JP4333524B2 - Loudspeaker - Google Patents

Description

  The present invention relates to a loudspeaker communication device used for an interphone or the like.

  Conventionally, it is not necessary to have a handset during a call, and a speaker who is away from the calling device outputs a voice signal transmitted from the other party's calling device through a speaker, and the voice emitted by the calling party is output. 2. Description of the Related Art There is provided a loudspeaker device that implements a loudspeaker call (hands-free call) by collecting sound with a microphone and transmitting the collected sound to a counterpart call device. In such a loudspeaker, a part of the voice uttered by the caller is reflected due to acoustic coupling from the speaker of the other call device to the microphone or impedance mismatch between the call device and the transmission path. For this reason, there may be a case where feedback is again superimposed on the received signal, and if the level of the feedback component is high, the caller hears it as an unpleasant echo (acoustic echo or line echo). In addition, when a closed loop is formed in the communication system due to the above acoustic coupling and reflection, and acoustic coupling at the terminal itself, and there is a frequency component in which the loop gain of the closed loop exceeds one time, howling occurs at that frequency, and stable It becomes impossible to continue the call. Therefore, how to suppress the above-mentioned unpleasant echo and howling is an important issue in designing a loudspeaker device as a telephone device.

  Conventionally, a voice switch that always estimates the call state (transmission state, reception state, etc.) and inserts losses into the transmission path and reception path with appropriate distribution based on the estimation results. A method of reducing closed loop loop gain and suppressing unpleasant echoes and howling has been widely used. FIG. 19 is a block diagram showing a conventional example of a so-called hands-free intercom comprising an interphone master unit (hereinafter abbreviated as “master unit”) M as a loudspeaker device and a doorphone slave unit S as a counterpart communication device. (See Patent Document 1). Base unit M includes microphone 1, speaker 2, two-wire / four-wire conversion circuit 30, microphone amplifier 31 that amplifies the transmission signal output from microphone 1, and reception signal that arrives from the other party's telephone equipment via the transmission system. Is composed of a line amplifier 32 and a voice switch VS ′. Moreover, although illustration is abbreviate | omitted, the doorphone subunit | mobile_unit S is comprised with a microphone, a speaker, a 2 wire | wire 4 line | wire conversion circuit, etc.

The voice switch VS ′ includes transmission side loss insertion means 33 for inserting loss into the transmission side signal path from the microphone 1 through the microphone amplifier 31 to the two-wire / four-wire conversion circuit 30, and the line amplifier 32 to the speaker 2. Receiving-side loss insertion means 34 for inserting a loss in the receiving-side signal path leading to, and insertion loss amount control means 35 for controlling the insertion loss amount in each of the transmission-side and reception-side loss insertion means 33, 34. . This insertion loss amount control means 35 estimates the power of, for example, a transmission signal and a reception signal, compares the estimated values, and compares the estimated values with the loss insertion means 33 and 34 on the side having a smaller instantaneous power. A process of switching between the transmission state and the reception state by inserting a loss is performed.
JP 2000-307745 A

  However, in the above conventional example, when the difference between the ambient noise level on the far end (doorphone slave unit S) side and the ambient noise level on the near end (master unit M) side is large, for example, the doorphone slave unit S installed outdoors. When a large noise such as wind noise or car noise is input to the microphone, the insertion loss amount control means 35 that monitors the transmission signal and the reception signal to estimate the call state, for example, has a far-end ambient noise level. In a large situation, it is always judged as a reception state, and in a situation where the ambient noise level at the near end is high, it is always judged as a transmission state, and it is either a reception state or a transmission state regardless of the actual call state. On the other hand, a phenomenon of fixing the call state (so-called voice switch one-sided fall) occurs.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a loudspeaker device capable of suppressing a fall of a voice switch.

