Systems and methods for noise reduction using speech recognition and speech synthesis

Patent metadata

Field	Value
Publication number	US-9390725-B2
Application number	US-201514788869-A
Country	US
Kind code	B2
Filing date	Jul 1, 2015
Priority date	Aug 26, 2014
Publication date	Jul 12, 2016
Grant date	Jul 12, 2016

Abstract

The present disclosure describes a system ( 100 ) for reducing background noise from a speech audio signal generated by a user. The system ( 100 ) includes a user device ( 102 ) receiving the speech audio signal, a noise reduction device ( 118 ) in communication with a stored data repository ( 208 ), where the noise reduction device is configured to convert the speech audio signal to text; generate synthetic speech based on the converted text; optionally determine the user as an actual subscriber based on a comparison between the speech audio signal with the synthetic speech; and selectively transmit the speech audio signal or the synthetic speech based on comparison between the predicted subjective quality of the recorded speech and the synthetic speech.

First claim

I claim the following invention:

1. A system using a user device in communication with a stored data repository, that reduces the background noise from a speech audio signal generated by a user, comprising:
a user device, with a processor and a memory, receiving a speech audio signal; and
a noise reduction device, in communication with a stored data repository, and in communication with said user device, is configured to:
convert said received speech audio signal to text;
generate synthetic speech based on a speech data corpus or speech model data of the user stored in said stored data repository and said converted text;
determine the predicted subjective quality of the received speech audio signal if that signal were to be transmitted to a far end listener;
determine the predicted subjective quality of said synthetic speech; and
transmit, selectively, said speech audio signal or said synthetic speech, whichever has higher predicted quality based on a comparison between the value of objective quality metrics computed for the speech audio signal and the synthetic speech signal.

2. The claim according to claim 1 , wherein said stored data repository is on said user device and or a server via a network.

3. The claim according to claim 1 , wherein said received speech audio signal is a live speech audio signal.

4. The claim according to claim 1 , wherein said user device is configured to pre-process said speech audio signal based on using a predetermined noise reduction algorithm.

5. The claim according to claim 1 , wherein said noise reduction device is integrated with said user device.

6. A method to manufacture a system using a user device in communication with a stored data repository, that reduces the background noise from a speech audio signal generated by a user, comprising:
providing a user device, with a processor and a memory, receiving a speech audio signal; and
providing a noise reduction device, in communication with a stored data repository, and in communication with said user device, is configured to:
convert said received speech audio signal to text;
generate synthetic speech based on a speech data corpus or speech model data of the user stored in said stored data repository and said converted text;
determine the predicted subjective quality of the received speech audio signal if that signal were to be transmitted to a far end listener;
determine the predicted subjective quality of said synthetic speech; and
transmit, selectively, said speech audio signal or said synthetic speech, whichever has higher predicted quality based on a comparison between the value of objective quality metrics computed for the speech audio signal and the synthetic speech signal.

7. The claim according to claim 6 wherein said stored data repository is on said user device and or a server via a network.

8. The claim according to claim 6 , wherein said received speech audio signal is a live speech audio signal.

9. The claim according to claim 6 , wherein said step of receiving said speech audio signal by said user device further comprises pre-processing said speech audio signal based on using a predetermined noise reduction algorithm.

10. The claim according to claim 6 , wherein said noise reduction device is integrated with said user device.

11. A method to use a system using a user device in communication with a stored data repository, that reduces the background noise from a speech audio signal generated by a user, comprising:
receiving a speech audio signal with a user device, said user device further comprises a processor and a memory; and
providing a noise reduction device, in communication with a stored data repository, and in communication with said user device, is configured to:
convert said received speech audio signal to text;
generate synthetic speech based on a speech data corpus or speech model data of the user stored in said stored data repository and said converted text;
determine the predicted subjective quality of the received speech audio signal if that signal were to be transmitted to a far end listener;
determine the predicted subjective quality of said synthetic speech; and
transmit, selectively, said speech audio signal or said synthetic speech, whichever has higher predicted quality based on a comparison between the value of objective quality metrics computed for the speech audio signal and the synthetic speech signal.

12. The claim according to claim 11 , wherein said stored data repository is on said user device and or a server via a network.

13. The claim according to claim 11 , wherein said received speech audio signal is a live speech audio signal.

14. The claim according to claim 11 , wherein said step of receiving said speech audio signal further comprises pre-processing said speech audio signal based on using a predetermined noise reduction algorithm.

15. The claim according to claim 11 , wherein said noise reduction device is integrated with said user device.

16. A non-transitory program storage device readable by a computing device that tangibly embodies a program of instructions executable by the computing device to perform a method to use a system using a user device in communication with a stored data repository, that reduces the background noise from a speech audio signal generated by a user, comprising:
receiving a speech audio signal with a user device, said user device further comprises a processor and a memory; and
providing a noise reduction device, in communication with a stored data repository, and in communication with said user device, is configured to:
convert said received speech audio signal to text;
generate synthetic speech based on a speech data corpus or speech model data of the user stored in said stored data repository and said converted text;
determine the predicted subjective quality of the received speech audio signal if that signal were to be transmitted to a far end listener;
determine the predicted subjective quality of said synthetic speech; and
transmit, selectively, said speech audio signal or said synthetic speech, whichever has higher predicted quality based on a comparison between the value of objective quality metrics computed for the speech audio signal and the synthetic speech signal.

17. The claim according to claim 16 , wherein said stored data repository is on said user device and or a server via a network.

18. The claim according to claim 16 , wherein said received speech audio signal is a live speech audio signal.

19. The claim according to claim 16 , wherein said step of receiving said speech audio signal further comprises pre-processing said speech audio signal based on using a predetermined noise reduction algorithm.

20. The claim according to claim 16 , wherein said noise reduction device is integrated with said user device.

Assignees

• Clearone Inc

Inventors

• Graham Derek

Classifications

• G10L25/60

  for measuring the quality of voice signals · CPC title

• G10L21/0264

  characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques · CPC title

• G10L13/027

  Concept to speech synthesisers; Generation of natural phrases from machine-based concepts (generation of parameters for speech synthesis out of text G10L13/08) · CPC title

• G10L15/26

  Speech to text systems (G10L15/08 takes precedence) · CPC title

• G10L21/0208 Primary

  Noise filtering · CPC title