Name synthesis

• US 8,719,027 B2
• Filed: 02/28/2007
• Issued: 05/06/2014
• Est. Priority Date: 02/28/2007
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Chen, Yining, Li, Yusheng, Chu, Min, Soong, Frank Kao-Ping
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): ADESANYA, OLUJIMI A
• Application Number:
  US11/712,298
• Publication Number:
  US 20080208574A1
• Time in Patent Office: 2,624 Days
• Field of Search: 704/260, 704/270
• US Class Current: 704/260
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination

Abstract

An automated method of providing a pronunciation of a word to a remote device is disclosed. The method includes receiving an input indicative of the word to be pronounced. The method further includes searching a database having a plurality of records. Each of the records has an indication of a textual representation and an associated indication of an audible representation. At least one output is provided to the remote device of an audible representation of the word to be pronounced.

Claims

1. A computer-implemented method of providing a pronunciation of a proper name to a remote device, the method comprising:

receiving, with a computer processor, a first textual input indicative of the proper name to be pronounced;

searching a database, with the computer processor, the database having a plurality of records each record having an indication of a textual representation of a proper name and an associated indication of an audible representation, of a proper name; and

identifying a record for a matching proper name, when the textual representation in the record matches the first textual input;

providing at least one output to the remote device of the audible representation of the record identified, for pronunciation of the proper name in the record identified;

receiving a second textual input indicative of a desired pronunciation from a remote device, the second textual input comprising a textual representation for a different word, other than the proper name indicated by the first textual input, the different word having a different spelling than the proper name indicated by the first textual input but a similar pronunciation;

generating a new audible representation from the textual representation of the different word using an automated text to speech engine;

associating, with the computer processor, the new audible representation with the first textual input indicative of the proper name to be pronounced; and

creating a record in the database, with the computer processor, including the proper name to be pronounced and the associated new audible representation.

2. The computer-implemented method of claim 1, wherein searching the database includes comparing the first textual input against the indication of the textual representation of at least one of the plurality of records.

3. The automated method of claim 1, wherein providing at least one output of an audible representation comprises:

retrieving an indication of an audible representation from the database; and

creating an audio representation from the retrieved indication of an audible representation.

4. The automated method of claim 3, wherein the database includes more than one record having an indication of a textual representation that matches the first textual input and further comprising:

retrieving an audible representation from each of the records having a textual representation that matches the first textual input; and

wherein providing at least one output to the remote device of an audible representation includes providing an output of each of the retrieved audible representations.

5. The automated method of claim 4, wherein providing an output of each of the retrieved audible representations includes providing the outputs according to a pre-established priority.

6. A computer-implemented method of providing a database of pronunciation information for use in an automated pronunciation system, the method comprising:

receiving, as an input at a computer processor, a plurality of indications of textual representations of a plurality of proper names for having pronunciations stored in the database;

using an automated text-to-speech synthesizer to automatically generate an indication of an audio representation associated with each of the proper names, the audio representation identifying a pronunciation;

associating, using the computer processor, the indication of an audio representation with the indication of a textual representation for the associated proper name;

storing the associated indications in a record in the database; and

for a given proper name,

retrieving a previously stored record including indications of a textual representation of the given proper name and an audio representation of the given proper name;

providing the audio representation of the given proper name to a remote device, that is remote from the database;

receiving data from the remote device including the indication of the textual representation of the given proper name and a textual representation of a different word having a different spelling than the given proper name;

creating an indication of an audio representation of the different word using the automated text-to-speech synthesizer; and

associating the indication of the audio representation of the different word with the textual representation of the previously stored record for the given proper name.

7. The method of claim 6, and further comprising:

for a given proper name, determining an origin for the given proper name; and

applying a set of pronunciation rules associated with the origin to the textual representation for the given proper name to create the indication of an audio representation.

8. The method of claim 6, wherein receiving the modified indication of the audio representation includes receiving a phoneme string and storing data in the database comprises storing the phoneme string in the database.

9. The method of claim 6, wherein receiving the modified indication of the audio representation includes receiving an audio file for the given proper name.

10. The method of claim 9 and further comprising:

generating a textual representation of the audio file for the given proper name; and

wherein storing the received data includes storing an indication of the textual representation for the given proper name.

11. A system adapted to provide an audible indication of a proper pronunciation of a proper name to a remote device that is remote from the system, the system comprising:

a database having a plurality of records each having a first data element indicative of a textual representation of a proper name and a second data element indicative of an audible representation of the proper name, wherein at least two records of the plurality of records in the database have first data elements indicative of a textual representation of a given proper name to be pronounced and second data elements indicative of different audible representations of the same given proper name to be pronounced, along with a separate metadata element indicative of a priority of each of the different audible representations based on an origin of the given proper name, wherein the at least two records in the database are prioritized using the metadata elements in a first order for a first origin of the given proper name and in a second order, that is different than the first order, for a second origin of the proper name;

a database manager communicating information with the database;

a text to speech engine that receives, as a text input, the textual representation of the given proper name to be pronounced and generates an audible representation of the text input; and

a communication device receiving an input from the remote device over a network indicative of the textual representation of the proper name to be pronounced and an origin indication from the remote device, the communication device providing the remote device an output over the network indicative of the audible representation of the proper name to be pronounced generated by the text to speech engine and prioritized using the origin indication and metadata elements in the database, wherein the communication device and text to speech engine are remote from the remote device.

12. The system of claim 11, wherein the second data element of at least one of the plurality of records includes information relating to a phoneme string.

13. The system of claim 11, wherein the second data element of at least one of the plurality of records includes information relating to an audio file.

14. The system of claim 11, wherein the proper name to be pronounced has multiple different origins, and wherein each of the at least two records includes a third data element that is associated with the first data element of the record and indicates one of the multiple different origins for the proper name.

15. The system of claim 14, wherein at least one of the plurality of records includes a fourth data element having information indicative of a priority of the record.