In order to achieve the above-mentioned object, the invention of claim 1 is a microphone and a speaker, a transmission state for transmitting a transmission signal to the transmission system and attenuating the reception signal, a transmission state of the reception signal to the speaker, and attenuating the transmission signal. A voice switch that switches between a receiving state and a receiving side signal path. The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side signal path from the transmission system to the speaker. A receiving side loss insertion unit that inserts a loss into the voice signal, a first voice zone detection unit that detects a voice zone of the transmission signal, a second voice zone detection unit that detects a voice zone of the reception signal, and a transmission signal Transmitting-side instantaneous power estimation unit for estimating the instantaneous power of the receiving signal, transmitting-side background noise power estimating unit for estimating the background noise power of the transmitting signal, and receiving-side instantaneous power estimating the instantaneous power of the receiving signal An estimation unit, a reception side background noise power estimation unit for estimating the background noise power of the reception signal, and an insertion loss amount control unit for controlling the insertion loss amount in each loss insertion unit on the transmission side and reception side, The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result A first attenuator for attenuating the instantaneous power estimate on the transmitting side and a first attenuation for increasing the attenuation of the first attenuator when the estimated background noise power on the transmitting side exceeds a predetermined threshold Quantity control unit, or A second attenuator for attenuating the talk side instantaneous power estimate and a second attenuator controller for increasing the attenuator of the second attenuator when the receiver background noise power estimate exceeds a predetermined threshold characterized by including a voice switch high-pass filter for removing at least one comprises one Rutotomoni low frequency band component than the main component band of a speech from the receiving signal to be referenced in the second speech section detecting unit of And

According to this invention, when the ambient noise level on the far end side (or near end side) is high, the instantaneous power estimation value on the receiving side (or transmitting side) also becomes large, so that the insertion loss amount control unit performs a call. The voice switch may fall down to the receiving side (or sending side) by misjudging the state as the receiving state (or sending state). In such a case, the background of the receiving side (or sending side) If the noise power estimate exceeds a predetermined threshold, the instantaneous power estimate on the receiver side (or transmitter side) is decreased by increasing the attenuation amount of the second (or first) attenuator. Therefore, it is possible to prevent the insertion loss amount control unit from erroneously determining the call state as the reception state (or the transmission state) and to suppress the fall of the voice switch. Furthermore, by removing noise having a frequency band component lower than the main component band of the voice from the received signal, erroneous detection of the voice section can be prevented and the instantaneous power estimate on the receiving side for the noise can be reduced. As a result, it is possible to prevent the voice switch from falling over.

According to a second aspect of the present invention, in the first aspect of the invention, there is provided means for specifying the other party's telephone device that communicates via the transmission system, and switching between valid / invalid of the high-pass filter according to the identified telephone device. It is characterized by that.

  According to the present invention, when there is a telephone apparatus installed outdoors and a telephone apparatus installed indoors, the high-pass filter is enabled when calling with a telephone apparatus installed outdoors. Therefore, the voice switch can be prevented from falling down by preventing the false detection of the voice section by removing the noise and reducing the instantaneous power estimation value on the receiver side for the noise, and the high-pass filter is disabled when talking to the telephone equipment installed indoors. By doing so, it is possible to maintain a balance of switching of the call state in the voice switch.

The invention of claim 3 is characterized in that, in the invention of claim 1 , there is provided means for taking in an operation input by a user's operation and switching valid / invalid of a high-pass filter in accordance with the operation input.

  According to the present invention, the user intentionally switches between enabling and disabling the high-pass filter, thereby preventing erroneous detection of the voice section and reducing the instantaneous power estimate on the receiving side for the noise, thereby suppressing the voice switch from falling over. As well as being able to improve usability.

