Speaking apparatus having handfree conversation function

A speaking apparatus having a handfree conversation function enabling speaking without holding a receiver in a hand by providing a microphone and a speaker in a casing of a system, said speaking apparatus having the handfree conversation function, comprising:

• a first echo canceler which cancels an echo caused by direct acoustic coupling between the speaker and the microphone positioned in the system comprised of the casing and a surface on which the casing is then emplaced;
• a speech switching circuit, connected to said first echo canceler, in which an insertion attenuation amount of a transmitted and received signal is set having an upper limit of an attenuation amount sufficient for cancellation of an indirect echo generated by indirect acoustic coupling caused by a location where the speaking apparatus is used,
• wherein the upper limit of the insertion attenuation amount of the speech switching circuit is set enabling two-way speaking;
• a second echo canceler connected to said speech switching circuit which cancels a tone echo generated by an anti-side tone circuit, wherein at least one of the first echo canceler and the second echo canceler comprises:
• an adaptive filter, wherein said adaptive filter comprises taps,
• coefficient correcting means, connected to said adaptive filter, for estimating, based on a difference between an input signal obtained at an input terminal receiving a reply and a pseudo echo generated based on an output signal using the adaptive filter, the adaptive filter using a filter input, and a coefficient of the adaptive filter to generate the pseudo echo, and
• coefficient selecting means for suspending a coefficient correction processing having a processing amount performed during a sampling period for the adaptive filter at a frequency in accordance with a predetermined rule reducing the processing amount performed during the sampling period, wherein said coefficient selecting means comprises suspending means for suspending the coefficient correction process for each of the taps of the adaptive filter;
• and
• a low pass filter having a low pass filter coefficient connected to said suspending means,
• wherein the suspending means for suspending the coefficient correction processing for each of the taps, comprises assumption means for assuming that an amplitude characteristic of an impulse response falls exponentially with respect to a predicted reverberation time of the unknown system and, when inputting a step input to the low pass filter having the low pass filter coefficient for suppressing an external disturbance, the suspending means suspends an estimation operation for each of the taps of the adaptive filter having a small response.

A speaking apparatus having a handfree conversation function enabling speaking without holding a receiver in a hand by providing a microphone and a speaker in a casing of a system, said speaking apparatus having the handfree conversation function, comprising:

• a first echo canceler which cancels an echo caused by direct acoustic coupling between the speaker and the microphone positioned in the system comprised of the casing and a surface on which the casing is then emplaced;
• a speech switching circuit, connected to said first echo canceler, in which an insertion attenuation amount of a transmitted and received signal is set having an upper limit of an attenuation amount sufficient for cancellation of an indirect echo generated by indirect acoustic coupling caused by a location where the speaking apparatus is used,
• wherein the upper limit of the insertion attenuation amount of the speech switching circuit is set enabling two-way speaking;
• a second echo canceler connected to said speech switching circuit which cancels a tone echo generated by an anti-side tone circuit, wherein at least one of the first echo canceler and the second echo canceler comprises:
• an adaptive filter, wherein said adaptive filter comprises sets of taps,
• coefficient correcting means, connected to said adaptive filter, for estimating, based on a difference between an input signal obtained at an input terminal receiving a reply and a pseudo echo generated based on an output signal using the adaptive filter, the adaptive filter using a filter input, and a coefficient of the adaptive filter to generate the pseudo echo, and
• coefficient selecting means for suspending a coefficient correction processing having a processing amount performed during a sampling period for the adaptive filter at a frequency in accordance with a predetermined rule reducing the processing amount performed during the sampling period, wherein said coefficient selecting means comprises suspending means for suspending a coefficient correction processing in units of sets of taps of the adaptive filter;
• and
• a low pass filter having a low pass filter coefficient connected to said suspending means,
• wherein the suspending means for suspending an operation in the units of sets of taps comprises assumption means for assuming that an amplitude characteristic of an impulse response falls exponentially with respect to a predicted reverberation time of the unknown system, said suspending means inputting a step input to the low pass filter having the low pass filter coefficient for suppressing external disturbances, the step input having one of an average value of one of the sets of taps of the adaptive filter and one of values in one of the sets of taps having an amplitude value, and suspends an estimation operation for the units of sets of taps having a small response.

