Method and device for audio input routing

1. A method for performing voice recognition, the method performed by an apparatus and comprising:

prior to receiving a first audio input:

configuring an audio processor of the apparatus with at least one first audio parameter that indicates a low-quality voice recognition engine of the apparatus should be utilized to process the first audio input;

receiving, by the low-quality voice recognition engine of the apparatus via one or more microphones of the apparatus, the first audio input, wherein the low-quality voice recognition engine is local to the apparatus;

determining, by the low-quality voice recognition engine, that the first audio input corresponds to a trigger phrase, the trigger phrase being a particular phrase;

initializing, by a processor of the apparatus and responsive to determining that the first audio input corresponds to the trigger phrase, a high-quality voice recognition engine of the apparatus, wherein the high-quality voice recognition engine is also local to the apparatus, wherein initializing the high-quality voice recognition engine of the apparatus and responsive to determining that the first audio input corresponds to the trigger phrase comprises:

configuring the audio processor of the apparatus with at least one second audio parameter that indicates the high-quality voice recognition engine of the apparatus should be utilized to process any subsequent audio input;

receiving, into a buffer of the apparatus while initializing the high-quality voice recognition engine, a second audio input;

receiving, by the high-quality voice recognition engine after the initialization of the high-quality recognition engine and from the buffer, the second audio input;

performing, by the high-quality voice recognition engine, the voice recognition based on the received second audio input;

determining, by the high-quality voice recognition engine and based on the performed voice recognition on the received second audio input, the second audio input corresponds to a command phrase; and

responsive to determining the received second audio input corresponds to the command phrase, performing a function based on the command phrase.

2. The method of claim 1 , wherein the first audio input is received via an audio input path that excludes the buffer of the apparatus, and wherein the second audio input is received via an audio input path that include the buffer of the apparatus.

3. The method of claim 2 , further comprising:

receiving, by the high-quality voice recognition engine, a third audio input via the audio input path that excludes the buffer and performing, by the high-quality voice recognition engine, voice recognition based on the received third audio input.

4. The method of claim 1 , wherein initializing the high-quality voice recognition engine comprises:

determining a latency parameter to be utilized as the at least one second parameter based at least in part on audio data associated with the first audio input; and

configuring the high-quality voice recognition engine using the latency parameter as the at least one second parameter.

5. The method of claim 4 , wherein the latency parameter is further based at least in part on the buffer of the apparatus.

6. The method of claim 1 , wherein initializing the high-quality voice recognition engine comprises:

determining a frequency range parameter to be utilized as the at least one second parameter based at least in part on audio data associated with the first audio input; and

configuring the high-quality voice recognition engine using the frequency range parameter as the at least one second parameter.

7. The method of claim 1 , wherein the high-quality voice recognition engine is configured using both a latency parameter and a frequency range parameter as the at least one second parameter.

8. The method of claim 7 , wherein initializing the high-quality voice recognition engine comprises:

determining the latency parameter and the frequency range parameter to be utilized as the at least one second parameter based at least in part on audio data associated with the first audio input; and

configuring the high-quality voice recognition engine using both the latency parameter and the frequency range parameter as the at least one second parameter.

9. The method of claim 7 , wherein the high-quality voice recognition engine is further configured using a noise canceling or noise separation feature as the at least one second parameter.

10. An apparatus comprising:

a microphone;

a buffer;

an audio processor that includes a low-quality voice recognition engine and a high-quality voice recognition engine, wherein both the low-quality voice recognition engine and the high-quality voice-recognition engine are local to the apparatus; and

a memory storing instructions that, when executed, cause the audio processor to:

prior to performing a first voice recognition:

configure the audio processor of the apparatus with at least one first audio parameter that indicates the low-quality voice recognition engine of the apparatus should be utilized to perform the first voice recognition;

perform, using the low-quality voice recognition engine, a first voice recognition to determine that a first audio input corresponds to a trigger phrase, wherein the first audio input is received via a first audio input path that excludes the buffer, and wherein the trigger phrase is a particular phrase;

in response to determining that the first audio input corresponds to the trigger phrase, initialize the high-quality voice recognition engine of the apparatus, wherein the instructions to initialize the high-quality engine of the apparatus cause the audio processor to:

configure the audio processor of the apparatus with at least one second audio parameter that indicates the high-quality voice recognition engine of the apparatus should be utilized to perform any subsequent voice recognition; and

subsequent to the initialization of the high-quality recognition engine, perform, by the high-quality voice recognition engine, a second voice recognition to determine that a second audio input corresponds to a command phrase, wherein the second audio input received via a second audio input path that includes the buffer, wherein the second audio input immediately follows the first audio input.

11. The apparatus of claim 10 , wherein the instructions to initialize the high-quality voice recognition engine include instructions that cause the audio processor to:

determine a latency parameter to be utilized as the at least one second parameter based at least in part on audio data associated with the first audio input; and

configure the high-quality voice recognition engine using the latency parameter as the at least one second parameter.

12. The apparatus of claim 10 , wherein the instructions to initialize the high-quality voice recognition engine include instructions that cause the audio processor to:

determine a frequency range parameter to be utilized as the at least one second parameter based at least in part on audio data associated with the first audio input; and

configure the high-quality voice recognition engine using the frequency range parameter as the at least one second parameter.

13. The apparatus of claim 10 , wherein the high-quality voice recognition engine is configured using both a latency parameter and a frequency range parameter as the at least one second parameter.

14. The apparatus of claim 13 , wherein the instructions to initialize the high-quality voice recognition engine include instructions that cause the audio processor to:

determine the latency parameter and the frequency range parameter to be utilized as the at least one second parameter based at least in part on audio data associated with the first audio input; and

configure the high-quality voice recognition engine using both the latency parameter and the frequency range parameter as the at least one second parameter.

15. The apparatus of claim 13 , wherein the high-quality voice recognition engine is further configured using a noise canceling or noise separation feature as the at least one second parameter.

16. The apparatus of claim 10 , wherein the apparatus is a smart phone, a smart watch, or a personal digital assistant.

17. The apparatus of claim 10 , wherein the buffer is a part of the microphone.

18. The apparatus of claim 10 , wherein the first audio input path and the second audio input are digital audio input paths.

19. The apparatus of claim 10 , wherein the first audio input path and the second audio input path are analog audio input paths.