In order to achieve the above-mentioned object , the invention according to claim 4 is a microphone and a speaker, a transmission state for transmitting a transmission signal to the transmission system and attenuating the reception signal, a transmission state of the reception signal to the speaker, and attenuating the transmission signal. A voice switch that switches between a receiving state and a receiving side signal path. The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side signal path from the transmission system to the speaker. A receiving side loss insertion unit that inserts a loss into the voice signal, a first voice zone detection unit that detects a voice zone of the transmission signal, a second voice zone detection unit that detects a voice zone of the reception signal, and a transmission signal Transmitting-side instantaneous power estimation unit for estimating the instantaneous power of the receiving signal, transmitting-side background noise power estimating unit for estimating the background noise power of the transmitting signal, and receiving-side instantaneous power estimating the instantaneous power of the receiving signal An estimation unit, a reception side background noise power estimation unit for estimating the background noise power of the reception signal, and an insertion loss amount control unit for controlling the insertion loss amount in each loss insertion unit on the transmission side and reception side, The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result A first attenuator for attenuating the instantaneous power estimate on the transmitting side and a first attenuation for increasing the attenuation of the first attenuator when the estimated background noise power on the transmitting side exceeds a predetermined threshold Quantity control unit, or A second attenuator for attenuating the talk side instantaneous power estimate and a second attenuator controller for increasing the attenuator of the second attenuator when the receiver background noise power estimate exceeds a predetermined threshold And the first and second speech section detection units include a first instantaneous power estimation unit that estimates the instantaneous power of the reference signal, and the power of background noise that is steadily present in the reference signal. A background noise power estimator for estimating the reference signal, a first determination unit for determining a speech section of the reference signal based on a ratio of the instantaneous power estimate value and the background noise power estimate value, A low-pass filter for removing high frequency band components, a second instantaneous power estimation unit for estimating the instantaneous power of the reference signal after the high frequency band components have been removed by the low-pass filter, and a first determination unit Is determined as a non-voice segment Is determined as a non-speech interval, and when the first determination unit determines a speech interval, the instantaneous power estimation value estimated by the first instantaneous power estimation unit is multiplied by a positive coefficient less than 1. And a second determination unit that determines whether or not the voice segment is based on a magnitude relationship between the value and the instantaneous power estimation value estimated by the second instantaneous power estimation unit.

According to this invention, when the ambient noise level on the far end side (or near end side) is high, the instantaneous power estimation value on the receiving side (or transmitting side) also becomes large, so that the insertion loss amount control unit performs a call. The voice switch may fall down to the receiving side (or sending side) by misjudging the state as the receiving state (or sending state). In such a case, the background of the receiving side (or sending side) If the noise power estimate exceeds a predetermined threshold, the instantaneous power estimate on the receiver side (or transmitter side) is decreased by increasing the attenuation amount of the second (or first) attenuator. Therefore, it is possible to prevent the insertion loss amount control unit from erroneously determining the call state as the reception state (or the transmission state) and to suppress the fall of the voice switch. Furthermore, it is possible to prevent the first and second voice section detection units from erroneously detecting the voice section due to noise having a frequency band component higher than the main component band of the voice.

The invention of claim 5 is characterized in that, in the invention of claim 4 , the coefficient in the second determination section is variable.

  According to the present invention, it is possible to optimize the detection operation of the first and second voice section detection units by changing the coefficient according to the intention of the other party's telephone device or the user.

According to a sixth aspect of the present invention, in the fourth or fifth aspect of the present invention, the first determination unit included in the first speech section detection unit is estimated by the first instantaneous power estimation unit with a predetermined time interval. The absolute value of the difference between the two instantaneous power estimation values is obtained, and the determination is made with reference to a comparison result between the absolute value of the difference and a predetermined threshold value.

  According to the present invention, it is possible to prevent erroneous detection of a section including noise with little temporal fluctuation of instantaneous power among non-stationary noises other than speech as a speech section.

