Granted Patent B2

US 12,334,065 · App. 17/315,704 · Granted Jun 17, 2025

Inventors: Ralph Birt (Santa Clara, CA); Robert Curtis (Los Gatos, CA); David H. Friedman (Austin, TX)

Assignee: ROKU, INC.

G10L15/22G10L15/08G10L21/034G10L21/0364G10L25/51G10L2015/088G10L2015/223

Abstract

Claims (72)

1. A computer implemented method, comprising:

receiving a voice command directed to controlling a device, the voice command comprising a wake command and an action command;

determining an amplitude of the wake command;

determining a previous gain adjustment based on historical information identifying calculated gain adjustments for a plurality of previous voice commands;

applying the previous gain adjustment to only the wake command;

adjusting an amplitude of the action command to generate an amplitude-adjusted action command based on the previous gain adjustment applied to the wake command and a difference in the amplitude of the wake command and the amplitude of the action command;

adjusting an amplitude of the amplitude-adjusted action command to generate a second amplitude-adjusted action command based on a difference in the amplitude of the amplitude-adjusted action command and a target amplitude;

identifying a device command for controlling the device based on the second amplitude-adjusted action command; and

providing the device command to the device.

2. The method of claim 1 , wherein the previous gain adjustment is a median of the calculated gain adjustments for the plurality of previous voice commands.

3. The method of claim 1 , further comprising:

determining that the calculated gain adjustment for the wake command is an anomaly, wherein the calculated gain adjustment is excluded from the historical information when processing a subsequent voice command.

4. The method of claim 1 , further comprising:

calculating a correction based on the difference in the amplitude of the amplitude-adjusted action command and the target amplitude, wherein the correction is applied to an amplitude of a subsequent action command of a subsequent voice command.

5. The method of claim 4 , wherein calculating the correction comprises:

calculating a first correction for a first user; and

calculating a second correction for a second user different from the first user.

6. The method of claim 1 , further comprising:

detecting the wake command prior to receiving the action command, wherein the device is configured to output an audible beep upon detecting the wake command;

determining that the voice command comprises a continuous stream of speech after detecting the wake command;

suppressing the audible beep based on the continuous stream of speech determination; and

detecting the wake command from the voice command.

7. The computer implemented method of claim 1 , further comprising:

retrieving the historical information via a network.

8. A system, comprising:

a memory; and

at least one processor coupled to the memory and configured to perform operations comprising:

receiving a voice command directed to controlling a device, the voice command comprising a wake command and an action command;

determining an amplitude of the wake command;

determining a previous gain adjustment based on historical information identifying calculated gain adjustments for a plurality of previous voice commands;

applying the previous gain adjustment to only the wake command;

adjusting an amplitude of the action command to generate an amplitude-adjusted action command based on the previous gain adjustment applied to the wake command and a difference in the amplitude of the wake command and the amplitude of the action command;

adjusting an amplitude of the amplitude-adjusted action command to generate a second amplitude-adjusted action command based on a difference in the amplitude of the amplitude-adjusted action command and a target amplitude;

identifying a device command for controlling the device based on the second amplitude-adjusted action command; and

providing the device command to the device.

9. The system of claim 8 , wherein the previous gain adjustment is a median of the calculated gain adjustments for the plurality of previous voice commands.

10. The system of claim 8 , wherein the operations further comprise:

determining that the calculated gain adjustment for the voice command is an anomaly, wherein the calculated gain adjustment for the wake command is excluded from the historical information when processing a subsequent voice command.

11. The system of claim 8 , wherein the operations further comprise:

calculating a correction based on the difference in the amplitude of the amplitude-adjusted action command and the target amplitude, wherein the correction is applied to an amplitude of a subsequent action command of a subsequent voice command.

12. The system of claim 11 , wherein calculating the correction comprises:

calculating a first correction for a first user; and

calculating a second correction for a second user different from the first user.

13. The system of claim 8 , wherein the operations further comprise:

detecting the wake command prior to receiving the action command, wherein the device is configured to output an audible beep upon detecting the wake command;

determining that the voice command comprises a continuous stream of speech after detecting the wake command;

suppressing the audible beep based on the continuous stream of speech determination; and

detecting the wake command from the voice command.

14. The system of claim 9 , wherein the operations further comprise:

retrieving the historical information via a network.

15. A non-transitory computer-readable medium having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:

receiving a voice command directed to controlling a device, the voice command comprising a wake command and an action command;

determining an amplitude of the wake command;

determining a previous gain adjustment based on historical information identifying calculated gain adjustments for a plurality of previous voice commands;

applying the previous gain adjustment to only the wake command;

adjusting an amplitude of the action command to generate an amplitude-adjusted action command based on the previous gain adjustment applied to the wake command and a difference in the amplitude of the wake command and the amplitude of the action command;

adjusting an amplitude of the amplitude-adjusted action command to generate a second amplitude-adjusted action command based on a difference in the amplitude of the amplitude-adjusted action command and a target amplitude;

identifying a device command for controlling the device based on the second amplitude-adjusted action command; and

providing the device command to the device.

16. The non-transitory computer-readable medium of claim 15 , wherein the previous gain adjustment is a median of the calculated gain adjustments for the plurality of previous voice commands.

17. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

determining that the calculated gain adjustment for the wake command is an anomaly, wherein the calculated gain adjustment is excluded from the historical information when processing a subsequent voice command.

18. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

calculating a correction based on the difference in the amplitude of the amplitude-adjusted action command and the target amplitude, wherein the correction is applied to an amplitude of a subsequent action command of a subsequent voice command.

19. The non-transitory computer-readable medium of claim 18 , wherein calculating the correction comprises:

calculating a first correction for a first user; and

calculating a second correction for a second user different from the first user.

20. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

detecting the wake command prior to receiving the action command, wherein the device is configured to output an audible beep upon detecting the wake command;

determining that the voice command comprises a continuous stream of speech after detecting the wake command;

suppressing the audible beep based on the continuous stream of speech determination; and

detecting the wake command from the voice command.

Assignments (2)

Continuity (1)

References Cited (17)