Recognition of voice-activated commands

• US 7,996,232 B2
• Filed: 02/19/2009
• Issued: 08/09/2011
• Est. Priority Date: 12/03/2001
• Status: Expired due to Fees

• Current Assignee: Synamedia Limited
• Original Assignee: Scientific-Atlanta LLC (Cisco Systems, Inc.)
• Inventors: Sedacca, David A., Garcia, Albert, Rodriguez, Arturo A.
• Primary Examiner(s): SKED, MATTHEW J
• Application Number:
  US12/388,733
• Publication Number:
  US 20090299752A1
• Time in Patent Office: 901 Days
• Field of Search: None
• US Class Current: 704/275
• CPC Class Codes: D06F 2103/08 Humidity
• CPC Class Codes: D06F 2103/12 Temperature
• CPC Class Codes: D06F 2103/64 Radiation, e.g. microwaves
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

Systems and methods for voice activated commands in a digital home communication terminal are disclosed. One example method includes storing a program audio signal corresponding to a program tuned by the digital home communication terminal. The method also includes storing an incoming audio signal carrying speech and removing from the incoming audio signal a portion of the incoming audio signal that corresponds to the program audio signal, this producing an improved version of the incoming audio signal. The method also includes selecting one of a plurality of voice-activated commands that corresponds to the improved version of the incoming audio signal, and performing a function corresponding to the selected voice-activated command.

Claims

1. A method performed in a digital home communication terminal (DHCT), the method comprising:

storing a program audio signal corresponding to a program tuned by the DHCT, wherein the program audio signal is stored as a plurality of time-shifted versions of the program audio signal;

storing an incoming audio signal carrying speech;

removing from the incoming audio signal a portion of the incoming audio signal that corresponds to the program audio signal, thus producing an improved version of the incoming audio signal;

selecting one of a plurality of voice-activated commands that corresponds to the improved version of the incoming audio signal; and

performing a function corresponding to the selected voice-activated command.

2. The method of claim 1, wherein the incoming audio signal is received from a remote control.

3. The method of claim 1, further comprising:

confirming the selected one of the voice-activated commands with a user before performing the function.

4. The method of claim 1, wherein the number of time-shifted versions is configurable.

5. The method of claim 1, wherein the improved version of the incoming audio signal includes a plurality of candidate incoming audio signals, and the removing further comprises:

generating each of the candidate incoming audio signals by subtracting a respective one of the time-shifted versions of the program audio signal from the incoming audio signal;

and the selecting further comprises;

selecting the one of the voice-activated commands by comparing each of the candidate incoming audio signals to a dictionary of voice-activated commands.

6. A method performed in a digital home communication terminal (DHCT), the method comprising:

storing a program audio signal, corresponding to a program tuned by the DHCT, as samples in a program audio buffer;

receiving an incoming audio signal carrying speech;

storing the incoming audio signal as samples in an incoming audio buffer;

generating a plurality of candidate signals by subtracting, from the samples in the incoming audio buffer, a portion of samples in the incoming audio signal that corresponds to the samples stored in the program audio buffer;

determining a best match within a dictionary of voice-activated commands based on the candidate signals; and

performing a function corresponding to the best match voice-activated command.

7. The method of claim 6, wherein the incoming audio signal is received from a remote control.

8. The method of claim 6, further comprising:

confirming the determined one of the voice-activated commands with a user before performing the function.

9. The method of claim 6, wherein the buffers are implemented as circular buffers.

10. The method of claim 6, wherein the program audio buffer includes a plurality of time-shifted portions, each time-shifted portion storing a different time-shifted version of the program audio signal, wherein the generating further comprises:

generating each of the candidate signals by subtracting the samples in a respective one of the time-shifted portions of the buffer from the samples in the incoming audio buffer.

11. The method of claim 10, wherein at least some of the time-shifted portions overlap.

12. The method of claim 10, wherein the number of time-shifted versions is configurable.

13. A method performed in a digital home communication terminal (DHCT), the method comprising:

demultiplexing a received transport stream into a compressed audio stream and a control stream;

decoding the compressed audio stream;

storing audio corresponding to a tuned program;

storing a start and length of a no-active-speech command, wherein the skipping utilizes the start and the length;

separating the stored audio from the compressed audio stream by subtracting the stored audio from the decoded audio stream, until the no-active-speech command is found in the control stream; and

responsive to the no-active-speech command, skipping the subtracting for a period corresponding to a duration of the no-active-speech command.

14. The method of claim 13, further comprising:

receiving the control stream from a remote control.

15. The method of claim 13, wherein the decoding begins upon finding a start command in the control stream.

16. The method of claim 13, wherein the decoding ends upon finding a stop command in the control stream.

17. The method of claim 13, the separating further comprising subtracting samples of the stored audio from a program buffer from the samples of the decoded audio stream from a speech buffer, the method further comprising:

responsive to the no-active-speech command, advancing a consumer index in the program buffer by an amount corresponding to a duration of the no-active-speech command.