System And Method For Synthesizing Human Speech

• US 20170125009A1
• Filed: 10/28/2016
• Published: 05/04/2017
• Est. Priority Date: 10/28/2015
• Status: Active Grant

• Current Assignee: United Arab Emirates University (Al Hikma Development Co. PJSC)
• Original Assignee: United Arab Emirates University (Al Hikma Development Co. PJSC)
• Inventors: Al Ahmad, Mahmoud F. Y.
• Granted Patent: US 10,134,384 B2
• Time in Patent Office: Days
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 13/04 Details of speech synthesis systems, e.g. synthesiser structure or memory management
• CPC Class Codes: G10L 15/24 Speech recognition using non-acoustical features
• CPC Class Codes: H04R 1/46 Special adaptations for use as contact microphones, e.g. on musical instrument, on stethoscope throat mountings H04R1/14
• CPC Class Codes: H04R 17/02 Microphones
• CPC Class Codes: H04R 3/005 for combining the signals of two or more microphones specially adapted for hearing aids H04R25/407

Abstract

A method and apparatus are described for detecting voice related vibration in the upper region of the chest and synthesizing human speech. The innovation finds its use in speech rehabilitation applications among others, specifically in speech impairments and speech disability arising due to accident, congenital defects or other reasons. A set of piezoelectric based sensors are placed on an upper region of the chest atop or near sound tendons. The sensors pick up the vibrations in the sound tendons and convert the vibrations into electrical output signals. These signals are filtered, amplified and processed using the signal recognition unit. Subsequently, a set of parameters are extracted and used to generate speech or a written text. The sensors incorporate piezoelectric or other transducing materials. These sensors are externally affixed to a human body surface corresponding to the position of the sounds tendons in the upper chest/neck region.

Claims

1. A method for synthesizing speech using piezoelectric material in contact with a chest of a subject, the method comprising:

obtaining a piezoelectric electrical signal from the piezoelectric material, the piezoelectric electrical signal based on mechanical movement of the chest, the mechanical movement related to movement of sound tendons of the subject during an act of speaking;

communicating the piezoelectric electrical signal to a signal processor;

manipulating the piezoelectric electrical signal in the signal processor using signal processing techniques;

extracting from the manipulated piezoelectric electrical signal a set of speech identifying parameters; and

generating a speech signal using the set of speech identifying parameters, the speech signal corresponding to the piezoelectric electrical signal.

2. The method of claim 1, the method further comprising conveying the generated speech signal using a speaker.

3. The method according to claim 1, wherein manipulating the piezoelectric electrical signal comprises:

segmenting the piezoelectric signal into a plurality of frames and identifying from at least one of the plurality of frames at least one parameter of the set of speech identifying parameters.

4. The method according to claim 1, the method further comprising powering the signal processor using electricity generated from movement of the chest due to mechanical movement related to movement of the sound tendons and other physiological activity.

5. The method according to claim 1, the method further comprising:

obtaining an audio electrical signal corresponding to the movement of the sound tendons of the subject during the act of speaking;

mapping the audio electrical signal to the piezoelectric signal; and

storing the mapped piezoelectric signal in a library for comparison with the set of speech identifying parameter.

6. The method according to claim 5, wherein the mapping of the piezoelectric electric signal to the audio electrical signal is performed using a linear one-to-one mapping technique.

7. The method according to claim 5, the method further comprising storing the library on a memory storage device, wherein the memory storage device is accessible by the processor.

8. The method according to claim 5, the method further comprising:

selecting the mapped piezoelectric signal when a positive correlation is established between the mapped piezoelectric signal and at least one parameter of the set of speech identifying parameters, wherein the selected mapped piezoelectric signal forming part of the speech signal.

9. The method according to claim 1, the method further comprising filtering and amplifying the piezoelectric signal before processing.

10. The method according to claim 1, wherein the step of manipulating the piezoelectric signal using signal processing techniques comprises:

wirelessly transmitting the piezoelectric signal using a transmitter; and

receiving the transmitted piezoelectric signal using a receiver located at a location away from the transmitter,

wherein the step of extracting from the manipulated piezoelectric signal the set of speech identifying parameter is performed at the location of the receiver.

11. The method according to claim 10, wherein the method further comprises comparing the piezoelectric signal to a pre-determined threshold before transmitting it by the transmitter and amplifying the piezoelectric signal before transmitting it if the piezoelectric signal is determined to be below the pre-determined threshold.

12. The method according to claim 11, wherein the method further comprises comparing the transmitted piezoelectric signal after receiving it by the receiver to a second pre-determined threshold and amplifying the transmitted piezoelectric signal after receiving it by the receiver if the transmitted piezoelectric signal is determined to be below the second pre-determined threshold.

13. A system for synthesizing human speech, the system comprising:

piezoelectric based sensors for placement in contact with a chest of a subject, the piezoelectric based sensors sensing mechanical movement of the chest, the mechanical movement related to movement of sound tendons of the subject during an act of speaking;

the piezoelectric based sensors generating a piezoelectric signal corresponding to the mechanical movement of the chest; and

a processor for receiving the piezoelectric signal generated from the piezoelectric based sensors, the processor configured to manipulate the received piezoelectric signal generated, extract a set of speech identifying parameters and generate a speech signal using the set of speech identifying parameter, the speech signal corresponding to the piezoelectric signal.

14. The system according to claim 13, the system further comprising a power storage unit for storing transduced electricity from piezoelectric material, the transduced electricity transduced from mechanical movement on the chest of the subject due to movement of the sound tendons and other physiological activity, the power storage unit powering the processor.

15. The system according to claim 14, wherein the physiological activity comprises one or more of heart activity and lungs activity.

16. The system according to claim 13, the system further comprising at least one speaker for conveying the speech signal.

17. The system according to claim 16, wherein the at least one speaker comprises a first speaker proximate to the processor and a second speaker away from the processor and wherein the speech signal is conveyed by the first speaker and the second speaker simultaneously.

18. The system according to claim 13, the system further comprising at least one transmitter for wirelessly transmitting the piezoelectric signal to at least one receiver located at a location away from the at least one transmitter and wherein the processor is located at the same location as the at least one receiver.

19. The system according to claim 18, the system further comprising a first conditional circuit for filtering and amplifying the piezoelectric signal before transmitting it to the at least one receiver and a second conditional circuit for filtering and amplifying the received transmitted piezoelectric signal before conveying it to the processor.

20. A processor for synthesizing human speech, the processor configured to:

receive piezoelectric signals generated based on mechanical movement of a chest portion of a subject, the mechanical movement related to movement of sound tendons of the subject during an act of speaking;

manipulate the received piezoelectric signal;

extract from the manipulated piezoelectric signal a set of speech identifying parameter; and

generate a speech signal using the set of speech identifying parameter, the speech signal corresponding to the piezoelectric signal.