In order to achieve the above-mentioned object , the invention according to claim 7 provides a microphone and a speaker, a transmission state for transmitting a transmission signal to the transmission system and attenuating the reception signal, a transmission state of the reception signal to the speaker, and attenuating the transmission signal. A voice switch that switches between a receiving state and a receiving side signal path. The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side signal path from the transmission system to the speaker. A receiving side loss insertion unit that inserts a loss into the voice signal, a first voice zone detection unit that detects a voice zone of the transmission signal, a second voice zone detection unit that detects a voice zone of the reception signal, and a transmission signal Transmitting-side instantaneous power estimation unit for estimating the instantaneous power of the receiving signal, transmitting-side background noise power estimating unit for estimating the background noise power of the transmitting signal, and receiving-side instantaneous power estimating the instantaneous power of the receiving signal An estimation unit, a reception side background noise power estimation unit for estimating the background noise power of the reception signal, and an insertion loss amount control unit for controlling the insertion loss amount in each loss insertion unit on the transmission side and reception side, The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result A first attenuator for attenuating the instantaneous power estimate on the transmitting side and a first attenuation for increasing the attenuation of the first attenuator when the estimated background noise power on the transmitting side exceeds a predetermined threshold Quantity control unit, or A second attenuator for attenuating the talk side instantaneous power estimate and a second attenuator controller for increasing the attenuator of the second attenuator when the receiver background noise power estimate exceeds a predetermined threshold The voice switch includes a high-pass filter that removes a frequency band component lower than the main component band of the voice from the received signal referred to by the insertion loss amount control unit, and the first voice section The detector includes a first instantaneous power estimator that estimates the instantaneous power of the reference signal, a background noise power estimator that estimates the power of background noise that is constantly present in the reference signal, an instantaneous power estimate, and a background A first determination unit that determines a speech section of a reference signal based on a ratio of noise power estimation values, a low-pass filter that removes a frequency band component higher than a main component band of speech from the reference signal, and a low-pass filter A second instantaneous power estimator that estimates the instantaneous power of the reference signal after the high frequency band component is removed in step 1 and when the first determination unit determines that it is a non-voice interval, When the first determination unit determines that the speech section is present, the second instantaneous power estimation unit estimates the value obtained by multiplying the instantaneous power estimation value estimated by the first instantaneous power estimation unit by a positive coefficient less than 1. And a second determination unit that determines whether or not the voice section is based on the magnitude relationship with the instantaneous power estimation value.

According to this invention, when the ambient noise level on the far end side (or near end side) is high, the instantaneous power estimation value on the receiving side (or transmitting side) also becomes large, so that the insertion loss amount control unit performs a call. The voice switch may fall down to the receiving side (or sending side) by misjudging the state as the receiving state (or sending state). In such a case, the background of the receiving side (or sending side) If the noise power estimate exceeds a predetermined threshold, the instantaneous power estimate on the receiver side (or transmitter side) is decreased by increasing the attenuation amount of the second (or first) attenuator. Therefore, it is possible to prevent the insertion loss amount control unit from erroneously determining the call state as the reception state (or the transmission state) and to suppress the fall of the voice switch. Furthermore, by removing noise having a frequency band component lower than the main component band of speech from the received signal using a high-pass filter, erroneous detection of the speech section can be prevented and the instantaneous power estimation on the receiving side for the noise The value can be reduced, and as a result, the fall of the voice switch can be suppressed. Further, by removing noise having a frequency band component higher than the main component band of speech using a low-pass filter, erroneous detection of the speech section in the first speech section detection unit can be prevented.

The invention according to claim 8 is the invention according to claim 7 , wherein the first determination unit is an absolute value of a difference between the two instantaneous power estimation values estimated by the first instantaneous power estimation unit at a predetermined time interval. And determining with reference to a comparison result between the absolute value of the difference and a predetermined threshold value.

  According to the present invention, it is possible to prevent erroneous detection of a section including noise with little temporal fluctuation of instantaneous power among non-stationary noises other than speech as a speech section.

In order to achieve the above object , the invention according to claim 9 provides a microphone and a speaker, a transmission state for transmitting a transmission signal to the transmission system and attenuating the reception signal, a transmission state of the reception signal to the speaker, and attenuating the transmission signal. A voice switch that switches between a receiving state and a receiving side signal path. The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side signal path from the transmission system to the speaker. A receiving side loss insertion unit that inserts a loss into the voice signal, a first voice zone detection unit that detects a voice zone of the transmission signal, a second voice zone detection unit that detects a voice zone of the reception signal, and a transmission signal Transmitting-side instantaneous power estimation unit for estimating the instantaneous power of the receiving signal, transmitting-side background noise power estimating unit for estimating the background noise power of the transmitting signal, and receiving-side instantaneous power estimating the instantaneous power of the receiving signal An estimation unit, a reception side background noise power estimation unit for estimating the background noise power of the reception signal, and an insertion loss amount control unit for controlling the insertion loss amount in each loss insertion unit on the transmission side and reception side, The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result A first attenuator for attenuating the instantaneous power estimate on the transmitting side and a first attenuation for increasing the attenuation of the first attenuator when the estimated background noise power on the transmitting side exceeds a predetermined threshold Quantity control unit, or A second attenuator for attenuating the talk side instantaneous power estimate and a second attenuator controller for increasing the attenuator of the second attenuator when the receiver background noise power estimate exceeds a predetermined threshold A low-pass filter that removes a frequency band component higher than the main component band of the speech from the transmission signal referred to by the first speech section detection unit, and a second speech section detection unit The voice switch is provided with a high-pass filter that removes a frequency band component lower than the main component band of the voice from the received reception signal.

