ELECTRONIC DEVICES AND METHODS THAT ADAPT FILTERING OF A MICROPHONE SIGNAL RESPONSIVE TO RECOGNITION OF A TARGETED SPEAKER'S VOICE

• US 20090216529A1
• Filed: 02/27/2008
• Published: 08/27/2009
• Est. Priority Date: 02/27/2008
• Status: Active Grant

• Current Assignee: Sony Ericsson Mobile Communications USA Incorporated (Sony Group Corp.)
• Original Assignee: Sony Ericsson Mobile Communications USA Incorporated (Sony Group Corp.)
• Inventors: Bengtsson, Henrik Sven
• Granted Patent: US 7,974,841 B2
• Time in Patent Office: Days
• US Class Current: 704/233
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 21/02 Speech enhancement, e.g. noise reduction or echo cancellation reducing echo effects in line transmission systems H04B3/20; echo suppression in hands-free telephones H04M9/08

Abstract

Electronic devices and methods are disclosed that adaptively filter a microphone signal responsive to recognition of a targeted speaker'"'"'s voice. An electronic device can include a microphone, a speaker characterization circuit, an adaptive sound filter circuit, and a speaker recognition circuit. The speaker characterization circuit operates in a training mode to learn characteristics of the targeted speaker'"'"'s voice component in the microphone signal, and to store the learned characteristics. The adaptive sound filter circuit adaptively filters the microphone signal responsive to a control signal. The speaker recognition circuit uses the learned characteristics to recognize the presence of the targeted speaker'"'"'s voice in the microphone signal and to regulate the control signal to cause the adaptive sound filter circuit to adapt the filtering to increase the targeted speaker'"'"'s voice component of the microphone signal relative to other components.

Claims

1. An electronic device comprising:

a microphone that generates a microphone signal responsive to incident sound; and

a sound processor that is configured to recognize a targeted speaker'"'"'s voice within the microphone signal and to regulate filtering of the microphone signal responsive to the recognition of the targeted speaker'"'"'s voice.

2. The electronic device of claim 1, wherein the sound processor comprises:

a speaker characterization circuit that is configured to operate in a training mode to learn characteristics of the targeted speaker'"'"'s voice component in the microphone signal, and to store the learned characteristics;

an adaptive sound filter circuit that is configured to adaptively filter the microphone signal responsive to a control signal; and

a speaker recognition circuit that is configured to use the learned characteristics to recognize the presence of the targeted speaker'"'"'s voice in the microphone signal and to regulate the control signal to cause the adaptive sound filter circuit to adapt the filtering to increase the targeted speaker'"'"'s voice component of the microphone signal relative to other components.

3. The electronic device of claim 2, wherein:

the adaptive sound filter circuit comprises a variable bandpass filter having bandpass frequencies that are varied in response to the control signal; and

the speaker recognition circuit is further configured to respond to recognition of the presence of the targeted speaker'"'"'s voice in the microphone signal by using the learned characteristics to determine a frequency range of the targeted speaker'"'"'s voice and by regulating the control signal to vary the bandpass frequencies of the variable bandpass filter in response to the determined frequency range of the targeted speaker'"'"'s voice.

4. The electronic device of claim 2, wherein:

the adaptive sound filter circuit comprises a high-pass filter having a cutoff frequency that is varied in response to the control signal; and

the speaker recognition circuit is further configured to respond to recognition of the presence of the targeted speaker'"'"'s voice in the microphone signal by using the learned characteristics to determine a lower frequency range of the targeted speaker'"'"'s voice and by regulating the control signal to vary the cutoff frequency of the high-pass filter in response to the determined lower frequency range of the targeted speaker'"'"'s voice.

5. The electronic device of claim 2, wherein:

the speaker characterization circuit is further configured to operate during the training mode to learn, from the targeted speaker'"'"'s voice component in the microphone signal, a time-varying spectral pattern of the targeted speaker'"'"'s voice as the targeted speaker vocalizes a particular song/speech; and

the speaker recognition circuit is further configured to use the learned time-varying spectral pattern to recognize the presence of the targeted speaker'"'"'s voice in the microphone signal vocalizing the particular song/speech, and to respond to the recognized presence by using the learned time-varying spectral pattern to vary the control signal over time to dynamically tune the adaptive sound filter circuit to track the expected time-varying spectral pattern of the targeted speaker'"'"'s voice during the song/speech and to increase the targeted speaker'"'"'s voice component in the microphone relative to other components.

