Method and apparatus for facilitating speech barge-in in connection with voice recognition systems

• US 6,266,398 B1
• Filed: 03/12/1998
• Issued: 07/24/2001
• Est. Priority Date: 05/21/1996
• Status: Expired due to Term

• Current Assignee: SpeechWorks International, Inc. (Microsoft Corporation)
• Original Assignee: SpeechWorks International, Inc. (Microsoft Corporation)
• Inventors: Nguyen, John N.
• Primary Examiner(s): Tsang, Fan
• Assistant Examiner(s): Foster, Roland G.
• Application Number:
  US09/041,419
• Time in Patent Office: 1,230 Days
• Field of Search: 379/88.01, 379/88.04, 379/88.07, 379/88.08, 379/88.16, 379/88.24, 379/208, 379/351, 379/406, 379/407, 379/410, 379/411, 379/88.03, 379/88.27, 379/88.28, 704/275, 704/273, 704/251, 704/246, 704/270, 704/233, 704/253, 704/214, 704/248, 704/228
• US Class Current: 379/88.01
• CPC Class Codes: G10L 25/21 the extracted parameters being power information
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10

Abstract

A barge-in detector for use in connection with a speech recognition system forms a prompt replica for use in detecting the presence or absence of user input to the system. The replica is indicative of the prompt energy applied to an input of the system. The detector detects the application of user input to the system, even if concurrent with a prompt, and enables the system to quickly respond to the user input.

Claims

1. A telephone line interface card comprising:

means for coupling the line card to at least one telephone line allowing the line card to receive a line signal indicative of user speech;

means for transmitting prompts to a user on the at least one telephone line; and

means for detecting the presence of user speech on the at least one telephone line concurrent with the transmission of a prompt, the detecting means comprising;

means for measuring a prompt residue resulting from imperfect cancellation of the prompt during at least a portion of a first interval in which said prompt residue is present and said speech is absent;

means for forming, over a subsequent interval, a prompt residue model based on said prompt and the measured residue; and

means for providing an indication of the presence of speech on the at least one telephone line when the signal on said at least one line differs from said prompt residue model by a threshold.

2. The telephone line interface card of claim 1 further comprising means for time-varying the threshold.

3. The telephone line interface card of claim 1 wherein the measuring means measures the difference in amplitude between the prompt and the line signal in the absence of user speech.

4. The telephone line interface card of claim 1 wherein the measuring means measures the difference in energy between the prompt and the line signal in the absence of user speech.

5. The telephone line interface card of claim 1 wherein the means for forming a prompt residue model includes means for subtracting the measured residue from the prompt.

6. A telephone line interface card comprising:

an input configured to couple the line card to at least one telephone line and to receive an input signal indicative of user speech;

a prompt generator configured to transmit prompts to a user on the at least one telephone line; and

a barge-in detector configured to detect the presence of user speech on the input concurrent with the transmission of a prompt by the prompt generator, the barge-in detector comprising;

a residue detector configured to measure a prompt residue resulting from imperfect cancellation of the prompt during at least a portion of a first interval in which the prompt residue is present and the speech is absent; and

a beginning-of-speech detector coupled to the residue detector and configured to, over a subsequent interval, form a prompt residue model based on the prompt and the measured residue, and further configured to provide an indication of the presence of speech on the input when the input signal differs from the prompt residue model by a threshold.

7. The telephone line interface card of claim 6 further comprising a threshold device configured to vary the threshold over time.

8. The telephone line interface card of claim 7 wherein the beginning-of-speech detector is configured to calculate the difference in amplitude between the prompt and the input signal in the absence of user speech.

9. The telephone line interface card of claim 6 wherein the beginning-of-speech detector is configured to calculate the difference in energy between the prompt and the input signal in the absence of user speech.

10. The telephone line interface card of claim 6 wherein the beginning-of-speech detector is configured to subtract the measured residue from the prompt.

11. A telephone line interface card comprising:

an input configured to couple the line card to at least one telephone line and to receive an input signal indicative of user speech;

a prompt generator configured to transmit prompts to a user on the at least one telephone line; and

a barge-in detector configured to detect the presence of user speech on the input concurrent with the transmission of a prompt by the prompt generator, the barge-in detector comprising;

a residue detector configured to measure a prompt residue resulting from imperfect cancellation of the prompt during at least a portion of a first interval in which the prompt residue is present and the speech is absent; and

means for forming, over a subsequent interval, a prompt residue model based on the prompt and the measured residue;

wherein the barge-in detector is configured to provide an indication of the presence of speech on the input when the input signal differs from the prompt residue model by a threshold.

12. The telephone line interface card of claim 11 further comprising a threshold device configured to vary the threshold over time.

13. The telephone line interface card of claim 12 wherein the means for forming the prompt residue model is configured to calculate the difference in amplitude between the prompt and the input signal in the absence of user speech.

14. The telephone line interface card of claim 13 wherein the means for forming the prompt residue model is configured to subtract the calculated difference from the prompt to form the prompt residue model.

15. The telephone line interface card of claim 11 wherein the means for forming the prompt residue model is configured to calculate the difference in energy between the prompt and the input signal in the absence of user speech.

16. The telephone line interface card of claim 11 wherein the means for forming the prompt residue model is configured to subtract the measured residue from the prompt.