TRANSITION FROM LOW POWER ALWAYS LISTENING MODE TO HIGH POWER SPEECH RECOGNITION MODE

• US 20150221307A1
• Filed: 12/20/2013
• Published: 08/06/2015
• Est. Priority Date: 12/20/2013
• Status: Abandoned Application

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Shah, Saurin, Tharappel, Francis M., Dadu, Saurabh, Bossart, Pierre-Louis, Worrell, Devon, Gamsaragan, Edward V., Nallasellan, Singaravelan, Falik, Ohad, Peebler, Bryan R., Le Hin, Ivan, Ughreja, Rakesh A., Joshi, Mandar S.
• Application Number:
  US14/360,072
• Publication Number:
  US 20150221307A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 1/325 Power saving in peripheral device
• CPC Class Codes: G06F 1/3293 by switching to a less power-consuming processor, e.g. sub-CPU
• CPC Class Codes: G06F 3/165 Management of the audio stream, e.g. setting of volume, audio stream path
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• CPC Class Codes: G10L 15/285 Memory allocation or algorithm optimisation to reduce hardware requirements
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management
• CPC Class Codes: Y02D 30/50 in wire-line communication networks, e.g. low power modes or reduced link rate

Abstract

Disclosed are embodiments for seamless, single-step, and speech-triggered transition of a host processor and/or computing device from a low functionality mode to a high functionality mode in which full vocabulary speech recognition can be accomplished. First audio samples are captured by a low power audio processor while the host processor is in a low functionality mode. The low power audio processor may identify a predetermined audio pattern. The low power audio processor, upon identifying the predetermined audio pattern, triggers the host processor to transition to a high functionality mode. An end portion of the first audio samples that follow an end-point of the predetermined audio pattern may be stored in system memory accessible by the host processor. Second audio samples are captured and stored with the end portion of the first audio samples. Once the host processor transitions to a high functionality mode, multi-channel full vocabulary speech recognition can be performed and functions can be executed based on detected speech interaction phrases.

Claims

1-26. -26. (canceled)

27. A computing system that transitions from a low functionality always listening mode to a higher functionality speech recognition mode, comprising:

a host processor having a low functionality mode and a high functionality mode;

a buffer to store audio samples;

a low power audio processor to capture first audio samples from an audio signal received through a microphone while the host processor is in the low functionality mode and to store the first audio samples in the buffer, wherein the low power audio processor is configured to identify a predetermined audio pattern in the first audio samples, including an end-point of the predetermined audio pattern, and to trigger the host processor to transition to the high functionality mode,

wherein the system is configured to, upon the low power audio processor triggering the host processor, capture second audio samples from audio signals received through one or more microphones and store the second audio samples, and

wherein the host processor is configured to, in the high functionality mode, perform speech recognition processing on at least a portion of the first audio samples in the buffer that follow the end-point of the predetermined audio pattern and on the second audio samples.

28. The system of claim 27, further comprising one or more onboard microphones each configured to receive an audio signal, wherein the one or more onboard microphones include the microphone and the one or more microphones.

29. The system of claim 27, wherein the second audio samples are stored in the buffer following the end-point of the predetermined audio pattern.

30. The system of claim 27, wherein the buffer comprises a first buffer to store audio samples captured while the host processor is in the low functionality mode, and wherein the system further comprises:

a second buffer accessible to the host processor to store audio samples,

wherein the second audio samples are stored in the second buffer, and

wherein the system is configured to, upon the low power audio processor triggering the host processor, copy to the second buffer the at least a portion of the first audio samples that follow the end-point of the predetermined audio pattern.

31. The system of claim 27, wherein the low power audio processor, comprises:

a capture module to monitor the audio signal received by the onboard microphone while the host processor is in the low functionality mode and to capture audio samples of the audio signal;

a language module to identify the predetermined audio pattern in the captured audio samples; and

a trigger module to trigger the host processor of the computing device to transition to the high functionality mode based on the predetermined audio pattern.

32. The system of claim 27, further comprising a single channel noise suppression module to perform noise suppression on the first audio samples.

33. The system of claim 27, further comprising a multi-channel noise suppression module to perform noise suppression on the second audio samples.

34. The system of claim 27, wherein the host processor is configured to, in the high functionality mode, perform speech recognition processing to identify a command.

35. The system of claim 34, wherein the host processor is further configured to perform an additional function based on the identified command

36. The system of claim 34, wherein the host processor is further configured to identify a silence period after determining the command and, during the silence period, switch the system from single-channel processing to multi-channel processing of second audio samples.

