Baseband modem for speech recognition and mobile communication terminal using the same

• US 7,593,853 B2
• Filed: 09/07/2005
• Issued: 09/22/2009
• Est. Priority Date: 09/07/2004
• Status: Expired due to Fees

• Current Assignee: LG Electronics, Inc. (LG Corporation)
• Original Assignee: LG Electronics, Inc. (LG Corporation)
• Inventors: Kim, Chan Woo
• Primary Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US11/221,463
• Publication Number:
  US 20060053011A1
• Time in Patent Office: 1,476 Days
• Field of Search: 704/251
• US Class Current: 704/251
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00

Abstract

A baseband modem and method for voice recognition and a mobile communication terminal using the baseband modem and method are disclosed. A speech recognition rate may be increased by selecting a sampling rate suitable for speech recognition and portions of the speech recognition process may be implemented in hardware. The present invention includes an audio codec modulating a received voice signal using either a sampling rate for speech recognition or a sampling rate for voice communication. A feature vector extraction block extracts one or more feature vectors from the modulated voice signal and a speech recognition block performs speech recognition using an extracted feature vector when the voice signal is determined as a voice command. A vocoder vocodes an output of the audio codec when the voice signal is determined as voice communication.

Claims

1. A baseband modem comprising:

an audio codec to modulate a voice signal using a first sampling rate or a second sampling rate;

means for speech recognition;

means for speech encoding, wherein the audio codec is to encode the voice signal using the first sampling rate and the speech recognition means is for performing speech recognition of the encoded voice signal, if the voice signal is a voice command, and the audio codec is to encode the voice signal using the second sampling rate and the speech encoding means is for performing vocoding of the encoded voice signal, if the voice signal is voice communication, and wherein the means for speech recognition comprises;

a feature vector extraction block to extract at least one feature vector from the encoded voice signal and

a speech recognition block to perform speech recognition using the at least one feature vector extracted by the feature vector extraction block; and

a controller to determine whether the voice signal is a voice command or a voice communication and to power on registers of the feature vector extraction block and speech recognition block, if the voice signal is a voice command, and to power off registers of the feature vector extraction block and speech recognition block, if the voice signal is a voice communication.

2. The baseband modem of claim 1, wherein the speech recognition means further comprises a buffer to store the encoded voice signal.

3. The baseband modem of claim 2, wherein the feature vector extraction block is further to extract the at least one feature vector from data stored in the buffer.

4. The baseband modem of claim 2, wherein the buffer comprises a ping-pong buffer.

5. The baseband modem of claim 1, wherein the speech recognition means further comprises a buffer to store the at least one feature vector extracted from the encoded voice signal.

6. The baseband modem of claim 1, wherein the feature vector extraction block is implemented in hardware.

7. The baseband modem of claim 1, wherein the feature vector extraction block is implemented in software.

8. The baseband modem of claim 1, wherein the speech encoding means further comprises a vocoder to vocode the encoded voice signal.

9. The baseband modem of claim 1, wherein the controller is further to determine the sampling rate used by the audio codec according to the determination.

10. The baseband modem of claim 1, wherein the first sampling rate is optimized for speech recognition.

11. The baseband modem of claim 10, wherein the first sampling rate is in a range of approximately 12 kHz to approximately 32 kHz.

12. The baseband modem of claim 11, wherein the first sampling rate is approximately 16 kHz.

13. The baseband modem of claim 1, wherein the second sampling rate is optimized for voice communication.

14. The baseband modem of claim 13, wherein the second sampling rate is approximately 8 kHz.

15. The baseband modem of claim 1, wherein the audio codec is further to perform pulse code modulation on the voice signal.

