Speech coding apparatus and method for generating acoustic feature vector component values by combining values of the same features for multiple time intervals

• US 5,544,277 A
• Filed: 07/28/1993
• Issued: 08/06/1996
• Est. Priority Date: 07/28/1993
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bakis, Raimo, Nadas, Arthur J., Kanevsky, Dimitri, Nahamoo, David, Sedivy, Jan, Picheny, Michael A., Gopalakrishnan, Ponani S.
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): Doerrler, Michelle
• Application Number:
  US08/098,682
• Time in Patent Office: 1,105 Days
• Field of Search: 381/41-45, 381/36, 395/2-2.31, 395/2.39, 395/2.4, 395/2.52-2.54, 395/2.59, 395/2.64-2.66
• US Class Current: 704/211
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence

Abstract

A speech coding apparatus and method measures the values of at least first and second different features of an utterance during each of a series of successive time intervals. For each time interval, a feature vector signal has a first component value equal to a first weighted combination of the values of only one feature of the utterance for at least two time intervals. The feature vector signal has a second component value equal to a second weighted combination, different from the first weighted combination, of the values of only one feature of the utterance for at least two time intervals. The resulting feature vector signals for a series of successive time intervals form a coded representation of the utterance. In one embodiment, a first weighted mixture signal has a value equal to a first weighted mixture of the values of the features of the utterance during a single time interval. A second weighted mixture signal has a value equal to a second weighted mixture, different from the first weighted mixture, of the values of the features of the utterance during a single time interval. The first component value of each feature vector signal is equal to a first weighted combination of the values of only the first weighted mixture signals for at least two time intervals, and the second component value of each feature vector signal is equal to a second weighted combination, different from the first weighted combination, of the values of only the second weighted mixture for at least two time intervals.

Claims

1. A speech coding apparatus comprising:

means for measuring the values of at least first and second different features of an utterance during each of a series of successive time intervals; and

means for generating a feature vector signal for each time interval, each feature vector signal having a first component value equal to a first weighted combination of the values of only one feature of the utterance for at least two time intervals, each feature vector signal having a second component value equal to a second weighted combination, different from the first weighted combination, of the values of only one feature of the utterance for at least two time intervals, said feature vector signals forming a coded representation of the utterance.

2. A speech coding apparatus as claimed in claim 1, characterized in that the first weighted combination is for the same time intervals as the second weighted combination.

3. A speech coding apparatus as claimed in claim 2, characterized in that each component value of a feature vector signal for a selected time interval is equal to a weighted combination of the values of only one feature of the utterance for the selected time interval and at least one other time interval.

4. A speech coding apparatus as claimed in claim 3, characterized in that each component value of a feature vector signal for a selected time interval is equal to a weighted combination of the values of only one feature of the utterance for the selected time interval, four immediately preceding time intervals and four immediately following time intervals.

5. A speech coding apparatus as claimed in claim 2, characterized in that:

the first component value is equal to the first weighted combination of the values of only the first feature of the utterance for at least two time intervals; and

the second component value is equal to the second weighted combination, different from the first weighted combination, of the values of only the first feature of the utterance for at least two time intervals.

6. A speech coding apparatus as claimed in claim 2, characterized in that:

the first component value is equal to the first weighted combination of the values of only the first feature of the utterance for at least two time intervals; and

the second component value is equal to the second weighted combination, different from the first weighted combination, of the values of only the second feature of the utterance, different from the first feature, for at least two time intervals.

7. A speech coding apparatus as claimed in claim 2, characterized in that:

the feature vector signal generating means further comprises means for generating first and second weighted mixture signals for each time interval, said first weighted mixture signal having a value equal to a first weighted mixture of the values of the features of the utterance during a single time interval, said second weighted mixture signal having a value equal to a second weighted mixture, different from the first weighted mixture, of the values of the features of the utterance during a single time interval;

the first component value of each feature vector signal is equal to a first weighted combination of the values of only the first weighted mixture signals for at least two time intervals; and

the second component value of each feature vector signal is equal to a second weighted combination, different from the first weighted combination, of the values of only the second weighted mixture for at least two time intervals.

8. A speech coding apparatus as claimed in claim 7, characterized in that the first weighted mixture is for the same single time interval as the second weighted mixture.