6. The electronic device of claim 5, wherein:

the adaptive sound filter circuit comprises a variable bandpass filter having bandpass frequencies that are varied in response to the control signal;

the speaker characterization circuit is further configured to learn the time-varying spectral pattern by characterizing differences over time in the frequency range of the targeted speaker'"'"'s voice in the microphone signal as the targeted speaker vocalizes the particular song/speech; and

the speaker recognition circuit is further configured to use the learned time-varying spectral pattern to tune the bandpass frequencies of the bandpass filter to follow the differences over time in the frequency range of the targeted speaker'"'"'s voice so as to increase the targeted speaker'"'"'s voice component in the microphone signal relative to other components during the particular song/speech.

7. The electronic device of claim 5, wherein:

the speaker characterization circuit is further configured to learn timing of gaps where the targeted speaker'"'"'s voice in the microphone signal has less than a threshold amplitude as the targeted speaker vocalizes the particular song/speech; and

the speaker recognition circuit is further configured to use the learned gap timing to vary the control signal over time to increase the other components of the microphone signal relative to the component of the targeted speaker'"'"'s voice in the microphone signal during each of the gaps and to increase the component of the targeted speaker'"'"'s voice relative to the other components of the microphone signal in-between each of the gaps.

8. The electronic device of claim 2, wherein:

the adaptive sound filter circuit comprises a variable gain amplifier that amplifies the microphone signal and has a gain that is varied in response to the control signal;

the speaker characterization circuit is further configured to operate during the training mode to learn, from the targeted speaker'"'"'s voice component in the microphone signal, a time-varying amplitude pattern of the targeted speaker'"'"'s voice as the targeted speaker vocalizes a particular song/speech; and

the speaker recognition circuit is further configured to respond to the recognized presence of the targeted speaker'"'"'s voice by using the learned time-varying amplitude pattern to vary the control signal over time to dynamically tune the gain of the variable gain amplifier to track the expected time-varying amplitude pattern of the targeted speaker'"'"'s voice during the song/speech.

9. The electronic device of claim 2, further comprising:

a radio frequency receiver that is configured to receive a wireless communication signal from a remote transmitter that is carried by the targeted speaker; and

a distance determination circuit that is configured to determine a signal strength of the received communication signal, and to estimate distance from the electronic device to the remote transmitter in response to the received signal strength,

wherein the speaker characterization circuit is further configured to adapt the learned characteristics of the targeted speaker'"'"'s voice in response to the estimated distance from the electronic device to the remote transmitter to compensate for expected effects of distance on characteristics of the targeted speaker'"'"'s voice component during the filtering of the microphone signal.

10. The electronic device of claim 2, further comprising:

an image sensor that is configured to generate an image signal; and

a distance determination circuit that is configured to recognize presence of a person in the image signal that is expected to correspond or be proximately located to the targeted speaker, and to estimate distance from the electronic device to the person in response to size of the person relative to a plurality of threshold values,

wherein the speaker characterization circuit is further configured to adapt the learned characteristics of the targeted speaker'"'"'s voice in response to the estimated distance from the electronic device to the recognized person to compensate for expected effects of distance on characteristics of the targeted speaker'"'"'s voice component during the filtering of the microphone signal.

11. The electronic device of claim 2, further comprising:

a radio frequency receiver that is configured to receive a wireless communication signal from a remote transmitter that is carried by the targeted speaker; and

a distance determination circuit that is configured to determine a signal strength of the received communication signal, and to estimate distance from the electronic device to the remote transmitter in response to the received signal strength,

wherein the microphone comprises a zoom/beam steerable microphone having a plurality of microphone elements and a controllable phase combiner circuit that adjusts relative phase between signals from the microphone elements in response to a phase control signal and combines the adjusted phase signals to generate the microphone signal; and

wherein the speaker recognition circuit is further configured to regulate the phase control signal in response to recognizing the presence of the targeted speaker'"'"'s voice in the microphone signal and the estimated distance from the electronic device to the remote transmitter to increase the targeted speaker'"'"'s voice component in the microphone signal relative to other components.

12. The electronic device of claim 2, further comprising:

an image sensor that is configured to generate an image signal; and

a distance determination circuit that is configured to recognize the presence of a person in the image signal, and to estimate distance from the electronic device to the person in response to size of the person relative to a plurality of threshold values,

wherein the microphone comprises a zoom/beam steerable microphone having a plurality of microphone elements and a controllable phase combiner circuit that adjusts relative phase between signals from the microphone elements in response to a phase control signal and combines the adjusted phase signals to generate the microphone signal; and

wherein the speaker recognition circuit is further configured to regulate the phase control signal in response to recognizing the presence of the targeted speaker'"'"'s voice in the microphone signal and the estimated distance from the electronic device to the recognized person to increase the targeted speaker'"'"'s voice component in the microphone signal relative to other components.