16. The baseband modem of claim 1, wherein the baseband modem is implemented in a mobile communication terminal.

17. A mobile communication terminal comprising:

an audio codec to modulate a voice signal using a first sampling rate or a second sampling rate;

a feature vector extraction block to extract at least one feature vector from the modulated voice signal;

a speech recognition block to perform speech recognition using the at least one feature vector extracted by the feature vector extraction block;

a vocoder to vocode the modulated voice signal, wherein the audio codec is to encode the voice signal using the first sampling rate, if the voice signal is a voice command and the audio codec is to encode, the voice signal using the second sampling rate, if the voice signal is voice communication; and

a controller to determine whether the voice signal is a voice command or a voice communication and to power on registers of the feature vector extraction block and speech recognition block, if the voice signal is a voice command, and to power off registers of the feature vector extraction block and the speech recognition block, if the voice signal is a voice communication.

18. The mobile communication terminal of claim 17, further comprising a buffer to store the encoded voice signal.

19. The mobile communication terminal of claim 18, wherein the buffer comprises a ping-pong buffer.

20. The mobile communication terminal of claim 17, further comprising a buffer to store the at least one feature vector extracted from the modulated voice signal.

21. The mobile communication terminal of claim 17, wherein the controller is further to determine whether the voice signal is a voice command or a voice communication comprises receiving a user selection.

22. The mobile communication terminal of claim 21, wherein the controller is further to determine the sampling rate used by the audio codec according to the determination.

23. The mobile communication terminal of claim 17, wherein the first sampling rate is optimized for speech recognition.

24. The mobile communication terminal of claim 23, wherein the first sampling rate is in a range of approximately 12 kHz to approximately 32 kHz.

25. The mobile communication terminal of claim 23, wherein the first sampling rate is approximately 16 kHz.

26. The mobile communication terminal of claim 17, wherein the second sampling rate is optimized for voice communication.

27. The mobile communication terminal of claim 26, wherein the second sampling rate is approximately 8 kHz.

28. The mobile communication terminal of claim 26, wherein the feature vector extraction block is implemented in hardware.

29. The mobile communication terminal of claim 17, wherein the feature vector extraction block is implemented in software.

30. The mobile communication terminal of claim 17, wherein the audio codec is further to perform pulse code modulation on the voice signal.

31. A method of performing speech recognition and speech communication in a baseband modem, the method comprising:

determining whether a voice signal is a voice command or a voice communication with a controller;

modulating the voice signal with an audio codec using a first sampling rate and performing speech recognition of the modulated voice signals if the voice signal is determined to be a voice command, and modulating the voice signal using a second sampling rate and performing vocoding of the modulated voice signals, if the voice signal is determined to be voice communication, and

controlling activation of a feature vector extraction block and a speech recognition block with the controller by powering on registers of the feature vector extraction block and the speech recognition block, if the voice signal is a voice command, and powering off registers of the feature vector extraction block and the speech recognition block, if the voice signal is voice communication.

32. The method of claim 31, wherein performing speech recognition of the modulated voice signal comprises:

extracting at least one feature vector from the modulated voice signal; and

performing speech recognition using the at least one feature vector.

33. The method of claim 31, wherein performing speech recognition of the modulated voice signal farther comprises:

storing the modulated voice signal in a buffer; and

extracting the at least one feature vector from data stored in the buffer.

34. The method of claim 32, wherein performing speech recognition of the modulated voice signal further comprises storing the at least one feature vector extracted from the modulated voice signal in a buffer.

35. The method of claim 32, wherein extracting the at least one feature vector from the modulated voice signal is implemented in hardware.

36. The method of claim 32, wherein extracting the at least one feature vector from the modulated voice signal is implemented in software.

37. The method of claim 31, wherein determining whether the voice signal is a voice command or a voice communication comprises receiving a user selection.

38. The method of claim 31, wherein the first sampling rate optimized for speech recognition.

39. The method of claim 38, wherein the first sampling rate in a range of approximately 12 kHz to approximately 32 kHz.

40. The method of claim 38, wherein the first sampling rate of approximately 16 kHz.

41. The method of claim 31, wherein the second sampling rate is optimized for voice communication.

42. The method of claim 41, wherein the voice signal using a second sampling rate is approximately 8 kHz.

43. The method of claim 31, further comprising performing pulse code modulation on the voice signal.

44. The method of claim 31, wherein the baseband modem is implemented in a mobile communication terminal.