9. A speech coding apparatus as claimed in claim 8, characterized in that:

each weighted combination is a linear weighted combination; and

each weighted mixture is a linear weighted mixture.

10. A speech coding apparatus as claimed in claim 2, further comprising:

means for storing a plurality of prototype vector signals, each prototype vector signal having at least two parameter values and having a unique identification value;

means for comparing the closeness of the component values of a selected feature vector signal to the parameter values of the prototype vector signals to obtain prototype match scores for the selected feature vector signal and each prototype vector signal; and

means for outputting at least the identification value of at least the prototype vector signal having the best prototype match score as a coded utterance representation signal of the selected feature vector signal.

11. A speech coding apparatus as claimed in claim 2, characterized in that the measuring means comprises a microphone.

12. A speech coding apparatus as claimed in claim 11, characterized in that the measuring means comprises a spectrum analyzer for measuring the amplitudes of the utterance in two or more frequency bands during each of a series of successive time intervals.

13. A speech coding method comprising the steps of:

measuring the values of at least first and second different features of an utterance during each of a series of successive time intervals; and

generating a feature vector signal for each time interval, each feature vector signal having a first component value equal to a first weighted combination of the values of only one feature of the utterance for at least two time intervals, each feature vector signal having a second component value equal to a second weighted combination, different from the first weighted combination, of the values of only one feature of the utterance for at least two time intervals, said feature vector signals forming a coded representation of the utterance.

14. A speech coding method as claimed in claim 13, characterized in that the first weighted combination is for the same time intervals as the second weighted combination.

15. A speech coding method as claimed in claim 14, characterized in that each component value of a feature vector signal for a selected time interval is equal to a weighted combination of the values of only one feature of the utterance for the selected time interval and at least one other time interval.

16. A speech coding method as claimed in claim 15, characterized in that each component value of a feature vector signal for a selected time interval is equal to a weighted combination of the values of only one feature of the utterance for the selected time interval, four immediately preceding time intervals and four immediately following time intervals.

17. A speech coding method as claimed in claim 14, characterized in that:

the first component value is equal to the first weighted combination of the values of only the first feature of the utterance for at least two time intervals; and

the second component value is equal to the second weighted combination, different from the first weighted combination, of the values of only the first feature of the utterance for at least two time intervals.

18. A speech coding method as claimed in claim 14, characterized in that:

the first component value is equal to the first weighted combination of the values of only the first feature of the utterance for at least two time intervals; and

the second component value is equal to the second weighted combination, different from the first weighted combination, of the values of only the second feature of the utterance, different from the first feature, for at least two time intervals.

19. A speech coding method as claimed in claim 14, characterized in that:

the step of generating a feature vector signal further comprises the step of generating first and second weighted mixture signals for each time interval, said first weighted mixture signal having a value equal to a first weighted mixture of the values of the features of the utterance during a single time interval, said second weighted mixture signal having a value equal to a second weighted mixture, different from the first weighted mixture, of the values of the features of the utterance during a single time interval;

the first component value of each feature vector signal is equal to a first weighted combination of the values of only the first weighted mixture signals for at least two time intervals; and

the second component value of each feature vector signal is equal to a second weighted combination, different from the first weighted combination, of the values of only the second weighted mixture for at least two time intervals.

20. A speech coding method as claimed in claim 19, characterized in that the first weighted mixture is for the same single time interval as the second weighted mixture.

21. A speech coding method as claimed in claim 20, characterized in that:

each weighted combination is a linear weighted combination; and

each weighted mixture is a linear weighted mixture.

22. A speech coding method as claimed in claim 14, further comprising the steps of:

storing a plurality of prototype vector signals, each prototype vector signal having at least two parameter values and having a unique identification value;

comparing the closeness of the component values of a selected feature vector signal to the parameter values of the prototype vector signals to obtain prototype match scores for the selected feature vector signal and each prototype vector signal; and

outputting at least the identification value of at least the prototype vector signal having the best prototype match score as a coded utterance representation signal of the selected feature vector signal.

23. A speech coding method as claimed in claim 14, characterized in that the step of measuring comprises measuring the amplitudes of the utterance in two or more frequency bands during each of a series of successive time intervals.