A speaking apparatus having a handfree conversation function enabling speaking without holding a receiver in a hand by providing a microphone and a speaker in a casing of a system, said speaking apparatus having the handfree conversation function, comprising:

• a first echo canceler which cancels an echo caused by direct acoustic coupling between the speaker and the microphone positioned in the system comprised of the casing and a surface on which the casing is then emplaced;
• a speech switching circuit, connected to said first echo canceler, in which an insertion attenuation amount of a transmitted and received signal is set having an upper limit of an attenuation amount sufficient for cancellation of an indirect echo generated by indirect acoustic coupling caused by a location where the speaking apparatus is used,
• wherein the upper limit of the insertion attenuation amount of the speech switching circuit is set enabling two-way speaking;
• a second echo canceler connected to said speech switching circuit which cancels a tone echo generated by an anti-side tone circuit, wherein at least one of the first echo canceler and the second echo canceler comprises:
• an adaptive filter,
• coefficient correcting means, connected to said adaptive filter, for estimating, based on a difference between an input signal obtained at an input terminal receiving a reply and a pseudo echo generated based on an output signal using the adaptive filter, the adaptive filter using a filter input, and a coefficient of the adaptive filter to generate the pseudo echo, and
• coefficient selecting means for suspending a coefficient correction processing having a processing amount performed during a sampling period for the adaptive filter at a frequency in accordance with a predetermined rule reducing the processing amount performed during the sampling period;
• and
• a secondary noncyclic type filter having an output, connected to said speech switching circuit, and including a shift register and multiplier,
• wherein said coefficient selecting means comprises:
• transmitting means, provided in the speaking apparatus for transmitting the output signal to the unknown system with the unknown characteristics, and estimating a transmission function which the unknown system provides in a response, and forming an equivalent transmission function equivalent to the transmission function, and detecting that the output signal transmitted to the unknown system is a narrow band signal for indicating suspension of the estimation of the transmission function, and
• coefficient calculating means for calculating a coefficient value of the coefficient producing a minimum difference between the output of the secondary noncyclic type filter comprised of the shift register producing a delay of one sampling period and the multiplier which multiplies a separately determined coefficient by the filter input,
• the coefficient correcting means uses a fact indicating that a ratio between a level of a difference and a filter level of the filter input becomes less than a fixed value determined separately, adds to a differential output a primary cyclic filter having as a cyclic coefficient 1/2 of a primary coefficient of a secondary noncyclic type filter which is extracted, and uses an output for a calculation of a level ratio.

A speaking apparatus having a handfree conversation function enabling speaking without holding a receiver in a hand by providing a microphone and a speaker in a casing of a system, said speaking apparatus having the handfree conversation function, comprising:

• a first echo canceler which cancels an echo caused by direct acoustic coupling between the speaker and the microphone positioned in the system comprised of the casing and a surface on which the casing is then emplaced;
• a speech switching circuit, connected to said first echo canceler, in which an insertion attenuation amount of a transmitted and received signal is set having an upper limit of an attenuation amount sufficient for cancellation of an indirect echo generated by indirect acoustic coupling caused by a location where the speaking apparatus is used,
• wherein the upper limit of the insertion attenuation amount of the speech switching circuit is set enabling two-way speaking;
• a second echo canceler connected to said speech switching circuit which cancels a tone echo generated by an anti-side tone circuit, wherein at least one of the first echo canceler and the second echo canceler comprises:
• an adaptive filter,
• coefficient correcting means, connected to said adaptive filter, for estimating, based on a difference between an input signal obtained at an input terminal receiving a reply and a pseudo echo generated based on an output signal using the adaptive filter, the adaptive filter using a filter input, and a coefficient of the adaptive filter to generate the pseudo echo, and
• coefficient selecting means for suspending a coefficient correction processing having a processing amount performed during a sampling period for the adaptive filter at a frequency in accordance with a predetermined rule reducing the processing amount performed during the sampling period;
• a secondary noncyclic type filter having an output, connected to said speech switching circuit, and including a shift register and multiplier,
• wherein said coefficient selecting means comprises:
• transmitting means, provided in the speaking apparatus for transmitting the output signal to the unknown system with the unknown characteristics, and estimating a transmission function which the unknown system provides in a response, and forming an equivalent transmission function equivalent to the transmission function, and detecting that the output signal transmitted to the unknown system is a narrow band signal for indicating suspension of the estimation of the transmission function, and
• coefficient calculating means for calculating a coefficient value of the coefficient producing a minimum difference between the output of the secondary noncyclic type filter comprised of the shift register producing a delay of one sampling period and the multiplier which multiplies a separately determined coefficient by the filter input,
• the coefficient correcting means uses a fact indicating that a ratio of a level of a difference and a filter level of the filter input becomes less than a fixed value determined separately, and constitutes the secondary cyclic type filter from a separately determined constant determining a bandwidth, connects an input of the secondary noncyclic type filter to the filter input generating a resultant output, and uses the resultant output for a calculation of the ratio.