13. The electronic device of claim 2, further comprising an audio recorder that records the filtered microphone signal from the adaptive sound filter.

14. A method comprising:

computer-implemented recognition of a targeted speaker'"'"'s voice within a microphone signal and regulation of filtering of the microphone signal responsive to the recognition of the targeted speaker'"'"'s voice.

15. The method of claim 14, further comprising:

computer-implemented learning of characteristics of the targeted speaker'"'"'s voice component in the microphone signal during a training mode, and using the learned characteristics to recognize the presence of the targeted speaker'"'"'s voice in the microphone signal and to regulate a control signal to cause an adaptive sound filter circuit to adapt filtering of the microphone signal to increase the targeted speaker'"'"'s voice component of the microphone signal relative to other components.

16. The method of claim 15, further comprising:

computer-implemented learning characteristic differences over time in the frequency range of the targeted speaker'"'"'s voice in the microphone signal as the targeted speaker vocalizes a particular song/speech during the training mode; and

computer-implemented responding to the recognized presence of the targeted speaker'"'"'s voice in the microphone signal by using the learned characteristic differences over time in the frequency range of the targeted speaker'"'"'s voice to vary the control signal over time to dynamically tune the adaptive sound filter circuit to track the expected characteristic differences over time in the frequency range of the targeted speaker'"'"'s voice during the particular song/speech and to increase the targeted speaker'"'"'s voice component in the microphone relative to other components.

17. The method of claim 15, further comprising:

computer-implemented learning characteristic differences over time in the amplitude pattern of the targeted speaker'"'"'s voice in the microphone signal as the targeted speaker vocalizes a particular song/speech during the training mode; and

computer-implemented responding to the recognized presence of the targeted speaker'"'"'s voice in the microphone signal by using the learned characteristic differences over time in the amplitude pattern of the targeted speaker'"'"'s voice to vary the control signal over time to dynamically tune a gain of a variable gain amplifier in the adaptive sound filter circuit that amplifies the microphone signal to track the expected characteristic differences over time in the amplitude pattern of the targeted speaker'"'"'s voice during the particular song/speech.

18. The method of claim 15, further comprising:

computer-implemented learning timing of gaps where the targeted speaker'"'"'s voice in the microphone signal has less than a threshold amplitude as the targeted speaker vocalizes the particular song/speech during the training mode; and

computer-implemented responding to the recognized presence of the targeted speaker'"'"'s voice in the microphone signal by using the learned timing of gaps to vary the control signal over time to dynamically tune the adaptive sound filter circuit to increase the other components of the microphone signal relative to the component of the targeted speaker'"'"'s voice in the microphone signal during each of the gaps and to increase the component of the targeted speaker'"'"'s voice relative to the other components of the microphone signal in-between each of the gaps.

19. The method of claim 15, further comprising:

computer-implemented estimation of distance from an electronic device to a remote transmitter that is carried by the targeted speaker in response to strength of a signal that is received by the electronic device from the remote transmitter; and

computer-implemented adaptation of the learned characteristics of the targeted speaker'"'"'s voice in response to the estimated distance from the electronic device to the remote transmitter to compensate for expected effects of distance on characteristics of the targeted speaker'"'"'s voice component during the filtering of the microphone signal.

20. An electronic device comprising:

a microphone that generates a microphone signal responsive to incident sound; and

a speaker characterization circuit that is configured to operate in a training mode to learn, from the targeted speaker'"'"'s voice component in the microphone signal, time-varying differences in the frequency range and/or amplitude pattern of the targeted speaker'"'"'s voice as the targeted speaker vocalizes a particular song/speech;

an adaptive sound filter circuit that is configured to adaptively filter the microphone signal responsive to a control signal, wherein the adaptive sound filter circuit comprises a variable bandpass filter having bandpass frequencies that are varied in response to the control signal and/or a variable gain amplifier having a gain that is varied in response to the control signal; and

a speaker recognition circuit that is configured to use the learned time-varying differences in the frequency range and/or amplitude pattern of the targeted speaker'"'"'s voice to recognize the presence of the targeted speaker'"'"'s voice in the microphone signal vocalizing the particular song/speech, and to respond to the recognized presence by using the time-varying differences in the frequency range and/or the amplitude pattern of the targeted speaker'"'"'s voice to vary the control signal over time to dynamically tune the bandpass frequencies of the bandpass filter to follow the differences over time in the frequency range of the targeted speaker'"'"'s voice and/or to dynamically tune the gain of the variable gain amplifier to follow the differences over time in the amplitude pattern of the targeted speaker'"'"'s voice.