to determine, based on the identity of the destination party, to generate a second transcript of the second audio stream; and

send, in response to the determination to generate the second transcript, the second audio stream to the audio analytics service to transcribe the second audio stream.

14. The system as defined in claim 13, further comprising the one or more computers configured to provide the audio analytics service to obtain a transcript document of an audio conversation occurring between the initiating party and a party at the pre-configured destination network address, with accurate speaker separation in realtime/neartime.

15. The system as defined in claim 13, further comprising:

the one or more computers configured to provide the audio analytics service to obtain a transcription of an audio conversation occurring between the initiating party and a party at the pre-configured destination network address, with accurate speaker separation in realtime/neartime, and to cause storage of the transcription in one or more transcription databases.

16. The system as defined in claim 13, further comprising:

the one or more computers configured to cause storage of a plurality of transcriptions from the audio analytics service in one or more databases; and

configured to perform data mining of one or more of the transcriptions stored in the one or more databases, based at least in part on one or more parameters provided by a requestor.

17. The system as defined in claim 13, further comprising:

one or more databases holding destination network addresses.

18. The system as defined in claim 13, further comprising a physical network tap configured to detect and create at least one selected from the group of the first audio stream of audio data packets and the second audio stream of audio data packets.

19. The system as defined in claim 13, further comprising one or more servers configured to detect and create at least one selected from the group of the first audio stream of audio data packets and the second audio stream of audio data packets.

20. A method for realtime/neartime call analytics, comprising:

receiving, by one or more computers via a network, an inbound call connection comprising inbound call data from an initiating party;

capturing a destination network address based on the inbound call data;

connecting, by the one or more computers, the inbound call to the destination network address to permit audio data transmission;

capturing, by the one or more computers, first data of the initiating party including audio data from the inbound call connection and producing from the first data, by the one or more computers via a Voice Over Internet Protocol (VOIP) signaling protocol, first time-stamped data packets including a first designation;

creating, by the one or more computers, a first audio stream comprising audio data packets from the first data packets;

using the first data of the initiating party to identify the initiating party and to determine, based on the identity of the initiating party, to generate a first transcript of the first audio stream;

sending, by the one or more computers responsive to the determination to generate the first transcript, the first audio stream to an audio analytics service to transcribe the first audio stream;

capturing, by the one or more computers from the destination network address, second data including audio data of the destination party and producing, by the one or more computers via the Voice Over Internet Protocol (VOIP) signaling protocol, second time-stamped data packets including a second designation;

creating, by the one or more computers, a second audio stream comprising audio data packets from the second data packets;

using the second data of the destination party to identify the destination party and to determine, based on the identity of the destination party, to generate a second transcript of the second audio stream; and

sending, by the one or more computers responsive to the determination to generate the second transcript, the second audio stream to the audio analytics service to transcribe the second audio stream;

storing the first and second transcripts in one or more transcription databases; and

performing data mining of at least one of the first and second transcripts based on one or more parameters provided by a requestor.