IP Library Granted Patent US 10,140,089
Granted Patent B1
US 10,140,089 · App. 15/672,808 · Granted Nov 27, 2018

Synthetic speech for in vehicle communication

View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 10,140,089
App. No.
15/672,808
Granted
Nov 27, 2018
Kind
B1
Abstract

A system and method that enhances spoken utterances by capturing one or more microphone signals. The system and method estimates a plurality of echo paths from each of the one or more microphone signals and synthesizes a speech reinforcement signal in response to and corresponding to the one or more microphone signals. The system and method concatenates portions of the synthesized reinforcement signal with the captured microphone signals and processes the captured microphone signals in response to the estimated plurality of echo paths.

Claims (40)

1. A method that enhances voice through synthetic speech reinforcement comprising:

capturing one or more microphone signals;

estimating a plurality of echo paths from each of the one or more microphone signals;

synthesizing a speech reinforcement signal in response to and corresponding to speech detected in the one or more microphone signals, wherein synthesizing the speech reinforcement signal includes inferring the speech reinforcement signal from a linguistic context of the captured one or more microphone signals;

concatenating portions of the synthesized speech reinforcement signal with the captured microphone signals to render a reinforcement signal; and

processing the captured microphone signals in response to the estimated plurality of echo paths by subtracting the echo contributions of each of the plurality of echo paths from the captured microphone signals.

2. The method of claim 1 further comprising inferring the speech reinforcement signal from the context of the captured microphone signals that are concealed by noise.

3. The method of claim 2 where the act of inferring the speech reinforcement signal comprises a probabilistic process.

4. The method of claim 2 where the act of inferring the speech reinforcement signal applies a linguistic knowledge pre-stored in a memory.

5. The method of claim 1 where the synthesized speech reinforcement signal contains only valid speech sounds.

6. The method of claim 1 where act of synthesizing a speech reinforcement signal is constrained to a single active talker.

7. The method of claim 1 where the act of concatenating portions of the synthesized reinforcement signal is executed by a plurality of bandpass filters.

8. A non-transitory machine-readable medium encoded with machine-executable instructions, wherein execution of the machine-executable instructions is for:

capturing one or more microphone signals;

estimating a plurality of echo paths from each of the one or more microphone signals;

synthesizing a speech reinforcement signal in response to and corresponding to speech detected in the one or more microphone signals, wherein synthesizing the speech reinforcement signal includes inferring the speech reinforcement signal from a linguistic context of the captured one or more microphone signals;

concatenating portions of the speech detected in the microphone signals with the synthesized reinforcement signal; and

processing the captured microphone signals in response to the estimated plurality of echo paths by subtracting the echo contributions of each of the plurality of echo paths from the captured microphone signals.

9. The non-transitory machine-readable medium of claim 8 further comprising machine-executable instructions is for inferring a portion of the speech reinforcement signal from the context of the captured microphone signals concealed by a noise.

10. The non-transitory machine-readable medium of claim 9 where the act of inferring the speech reinforcement signal comprises a probabilistic process.

11. The non-transitory machine-readable medium of claim 9 where the act of inferring the speech reinforcement signal applies a linguistic knowledge pre-stored in a memory.

12. The non-transitory machine-readable medium of claim 8 where the synthesized speech reinforcement signal contains only valid speech sounds.

13. The non-transitory machine-readable medium of claim 8 where act of synthesizing a speech reinforcement signal is constrained to a single active talker.

14. The non-transitory machine-readable medium of claim 8 where the act of concatenating portions of the synthesized reinforcement signal is executed by a plurality of bandpass filters.

15. A system that enhances voice through reinforcement comprising:

a plurality of microphones capturing one or more microphone signals;

a processor programmed to estimate a plurality of echo paths from each of the one or more microphone signals;

the processor further programmed to synthesize a speech reinforcement signal in response to and corresponding to speech detected in the one or more microphone signals, wherein synthesizing the speech reinforcement signal includes inferring the speech reinforcement signal from a linguistic context of the captured one or more microphone signals;

the processor further programmed to concatenate portions of the synthesized reinforcement signal with the speech detected in the captured microphone signals; and

the processor further programmed to process the captured microphone signals in response to the estimated plurality of echo paths by subtracting the echo contributions of each of the plurality of echo paths from the captured microphone signals.

16. The system of claim 15 where the processor comprises a neural network.

17. The system of claim 15 where the processor comprises a speech synthesizer.

18. The system of claim 15 where the processor comprises a speech generator.

19. A vehicle that enhances voice through synthetic speech reinforcement comprising:

a plurality of microphones within the vehicle capturing one or more microphone signals;

a processor within the vehicle programmed to estimate a plurality of echo paths from each of the one or more microphone signals;

the processor further programmed to synthesize a speech reinforcement signal in response to and corresponding to speech detected in the one or more microphone signals, wherein synthesizing the speech reinforcement signal includes inferring the speech reinforcement signal from a linguistic context of the captured one or more microphone signals;

the processor further programmed to concatenate portions of the synthesized reinforcement signal with the speech detected in the captured microphone signals; and

the processor further programmed to process the captured microphone signals in response to the estimated plurality of echo paths by subtracting the echo contributions of each of the plurality of echo paths from the captured microphone signals.

20. The vehicle of claim 19 where the processor comprises a neural network.

Assignments (4)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded May 22, 2020
From: 2236008 ONTARIO INC.
To: BLACKBERRY LIMITED
Reel/Frame 053313/0315 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 13, 2017
From: QNX SOFTWARE SYSTEMS LIMITED
To: 2236008 ONTARIO INC.
Reel/Frame 043575/0468 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 10, 2017
From: HETHERINGTON, PHILLIP ALAN; PARANJPE, SHREYAS
To: QNX SOFTWARE SYSTEMS LIMITED
Reel/Frame 043258/0471 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 10, 2017
From: LAYTON, LEONARD CHARLES
To: BLACKBERRY LIMITED
Reel/Frame 043259/0456 →
Cited By (6)
US 12,210,401 US 12,249,189 US 12,260,872 US 12,443,387 US 12,497,055 US 12,518,570