Use text to speech techniques to improve understanding when announcing search results

• US 8,856,007 B1
• Filed: 10/15/2012
• Issued: 10/07/2014
• Est. Priority Date: 10/09/2012
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Hamilton, Robert William, Stuttle, Matthew Nicholas
• Primary Examiner(s): Opsasnick, Michael N
• Assistant Examiner(s): ADESANYA, OLUJIMI A
• Application Number:
  US13/652,460
• Time in Patent Office: 722 Days
• Field of Search: 704/9, 704/260, 704/275
• US Class Current: 704/260
• CPC Class Codes: G06F 16/957 Browsing optimisation, e.g. caching or content distillation
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G10L 13/033 Voice editing, e.g. manipulating the voice of the synthesiser

Abstract

Disclosed are apparatus and methods for generating synthesized utterances related to output of commands. A command is received at a computing device. A textual output for the command is determined using the computing device. A spoken output of the computing device is generated that utilizes a plurality of vocal characteristic sets. At least a portion of the spoken output corresponds to the textual output. At least a first part of the spoken output utilizes vocal characteristics of a first vocal characteristic set. At least a second part of the spoken output utilizes vocal characteristics of a second vocal characteristic set, where at least some of the vocal characteristics of the first vocal characteristic set differ from the vocal characteristics of the second vocal characteristic set.

Claims

1. A method, comprising:

receiving a command at a computing device;

determining a textual response to the command using the computing device;

providing a graphical user interface of the computing device for specifying spoken output;

specifying one or more vocal characteristics of at least a first voice and a second voice using the graphical user interface, wherein a first vocal characteristic set is associated with the first voice, and a second vocal characteristic set is associated with the second voice, and wherein the first vocal characteristic set differs from the second vocal characteristic set;

generating a specification for spoken output specified using the graphical user interface, wherein the specification comprises a first selection to use the first voice in generating a first part of spoken output related to a first portion of the textual response to the command and a second selection to use the second voice in generating a second part of spoken output related to a second portion of the textual response to the command, and wherein the first part of spoken output differs from the second part of spoken output; and

generating a spoken output of the computing device using the specification for spoken output, wherein the first part of the spoken output utilizes vocal characteristics of the first voice, and wherein the second part of the spoken output utilizes vocal characteristics of the second voice.

2. The method of claim 1, wherein the spoken output comprises a salient portion and a standard portion, wherein the salient portion comprises a portion of the spoken output corresponding to the textual response, wherein the standard portion utilizes the first voice, and wherein the salient portion utilizes at least the second voice.

3. The method of claim 2, wherein the salient portion utilizes at least the second voice and a third voice, and wherein the second voice differs from the third voice.

4. The method of claim 3, wherein the textual response comprises a primary output and secondary output, wherein the salient portion utilizes the second voice for the primary output, and wherein the salient portion utilizes the third voice for the secondary output.

5. The method of claim 1, wherein generating the spoken output comprises:

using a machine learning technique to distinguish a salient portion of the textual response from a standard portion of the textual response; and

generating the spoken output by;

generating a spoken output for the standard portion of the textual response utilizing the first voice, and

generating a spoken output for the salient portion of the textual response utilizing the second voice.

6. The method of claim 1, wherein the command comprises a search query.

7. The method of claim 1, wherein the spoken output is generated utilizing a parametric speech model configured to modify a pitch and a duration in mid-syllable for at least a part of the spoken output.

8. A computing device, comprising:

a sound output device;

a processor; and

a computer-readable storage medium having stored thereon program instructions that, upon execution by the processor, cause the computing device to perform operations comprising;

receiving a command,

determining a textual response to the command,

providing a graphical user interface of the computing device for specifying spoken output,

specifying one or more vocal characteristics of at least a first voice and a second voice using the graphical user interface, wherein a first vocal characteristic set is associated with the first voice, and a second vocal characteristic set is associated with the second voice, and wherein the first vocal characteristic set differs from the second vocal characteristic set,

generating a specification for spoken output specified using the graphical user interface, wherein the specification comprises a first selection to use the first voice in generating a first part of spoken output related to a first portion of the textual response to the command and the second selection to use a second voice in generating a second part of spoken output related to a first portion of the textual response to the command, and wherein the first part of spoken output differs from the second part of spoken output, and

generating a spoken output using the specification for spoken output, wherein the first part of the spoken output utilizes vocal characteristics of the first voice, and wherein the second part of the spoken output utilizes vocal characteristics of the second voice.

9. The computing device of claim 8, wherein the spoken output comprises a salient portion and a standard portion, wherein the salient portion comprises a portion of the spoken output corresponding to the textual response, wherein the standard portion utilizes the first voice, and wherein the salient portion utilizes at least the second voice.

10. The computing device of claim 9, wherein the salient portion utilizes at least the second voice and a third voice, and wherein the second voice differs from the third voice.

11. The computing device of claim 10, wherein the textual response comprises a primary output and secondary output, wherein the salient portion utilizes the second voice for the primary output, and wherein the salient portion utilizes the third voice for the secondary output.

12. The computing device of claim 8, wherein generating the spoken output comprises:

using a machine learning technique to distinguish a salient portion of the textual response from a standard portion of the textual response; and

generating the spoken output by;

generating a spoken output for the standard portion of the textual response utilizing the first voice, and

generating a spoken output for the salient portion of the textual response utilizing the second voice.

13. The computing device of claim 8, wherein the command comprises a search query.

14. The computing device of claim 8, wherein the spoken output is generated utilizing a parametric speech model configured to modify a pitch and a duration in mid-syllable for at least a part of the spoken output.

15. An article of manufacture including a non-transitory computer-readable storage medium having stored thereon program instructions that, upon execution by a computing device, cause the computing device to perform operations comprising:

receiving a command;

determining a textual response to the command;

providing a graphical user interface of the computing device for specifying spoken output;

specifying one or more vocal characteristics of at least a first voice and a second voice using the graphical user interface, wherein a first vocal characteristic set is associated with the first voice, and a second vocal characteristic set is associated with the second voice, and wherein the first vocal characteristic set differs from the second vocal characteristic set;

generating a specification for spoken output specified using the graphical user interface, wherein the specification comprises a first selection to use the first voice in generating a first part of spoken output related to a first portion of the textual response to the command and a second selection to use the second voice in generating a second part of spoken output related to a second portion of the textual response to the command, and wherein the first part of spoken output differs from the second part of spoken output; and

generating a spoken output using the specification for spoken output, wherein the first part of the spoken output utilizes vocal characteristics of the first voice and wherein the second part of the spoken output utilizes vocal characteristics of the second voice.

16. The article of manufacture of claim 15, wherein the spoken output comprises a salient portion and a standard portion, wherein the salient portion comprises a portion of the spoken output corresponding to the textual response, wherein the standard portion utilizes the first voice, and wherein the salient portion utilizes at least the second voice.

17. The article of manufacture of claim 16, wherein the salient portion utilizes at least the second voice and a third voice, and wherein the second voice differs from the third voice.

18. The article of manufacture of claim 15, wherein generating the spoken output comprises:

using a machine learning technique to distinguish a salient portion of the textual response from a standard portion of the textual response; and

generating the spoken output by;

generating a spoken output for the standard portion of the textual response utilizing the first voice, and

generating a spoken output for the salient portion of the textual response utilizing the second voice.

19. The article of manufacture of claim 15, wherein the command comprises a search query.

20. The article of manufacture of claim 15, wherein the spoken output is generated utilizing a parametric speech model configured to modify a pitch and a duration in mid-syllable for at least a part of the spoken output.