Audio feedback time shift filter system and method

• US 8,738,382 B1
• Filed: 12/16/2005
• Issued: 05/27/2014
• Est. Priority Date: 12/16/2005
• Status: Active Grant

• Current Assignee: NVIDIA Corporation
• Original Assignee: NVIDIA Corporation
• Inventors: Herz, William Samuel
• Primary Examiner(s): Godbold, Douglas
• Assistant Examiner(s): ORTIZ SANCHEZ, MICHAEL
• Application Number:
  US11/303,556
• Time in Patent Office: 3,084 Days
• Field of Search: 704/270, 704/275, 704/233
• US Class Current: 704/275
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• 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 21/0216 characterised by the method used for estimating noise

Abstract

Audio feedback time shifted filtering systems and methods are presented. The systems and methods facilitate separation of program audio feedback from received environmental audio (e.g., audio sensed by a microphone.) The separation of the program audio feedback reduces interference from program content audio feedback on performance of voice recognition operations. In one embodiment of a personal video recorder audio filter method, environmental audio patterns are received, an audio feedback time shift filter process is executed for separating out program content from the environmental audio patterns, and voice recognition is performed on the filtered environment audio patterns (without interference from program audio content feedback). The time shift or deterministic delay provides a closer correlation between program audio content and program audio content feedback received at the microphone and permits input timing compensation to compensate for feedback loop delays.

Claims

1. A personal video recorder audio filter method comprising:

receiving program content at a program content receiving component that forwards the program content to a program content feedback separator and to an audio emitter;

receiving environmental audio patterns at a microphone wherein, upon an annunciation of the received program content by the audio emitter, the received environmental audio patterns comprise a feedback of an annunciated instance of the program content and wherein the received audio patterns are forwarded to the program content feedback separator;

separating a component of a portion of the received program content from a corresponding portion of the received environmental audio patterns, wherein the separating comprises;

compensating for a delay between the feedback of a portion of the annunciated program content instance and by time shifting a stored un-annunciated instance of the program content portion;

subtracting information corresponding to the time shifted un-annunciated program content portion component from the received environmental audio patterns; and

adjusting an amplitude of the program content;

performing voice recognition on a remainder of the environment audio patterns after the subtraction, wherein the voice recognition is performed independent of the subtracted program content portion; and

upon detecting, based on the performing the voice recognition, that the remainder environment audio patterns comprise at least one control command related audio pattern, executing one or more control commands corresponding to the control command related audio pattern;

wherein said program content feedback separation component stops one or more false commands, which correspond to at least an element of said program content, when user commands are not being issued.

2. A personal video recorder audio filter method of claim 1 wherein said method facilitates minimization of interference from program content audio feedback on performance of said voice recognition.

3. A personal video recorder audio filter method of claim 1 wherein a delay associated with said separating is tapped into an existing personal video time shift buffer and is user variable.

4. A personal video recorder audio filter method of claim 1 wherein said separating time shifts said filtered program content to compensate for a feedback loop delay.

5. A personal video recorder audio filter method of claim 1 wherein said separating comprises:

tracking program content audio; and

synchronizing said tracking of said program content audio to correspond with program audio content input feedback, wherein a is utilized to perform said synchronizing.

6. A personal video recorder audio filter method of claim 5 wherein said delay of said program content audio comprises a deterministic feedback time period.

7. A personal video recorder audio filter method of claim 6 wherein said deterministic feedback time period includes a time sufficient for program content audio to be emitted and sensed in a feed back loop.

8. A personal video recorder command recognition system comprising:

a program content receiving component operable for receiving a program content input;

a microphone component operable for receiving environmental audio patterns,

wherein the environmental audio patterns comprise a feedback of an annunciated instance of the program content upon an annunciation of a portion thereof;

an audio emitting component coupled to the program content receiving component and operable for annunciating the annunciation of the program content;

a program content feedback separation component coupled to the microphone component, and to the program content receiving component, and operable for separating the program content from the received environmental audio patterns, wherein the program content feedback compensation component comprises;

an input timing compensator operable for compensating for a delay between the feedback of a portion of the annunciated program content instance and by time shifting a stored un-annunciated instance of the program content portion;

a program content subtractor operable for subtracting information corresponding to the time shifted un-annunciated program content portion component from the received environmental audio patterns; and

an amplifier operable for amplifying the program content, wherein the amplifying comprises adjusting an amplitude of the program content to facilitate an accurate filtering;

a command recognition component coupled to the program content feedback separation component and operable for recognizing audio commands in a remainder of the environment audio patterns after the subtraction of the program content portion therefrom;

wherein the command recognition component comprises;

a memory for storing a table mapping to commands to voice patterns; and

a voice command recognizer operable for recognizing the voice patterns; and

a command implementing component coupled to the command recognition component and operable for implementing corresponding control command instructions;

wherein, upon the annunciation of the program content portion by the audio emitting component, the program content portion is sensed in a feedback loop by the microphone component and is deterred from being forwarded to the command recognition component and prevented from being misinterpreted as a user command and wherein the command recognition component stops one or more false commands, which correspond to at least an element of the program content, when user commands are not being issued.

9. A personal video recorder command recognition system of claim 8 wherein said input timing compensator delays said program content audio patterns over said second time period, during which said program content feedback compensation component is further operable for receiving said annunciated program content from said microphone component as an audio feedback input.

10. A personal video recorder command recognition system comprising:

means for receiving program content at a program content receiving component that forwards the program content to a means for separating program content feedback and to a means for emitting audio;

means for receiving environmental audio patterns at a microphone wherein, upon an annunciation of the received program content by the audio emitting means, the received environmental audio patterns comprise a feedback of an annunciated instance of the program content and wherein the received audio patterns are forwarded to the means for separating the program content feedback;

the means for separating a component of a portion of the received program content from a corresponding portion of the received environmental audio patterns, wherein the separating means comprises;

means for compensating for a delay between the feedback of a portion of the annunciated program content instance and by time shifting a stored un-annunciated instance of the program content portion;

means for subtracting information corresponding to the time shifted un-annunciated program content portion component from the received environmental audio patterns; and

means for adjusting an amplitude of the program content;

means for performing voice recognition on a remainder of the environment audio patterns after the subtraction, wherein the voice recognition is performed independent of the subtracted portion; and

means for executing, based on a detection by the means for performing the voice recognition, that the remainder environment audio patterns comprise at least one control command related audio pattern, one or more control commands corresponding to the control command related audio pattern;

wherein said subtraction means functions to stop one or more false commands, which correspond to at least an element of said program audio information, when user commands are not being issued.

11. A personal video recorder command recognition system of claim 10 wherein said means for separating comprises means for tracking an un-annunciated version of said program audio information over said first time period to allow an annunciated version of said program audio information to be sensed.

12. A personal video recorder command recognition system of claim 10 further comprising:

a means for mapping voice patterns to instructions; and

a means for recognizing said voice patterns in said annunciated version of said program audio information.

13. A personal video recorder command recognition system of claim 10 wherein an un-annunciated version of program audio information is received in a compressed digitized format that is decompressed by said means for receiving said program content.