According to this invention, when the ambient noise level on the far end side (or near end side) is high, the instantaneous power estimation value on the receiving side (or transmitting side) also becomes large, so that the insertion loss amount control unit performs a call. The voice switch may fall down to the receiving side (or sending side) by misjudging the state as the receiving state (or sending state). In such a case, the background of the receiving side (or sending side) If the noise power estimate exceeds a predetermined threshold, the instantaneous power estimate on the receiver side (or transmitter side) is decreased by increasing the attenuation amount of the second (or first) attenuator. Therefore, it is possible to prevent the insertion loss amount control unit from erroneously determining the call state as the reception state (or the transmission state) and to suppress the fall of the voice switch. Further, by removing noise having a frequency band component higher than the main component band of the voice using a low-pass filter, erroneous detection of the voice section in the first voice section detection unit can be prevented and reception of the noise is received. The instantaneous power estimation value on the side can be reduced, and as a result, the voice switch can be prevented from falling down. In addition, by removing noise having a frequency band component lower than the main component band of speech from the received signal, erroneous detection of the speech section can be prevented and the instantaneous power estimate on the receiving side for the noise can be reduced. As a result, it is possible to prevent the voice switch from falling over.

According to a tenth aspect of the present invention, in the ninth aspect of the invention, the low-pass filter having a variable cut-off frequency is provided in a voice switch, and an operation input by a user's operation is taken in and a low frequency is set according to the operation input. A means for changing the cut-off frequency of the pass filter is provided.

  According to the present invention, since an appropriate cutoff frequency can be set according to the use environment, it is possible to simultaneously prevent erroneous detection of voice sections and improve usability.

The invention of claim 11 is the invention of claim 9 , further comprising means for detecting the pitch of the voice included in the reference signal and changing the cutoff frequency of the low-pass filter in accordance with the detected voice pitch. It is characterized by.

  According to the present invention, it is possible to prevent an infant's voice whose voice pitch is generally higher than that of an adult voice from being removed by the low-pass filter.

The invention of claim 12 is characterized in that, in the invention of claim 9 , a high-pass filter for removing frequency band components lower than the main component band of speech is provided in series with the low-pass filter.

  According to the present invention, it is possible to prevent erroneous detection of a speech section by removing noise having frequency band components other than the main component band of speech from a transmission signal using a high-pass filter and a low-pass filter. It is possible to reduce the instantaneous power estimation value on the transmission side with respect to the noise, and as a result, it is possible to suppress the fall of the voice switch.

A thirteenth aspect of the invention is characterized in that, in the twelfth aspect of the invention, the filter for removing frequency band components higher or lower than the main component region of speech from the transmitted signal and the received signal is constituted by a digital filter.

  According to the present invention, a filter having desired characteristics can be easily realized without changing the circuit configuration.

  According to the present invention, when the ambient noise level on the far end side (or near end side) is high, the instantaneous power estimation value on the receiving side (or transmitting side) also increases, so that the insertion loss amount control unit can The voice switch may fall down to the receiving side (or sending side) by misjudging the state as the receiving state (or sending state). In such a case, the background of the receiving side (or sending side) If the noise power estimate exceeds a predetermined threshold, the instantaneous power estimate on the receiver side (or transmitter side) is decreased by increasing the attenuation amount of the second (or first) attenuator. Therefore, there is an effect that it is possible to prevent the voice switch from falling down by preventing the insertion loss amount control unit from erroneously determining the call state as the reception state (or transmission state).

Before describing embodiments of the present invention, reference examples of the present invention will be described.
The loudspeaker device of this reference example is common to the conventional loudspeaker device (interphone master M) in that it includes a microphone 1, a speaker 2, a microphone amplifier 5, a line amplifier 6, and a voice switch VS as shown in FIG. To do. A line amplifier 7 is inserted in the transmitting signal path between the voice switch VS and the 2-wire / four-wire conversion circuit 30, and a speaker amplifier is connected in the receiving signal path between the voice switch VS and the speaker 2. 8 is inserted.