37. The system of claim 27, further comprising a plurality of additional microphones operable to receive an audio signal when the host processor is in the high functionality mode,

wherein the one or more microphones comprise the plurality of additional microphones, and

wherein the second audio samples are captured from audio signals received through the plurality of additional microphones.

38. A method to transition a computing device from a low functionality mode to a high functionality mode, comprising:

capturing first audio samples from an audio signal received through a microphone while a host processor of the computing device is in a low functionality mode;

storing the first audio samples in a first buffer;

identifying by a low power audio processor a predetermined audio pattern in the first audio samples, including an end-point of the predetermined audio pattern;

in response to identifying the predetermined audio pattern, triggering the host processor of the computing device to transition to a high functionality mode;

capturing second audio samples from the audio signal received through one or more microphones;

storing the second audio samples; and

processing at least a portion of the first audio samples stored in the first buffer following the end-point of the predetermined audio pattern and the second audio samples by the host processor in the high functionality mode.

39. The method of claim 38, further comprising copying to a second buffer the at least a portion of the first audio samples in the first buffer that follow the end-point of the predetermined audio pattern,

wherein storing the second audio samples comprises storing the second audio samples in the second buffer.

40. The method of claim 38, further comprising performing single channel noise suppression on the first audio samples captured while the host processor is in the low functionality mode.

41. The method of claim 38, further comprising activating one or more microphones based on the predetermined audio pattern, wherein capturing second audio samples comprises capturing the second audio samples from audio signals received through the activated one or more microphones.

42. The method of claim 41, further comprising performing multi-channel noise suppression on the second audio samples captured while the host processor is in the high functionality mode.

43. The method of claim 38, wherein processing the at least a portion of the first audio samples and the second audio samples comprises performing speech recognition to determine a command.

44. The method of claim 43, further comprising executing the command by the host processor in the high functionality mode.

45. The method of claim 43, further comprising:

identifying a silence period after determining the command;

during the silence period, switching from single-mic processing to multi-mic processing of further audio samples.

46. A low power always listening digital signal processor, comprising:

a capture module to monitor an audio signal received by a microphone while a host processor is in a low functionality mode and to capture first audio samples of the audio signal;

a language module to identify a predetermined audio pattern in the first audio samples, including an end-point of the predetermined audio pattern; and

a trigger module to, in response to the language module identifying the predetermined audio pattern, trigger the host processor to transition to a high functionality mode and initiate speech recognition processing on a portion of the first audio samples captured after the end-point of the predetermined audio pattern and on second audio samples captured after the trigger module triggers the host processor.

47. The low power always listening digital signal processor of claim 46, further comprising a first buffer to store the first audio samples.

48. The low power always listening digital signal processor of claim 47, wherein the first buffer is accessible by the host processor.

49. The low power always listening digital signal processor of claim 46, further comprising an onboard microphone to receive the audio signal while the host processor is in the low functionality mode.

50. The low power always listening digital signal processor of claim 46, further comprising a flush module to copy to a second buffer a portion of the first audio samples captured after the end-point of the predetermined audio pattern, the second buffer being accessible by the host processor.

51. One or more machine-readable storage media having stored thereon a plurality of instructions that, in response to being executed by a computing device, result in the computing device performing operations comprising:

capturing first audio samples from an audio signal received through a microphone while a host processor of the computing device is in a low functionality mode;

storing the first audio samples in a first buffer;

identifying by a low power audio processor a predetermined audio pattern in the first audio samples, including an end-point of the predetermined audio pattern;

in response to identifying the predetermined audio pattern, triggering the host processor of the computing device to transition to a high functionality mode;

capturing second audio samples from the audio signal received through one or more microphones;

storing the second audio samples; and

processing at least a portion of the first audio samples stored in the first buffer following the end-point of the predetermined audio pattern and the second audio samples by the host processor in the high functionality mode.

52. The one or more machine-readable storage media of claim 51, further comprising instructions that, in response to being executed by the computing device, result in the computing device performing operations comprising:

copying to a second buffer the at least a portion of the first audio samples in the first buffer that follow the end-point of the predetermined audio pattern,

wherein storing the second audio samples comprises storing the second audio samples in the second buffer.

53. The one or more machine-readable storage media of claim 51, further comprising instructions that, in response to being executed by the computing device, result in the computing device performing operations comprising:

activating one or more microphones based on the predetermined audio pattern, wherein capturing second audio samples comprises capturing the second audio samples from audio signals received through the activated one or more microphones.

54. The one or more machine-readable storage media of claim 51, wherein processing the at least a portion of the first audio samples and the second audio samples comprises performing speech recognition to determine a command.