Building a Dictionary Based on Speech Signals that are Compressed

• US 20080091434A1
• Filed: 12/14/2007
• Published: 04/17/2008
• Est. Priority Date: 12/03/2001
• Status: Active Grant

• Current Assignee: Synamedia Limited
• Original Assignee: Scientific-Atlanta (Cisco Systems, Inc.)
• Inventors: Sedacca, David, Garcia, Albert, Rodriguez, Arturo
• Granted Patent: US 8,849,660 B2
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: G08C 2201/31 Voice input
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 19/00 Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis in musical instruments G10H
• CPC Class Codes: G10L 19/008 Multichannel audio signal coding or decoding using interchannel correlation to reduce redundancy, e.g. joint-stereo, intensity-coding or matrixing
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: H04N 21/42203 sound input device, e.g. microphone
• CPC Class Codes: H04N 21/42206 characterized by hardware details
• CPC Class Codes: H04N 21/4394 involving operations for analysing the audio stream, e.g. detecting features or characteristics in audio streams arrangements characterised by components specially adapted for monitoring, identification or recognition of audio in broadcast systems H04H60/58

Abstract

Systems and methods for training voice activation control of electronic equipment are disclosed. One example method includes receiving a selection corresponding to at least one command used to control the electronic equipment. The method further includes instructing a user to speak, and responsive to the instruction, receiving a digitized speech stream. The method further includes segmenting the speech stream into speech segments, storing at least one of the speech segments as an entry in a dictionary, and associating the dictionary entry with the selected command.

Claims

1. A method for training voice activation control of electronic equipment, the method comprising:

receiving a selection corresponding to at least one command used to control the electronic equipment;

instructing a user to speak;

responsive to the instruction, receiving a digitized speech stream;

segmenting the speech stream into speech segments;

storing one of the speech segments as an entry in a dictionary; and

associating the dictionary entry with the selected command.

2. The method of claim 1, further comprising:

storing a sequence from the speech segments as the entry in the dictionary.

3. The method of claim 1, further comprising:

receiving an indication from the user that more than one word is to be associated with the selected command; and

storing a sequence from the speech segments as the entry in the dictionary, responsive to the indication.

4. The method of claim 1, wherein the selection corresponds to a plurality of commands.

5. The method of claim 1, wherein the selection corresponds to a plurality of commands, the method further comprising:

receiving an indication from the user that more than one command is to be associated with the particular word; and

responsive to the indication, associating the dictionary entry with the plurality of commands.

6. The method of claim 1, wherein the instructing further comprises instructing the user to speak a particular word.

7. The method of claim 1, wherein the instructing further comprises instructing the user to speak one word in a list of predetermined words.

8. The method of claim 1, wherein the instructing further comprises instructing the user to speak a word multiple times, the method further comprising:

storing multiple speech segments as an entry in the dictionary; and

associating the dictionary entry with the selected command.

9. A digital home communication terminal (DHCT) comprising:

memory having stored thereon program code; and

a processor that is programmed by the program code to;

receive a selection corresponding to at least one command used to control electronic equipment;

instruct a user to speak;

responsive to the instruction, receive a digitized speech stream;

segment the speech stream into speech segments;

store one of the speech segments as an entry in a dictionary; and

associate the dictionary entry with the selected command.

10. The DHCT of claim 9, wherein the processor is further programmed by the program code to:

emit a plurality of varied-length pink noise pulses, each having a respective emission start time and emission stop time;

decode a first audio signal received from a source, the first decoded audio signal corresponding to the plurality of pink noise pulses;

determine, from the first decoded audio signal, a respective receive start time and receive stop time for each of the pink noise pulses; and

compute a roundtrip delay between the DHCT and the source, based on the plurality of emission start and stop times and the plurality of receive start and stop times.

11. The DHCT of claim 10, wherein the processor is further programmed by the program code to:

receive a voice activation training signal,

wherein the emission of the plurality of varied-length pink noise pulses is responsive to receiving the voice activation training signal.

12. The DHCT of claim 10, wherein the processor is further programmed by the program code to:

instruct the user to initiate voice activation training; and

receive a voice activation signal,

wherein the emission of the plurality of varied-length pink noise pulses is responsive to receiving the voice activation training signal.

13. The DHCT of claim 10, wherein the processor is further programmed by the program code to:

play program audio;

buffer the program audio;

decode a second audio signal received from the source, to produce a plurality of samples;

compare the buffered program audio with the plurality of samples, taking into account the computed roundtrip delay; and

produce an estimate of signal degradation based on the comparison.

14. The DHCT of claim 13, wherein the processor is further programmed by the program code to:

receive a third audio signal from the source, the third audio signal comprising a voice-activated command and program audio;

separate the program audio from the voice-activated command using the estimate of signal degradation.

15. The DHCT of claim 9, wherein the processor is further programmed by the program code to:

play program audio;

buffer the program audio;

decode an audio signal received from a source to produce a plurality of samples;

compute a roundtrip delay between the DHCT and the source;

compare the buffered program audio with the plurality of samples, taking into account the computed roundtrip delay; and

produce an estimate of signal degradation based on the comparison.

16. The DHCT of claim 15, wherein the processor is further programmed by the program code to:

receive a second audio signal from the source, the second audio signal comprising a voice-activated command and program audio; and

separate the program audio from the voice-activated command using the estimate of signal degradation.

17. A digital home communication terminal (DHCT) comprising:

logic configured to instruct a user to speak;

logic configured to responsive to the instruction, receive a digitized speech stream;

logic configured to segment the speech stream into speech segments;

logic configured to store one of the speech segments as an entry in a dictionary; and

logic configured to associate the dictionary entry with the selected command.

18. The DHCT of claim 17, further comprising:

logic configured to emit a plurality of varied-length pink noise pulses, each having a respective emission start time and emission stop time;

logic configured to decode a first audio signal received from a source, the first decoded audio signal corresponding to the plurality of pink noise pulses;

logic configured to determine, from the first decoded audio signal, a respective receive start time and receive stop time for each of the pink noise pulses;

logic configured to compute a roundtrip delay between the DHCT and the source, based on the plurality of emission start and stop times and the plurality of receive start and stop times.

logic configured to play program audio;

logic configured to buffer the program audio;

logic configured to decode a second encoded audio signal received from the source, to produce a plurality of samples;

logic configured to compare the buffered program audio with the plurality of samples, taking into account the computed roundtrip delay; and

logic configured to produce an estimate of signal degradation based on the comparison.

19. The DHCT of claim 18, further comprising:

logic configured to receive a voice activation training signal,

wherein the logic configured to emit is responsive to receiving the voice activation training signal.

20. The DHCT of claim 18, further comprising:

logic configured to receive a third audio signal from the source, the third audio signal comprising a voice-activated command and program audio; and

logic configured to separate the program audio from the voice-activated command using the estimate of signal degradation.