Altering audio to improve automatic speech recognition

• US 9,251,787 B1
• Filed: 09/26/2012
• Issued: 02/02/2016
• Est. Priority Date: 09/26/2012
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Hart, Gregory M., Worley, William Spencer III
• Primary Examiner(s): Paul, Disler
• Application Number:
  US13/627,890
• Time in Patent Office: 1,224 Days
• Field of Search: 381 56- 59, 381/110, 381104-109, 381 941- 949, 381/86, 704/273, 704275-276, 704/226, 704/233
• US Class Current: 1/1
• 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
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G11B 27/005 Reproducing at a different information rate from the information rate of recording for television signals H04N5/783
• CPC Class Codes: H03G 3/32 the control being dependent upon ambient noise level or sound level
• CPC Class Codes: H03G 5/02 Manually-operated control
• CPC Class Codes: H04R 3/12 for distributing signals to two or more loudspeakers specially adapted for hearing aids H04R25/407

Abstract

Techniques for altering audio being output by a voice-controlled device, or another device, to enable more accurate automatic speech recognition (ASR) by the voice-controlled device. For instance, a voice-controlled device may output audio within an environment using a speaker of the device. While outputting the audio, a microphone of the device may capture sound within the environment and may generate an audio signal based on the captured sound. The device may then analyze the audio signal to identify speech of a user within the signal, with the speech indicating that the user is going to provide a subsequent command to the device. Thereafter, the device may alter the output of the audio (e.g., attenuate the audio, pause the audio, switch from stereo to mono, etc.) to facilitate speech recognition of the user'"'"'s subsequent command.

Claims

1. An apparatus comprising;

a speaker to output audio in an environment;

a microphone unit to capture sound from the environment;

a processor; and

computer-readable media storing computer-executable instructions that, when executed on the processor, cause the processor to perform acts comprising;

receiving an audio signal generated by the microphone unit, the microphone unit having generated the audio signal based at least in part on the sound captured by the microphone unit, wherein the sound includes an utterance from a user in the environment, the utterance indicating that the user is going to provide a subsequent request to the apparatus;

identifying one or more characteristics associated with at least one of the audio signal or the audio being output by the speaker, the one or more characteristics at least including information indicative of a distance between the user and the apparatus;

determining, based at least in part on the one or more characteristics, an amount to attenuate the audio being output by the speaker to facilitate processing of the subsequent request, the amount to attenuate the audio increasing with increasing distance between the user and the apparatus; and

attenuating the audio being output by the speaker by the determined amount.

2. An apparatus as recited in claim 1, wherein the one or more characteristics further comprise information indicative of whether the audio that is being output by the speaker comprises a song or an audio book.

3. An apparatus as recited in claim 1, wherein the one or more characteristics further comprise at least one of:

information indicative of the identity of the user;

or

information indicative of a direction of the user relative to the apparatus.

4. An apparatus as recited in claim 1, wherein:

the apparatus further comprises an additional speaker, the speaker and the additional speaker outputting the audio in stereo;

the attenuating comprises, at least in part, altering a signal sent to the speaker and the additional speaker from a stereo signal to a mono signal.

5. A computer-implemented method comprising:

receiving, while a speaker of a device outputs audio, an indication that a user is going to provide a subsequent voice command to the device;

determining a distance between the user and the device;

determining an amount to attenuate the audio output by the speaker of the device, the amount to attenuate the audio increasing with increasing distance between the user and the device; and

attenuating the audio by the determined amount to increase an accuracy of speech recognition performed on an audio signal that includes the subsequent voice command by increasing a signal-to-noise ratio of the audio signal.

6. The computer-implemented method as recited in claim 5, wherein the attenuating comprises, at least in part, altering a signal sent to the speaker from a stereo signal to a mono signal.

7. The computer-implemented method as recited in claim 5, further comprising determining an identity of the user, and wherein the determining the amount to attenuate the audio is further based at least in part on the identity of the user.

8. The computer-implemented method as recited in claim 5, further comprising identifying a class of content being output, and wherein the determining the amount to attenuate the audio is further based at least in part on the type of content being output.

9. The computer-implemented method as recited in claim 5, further comprising determining a frequency range of the audio to attenuate.

10. The computer-implemented method as recited in claim 9, further comprising determining an identity of the user, and wherein the determining the frequency range of the audio to attenuate is further based at least in part on the identity of the user.

11. The computer-implemented method as recited in claim 5, the wherein the indication comprises the user speaking a predefined word or phrase.

12. A method comprising:

under control of an electronic device that includes a microphone, a speaker and executable instructions,

outputting audio via the speaker;

determining that a user is going to provide a voice command to the device based at least in part on an utterance from the user represented in an audio signal generated by the microphone;

determining a distance between the user and the electronic device;

determining an amount to attenuate the audio, the amount to attenuate the audio increasing with increasing distance between the user and the electronic device; and

attenuating the audio by the determined amount to increase an accuracy of speech recognition performed on an audio signal that includes the subsequent voice command by increasing a signal-to-noise ratio of the audio signal.

13. A method as recited in claim 12, wherein the attenuating comprises, at least in part, switching a signal sent to the speaker from a stereo signal to a mono signal.

14. A method as recited in claim 12, wherein the device includes the speaker and an additional speaker, and the attenuating comprises, at least in part, ceasing to send a signal to the speaker or the additional speaker.