Evaluating spoken skills

• US 8,775,184 B2
• Filed: 01/16/2009
• Issued: 07/08/2014
• Est. Priority Date: 01/16/2009
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Deshmukh, Om D., Verma, Ashish
• Primary Examiner(s): ROBERTS, SHAUN A
• Application Number:
  US12/354,849
• Publication Number:
  US 20100185435A1
• Time in Patent Office: 1,999 Days
• Field of Search: 704/251, 704/254, 434/185, 434/156
• US Class Current: 704/254
• CPC Class Codes: G09B 19/04 Speaking with audible presentation of the material to be studied G09B5/04
• CPC Class Codes: G09B 19/06 Foreign languages with audible presentation of material to be studied G09B5/04
• CPC Class Codes: G10L 15/1807 using prosody or stress
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

Techniques for evaluating one or more spoken language skills of a speaker are provided. The techniques include identifying one or more temporal locations of interest in a speech passage spoken by a speaker, computing one or more acoustic parameters, wherein the one or more acoustic parameters capture one or more properties of one or more acoustic-phonetic features of the one or more locations of interest, and combining the one or more acoustic parameters with an output of an automatic speech recognizer to modify an output of a spoken language skill evaluation.

Claims

1. A method for evaluating one or more spoken language skills of a speaker, comprising the steps of:

identifying one or more temporal locations of interest in an output of a spoken language skill evaluation performed on a speech passage spoken by a speaker;

computing one or more acoustic parameters to capture one or more properties of one or more acoustic-phonetic features of the one or more locations of interest, said computing comprising determining one or more spectral differences between the onset of a fricative, a vowel and a stop burst by computing a ratio of energy in a high frequency region of the speech passage to energy in a low frequency region of the speech passage over a predetermined time period via

2. The method of claim 1, wherein identifying one or more temporal locations of interest comprises locating one or more phones of interest.

3. The method of claim 1, wherein identifying one or more temporal locations of interest comprises locating one or more acoustic dissimilarities between a sample speech passage and the speech passage spoken by the speaker.

4. The method of claim 1, wherein identifying one or more temporal locations of interest comprises identifying a spoken utterance from a set of one or more acoustically similar options.

5. The method of claim 1, wherein combining the one or more acoustic parameters with the output of an automatic speech recognizer comprises combining information about an acoustic identity of a speech region obtained from the one or more acoustic parameters with an acoustic identity obtained from a speech recognition system.

6. The method of claim 1, wherein the one or more spoken language skills of a speaker comprise at least one of grammatical skill of the speaker and pronunciation skill of the speaker.

7. The method of claim 1, further comprising using time alignment information from an automatic speech recognition system to improve recognition accuracy.

8. The method of claim 7, further comprising allowing certain bi-directional temporal tolerance to the time alignment information from an automatic speech recognition system.

9. The method of claim 1, wherein the one or more temporal locations of interest comprises one or more regions corresponding to a pre-defined set of one or more impact sounds.

10. The method of claim 1, wherein the one or more temporal locations of interest comprises one or more regions where one or more acoustic dissimilarities are prominent.

11. The method of claim 1, further comprising phonetic profiling of a speaker with a speech disability.

12. A computer program product comprising a tangible computer readable storage memory device having computer readable program code for evaluating one or more spoken language skills of a speaker, said computer program product including:

computer readable program code for identifying one or more temporal locations of interest in an output of a spoken language skill evaluation performed on a speech passage spoken by a speaker;

computer readable program code for computing one or more acoustic parameters to capture one or more properties of one or more acoustic-phonetic features of the one or more locations of interest, said computing comprising determining one or more spectral differences between the onset of a fricative, a vowel and a stop burst by computing a ratio of energy in a high frequency region of the speech passage to energy in a low frequency region of the speech passage over a predetermined time period via

13. The computer program product of claim 12, wherein the computer readable program code for identifying one or more temporal locations of interest comprises computer readable program code for locating one or more phones of interest.

14. The computer program product of claim 12, wherein the computer readable program code for identifying one or more temporal locations of interest comprises computer readable program code for locating one or more acoustic dissimilarities between a sample speech passage and the speech passage spoken by the speaker.

15. The computer program product of claim 12, wherein the one or more spoken language skills of a speaker comprise at least one of grammatical skill of the speaker and pronunciation skill of the speaker.

16. The computer program product of claim 12, further comprising computer readable program code for using time alignment information from an automatic speech recognition system to improve recognition accuracy.

17. A system for evaluating one or more spoken language skills of a speaker, comprising:

a memory; and

at least one processor coupled to said memory and operative to;

identify one or more temporal locations of interest in an output of a spoken language skill evaluation performed on a speech passage spoken by a speaker;

compute one or more acoustic parameters to capture one or more properties of one or more acoustic-phonetic features of the one or more locations of interest, said computing comprising determining one or more spectral differences between the onset of a fricative, a vowel and a stop burst by computing a ratio of energy in a high frequency region of the speech passage to energy in a low frequency region of the speech passage over a predetermined time period via

18. The system of claim 17, wherein in identifying one or more temporal locations of interest the at least one processor coupled to said memory is further operative to locate one or more phones of interest.

19. The system of claim 17, wherein in identifying one or more temporal locations of interest the at least one processor coupled to said memory is further operative to locate one or more acoustic dissimilarities between a sample speech passage and the speech passage spoken by the speaker.

20. The system of claim 17, wherein the one or more spoken language skills of a speaker comprise at least one of grammatical skill of the speaker and pronunciation skill of the speaker.

21. The system of claim 17, wherein the at least one processor coupled to said memory is further operative to use time alignment information from an automatic speech recognition system to improve recognition accuracy.

22. An apparatus for evaluating one or more spoken language skills of a speaker, said apparatus comprising:

means for identifying one or more temporal locations of interest in an output of a spoken language skill evaluation performed on a speech passage spoken by a speaker;

means for computing one or more acoustic parameters to capture one or more properties of one or more acoustic-phonetic features of the one or more locations of interest, said computing comprising determining one or more spectral differences between the onset of a fricative, a vowel and a stop burst by computing a ratio of energy in a high frequency region of the speech passage to energy in a low frequency region of the speech passage over a predetermined time period via