Method and apparatus for detecting speech

• US 5,152,007 A
• Filed: 04/23/1991
• Issued: 09/29/1992
• Est. Priority Date: 04/23/1991
• Status: Expired due to Term

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Uribe, Juan R.
• Primary Examiner(s): Eisenzopf, Reinhard J.
• Assistant Examiner(s): BELZER, CHRISTINE
• Application Number:
  US07/690,033
• Time in Patent Office: 525 Days
• Field of Search: 455/116, 455/79, 455/78, 381/42, 381/46, 381/41, 381/47
• US Class Current: 455/116
• CPC Class Codes: H03G 3/342 Muting when some special characteristic of the signal is sensed which distinguishes it from noise, e.g. using speech detector H03G3/344, H03G3/345 take precedence
• CPC Class Codes: H04M 1/724 User interfaces specially adapted for cordless or mobile telephones
• CPC Class Codes: H04W 52/0251 using monitoring of local events, e.g. events related to user activity
• CPC Class Codes: Y02D 30/70 in wireless communication networks

Abstract

A method of generating a first signal (115) in a radiotelephone (101) representing the detection of speech on a second signal having background noise, a magnitude, a zero crossing rate and divided in time by frames having a predetermined time period. In absence of speech, the method characterizes the background noise (209) and computes background noise thresholds (211). The method computes an average of the magnitude of the second signal and a zero crossing rate of the second signal for a given frame (215). Then, compares said average of the magnitude (217) and said zero crossing rate (221) of the second signal to said background noise threshold. Upon one of said average of the magnitude and said zero crossing rate consecutively exceeding said background noise thresholds for a first predetermined number of frames, the transceiver(107) is turned on.

Claims

1. A method of generating a first signal representing the detection of speech on a second signal having background noise, a magnitude, a zero crossing rate and is divided in time by frames having a predetermined time period, the method comprising the steps of:

characterizing, in absence of speech, the background noise of the second signal comprising the steps of;

computing an average of the magnitude of the second signal for a second predetermined number of frames, and

computing an average of the zero crossing rate of the second signal for said second predetermined number of frames;

computing background noise magnitude and zero crossing rate thresholds for the second signal;

computing an average of the magnitude of the second signal for a first frame;

computing a zero crossing rate of the second signal for said first frame;

comparing said average of the magnitude of the second signal to said background noise magnitude threshold;

comparing said zero crossing rate of the second signal to said background noise zero crossing rate threshold; and

generating, upon said average of the magnitude or said zero crossing rate consecutively exceeding said background noise threshold a first predetermined number of frames, the first signal.

2. A method of generating a first signal in accordance with claim 1 wherein said computing said background noise magnitude and zero crossing rate thresholds further comprises the steps of:

computing a standard deviation of the magnitude of the second signal for said second predetermined number of frames;

computing a standard deviation of the zero crossing rate of the second signal for said second predetermined number of frames;

multiplying said standard deviation of the magnitude by a second predetermined number and adding said average of the magnitude; and

multiplying said standard deviation of the zero crossing rate by a third predetermined number and adding said average of the zero crossing rate.

3. A method of generating a first signal in accordance with claim 1 wherein said second signal is a speech channel signal.

4. A radiotelephone including a power saving apparatus, a transmitter, and a handset, the handset generating a second signal responding to a user'"'"'s speech, having a magnitude, a zero crossing rate, background noise and is divided in time by frames having a predetermined time period, the apparatus comprising:

means for converting the second signal to a digital representation;

means for detecting the user'"'"'s speech in said digital representation of the second signal comprising;

means for characterizing, in absence of the user'"'"'s speech, the background noise of the second signal, comprising;

computing an average of the magnitude of the second signal for a second predetermined number of frames, and

computing an average of the zero crossing rate of the second signal for said second predetermined number of frames,

means for computing background noise magnitude and zero crossing rate thresholds for the second signal,

means for computing an average of the magnitude of the second signal for a first frame,

means for computing a zero crossing rate of the second signal for said first frame,

means for comparing said average of the magnitude of the second signal to said background noise magnitude threshold,

means for comparing said zero crossing rate of the second signal to said background noise zero crossing rate threshold, and

means for generating, upon said average of the magnitude or said zero crossing rate consecutively exceeding said background noise threshold a first predetermined number of frames, a first signal, indicating detection of the user'"'"'s speech on the second signal; and

means, responsive to detection of the user'"'"'s speech, for turning on the transmitter.

5. An apparatus for generating a first signal representing the detection of speech on a second signal having background noise, a magnitude, a zero crossing rate and is divided in time by frames having a predetermined time period, the apparatus comprising:

means for characterizing, in absence of speech, the background noise of the second signal comprising;

means for computing an average of the magnitude of the second signal for a second predetermined number of frames, and

means for computing an average of the zero crossing rate of the second signal for said second predetermined number of frames;

means for computing background noise magnitude and zero crossing rate thresholds for the second signal;

means for computing an average of the magnitude of the second signal for a first frame;

means for computing an average of the zero crossing rate of the second signal for said first frame;

means for comparing said average of the magnitude of the second signal to said background noise magnitude threshold;

means for comparing said average of the zero crossing rate of the second signal to said background noise zero crossing rate threshold; and

means for generating, upon said average of the magnitude or said zero crossing rate consecutively exceeding said background noise threshold a first predetermined number of frames, the first signal.

6. An apparatus for generating a first signal in accordance with claim 5 wherein said computing said background noise magnitude and zero crossing rate thresholds further comprises:

means for computing a standard deviation of the magnitude of the second signal for said second predetermined number of frames;

means for computing a standard deviation of the zero crossing rate of the second signal for said second predetermined number of frames;

means for multiplying said standard deviation of the magnitude by a second predetermined number and adding said average of the magnitude; and

means for multiplying said standard deviation of the zero crossing rate by a third predetermined number and adding said average of the zero crossing rate.

7. An apparatus for generating a first signal in accordance with claim 5 wherein said second signal is a speech channel signal.