The voice switch VS in this reference example includes a transmission side loss insertion unit 3 inserted into a transmission side signal path for transmitting a transmission signal to a line, and a reception side signal for transmitting a reception signal to the speaker 2. A receiving side loss insertion unit 4 to be inserted into the path, a first voice zone detection unit 11 for detecting a voice zone of a transmission signal, a second voice zone detection unit 12 for detecting a voice zone of a reception signal, An insertion loss amount control unit 10 that controls the insertion loss amount of the transmission side loss insertion unit 3 and the reception side loss insertion unit 4 according to the call state is provided. Note that both the transmission side loss insertion unit 3 and the reception side loss insertion unit 4 are composed of amplifiers, and the insertion loss amount control unit 10 controls each amplification degree G _T , G _R (the loss amounts are −G _T , − G _R ) is variable.

  The first and second voice section detection units 11 and 12 have a common configuration, and as shown in FIG. 2, a reference signal (a transmission signal input to the transmission side loss insertion unit 3 or a reception side loss insertion unit 4). The instantaneous power of the reference signal), the background noise power estimation unit 21 for estimating the background noise power of the reference signal, and the ratio of the instantaneous power estimation value Ps and the background noise power estimation value Pn. The first determination unit 22 compares (= Ps / Pn) with a predetermined threshold value and determines that the voice section is used when the ratio is equal to or greater than the threshold value. When the first and second voice section detection units 11 and 12 detect the voice section, the voice section detection signals TSD and RSD on the transmission side and the reception side are set to “1”, respectively, and the voice section is not detected. In a non-voice section, the voice section detection signals TSD and RSD are set to “0”. The instantaneous power estimation unit 20 on the transmitting side and the receiving side is configured by a filter having a characteristic that the rise is steep and the fall is gentle, and the background noise power estimation unit 21 on the transmission side and the reception side is a rise. Is composed of a filter or the like having a characteristic of a gradual fall and a steep fall.

  Further, the insertion loss amount control unit 10 returns from the input point to the transmission side loss insertion unit 3 to the input point to the reception side loss insertion unit 4 through the transmission side loss insertion unit 3 and the loop on the line side. A line feedback gain multiplier (not shown) whose coefficient is a value determined in accordance with the gain of the receiver, a receiving side loss inserting unit 4 and an acoustic side (microphone 1 and speaker) from the input point to the receiving side loss inserting unit 4 An acoustic coupling gain multiplier (not shown) having as a coefficient a value determined in accordance with the gain of the path that reaches the input point to the transmission side loss insertion unit 3 through the wraparound in 2); The receiving side instantaneous power estimate value Ps (R) output from the speech segment detector 12 is input to the acoustic coupling gain multiplier and the transmission side output from the first speech segment detector 11. A first comparator for comparing the magnitude relationship with the instantaneous power estimated value Ps (T) (see FIG. And the magnitude relationship between the output signal P1 obtained by inputting the transmission side instantaneous power estimation value Ps (T) to the line feedback gain multiplier and the reception side instantaneous power estimation value Ps (R). 2 comparators (not shown), the output signals C1 and C2 of the first comparator and the second comparator, and the output signals C3 of the first speech section detector 11 and the second speech section detector 12 (= TSD), an insertion loss amount distribution processing unit (not shown) that determines a call state based on C4 (= RSD) and controls the loss loss amount of the transmission side loss insertion unit 3 and the reception side loss insertion unit 4 A). Here, the output signal C1 of the first comparator becomes “0” when Ps (T) <P2, and becomes “1” when Ps (T) ≧ P2. Further, the output signal C2 of the second comparator becomes “0” when Ps (R) ≧ P1, and becomes “1” when Ps (R) <P1.

  Thus, the insertion loss amount control unit 10 determines the call state by referring to the four binary signals C1 to C4, and the insertion loss amounts of the transmission side loss insertion unit 3 and the reception side loss insertion unit 4 are determined. To decide. Here, when C1 = C2 = 1 and C3 = 1, the transmission mode, when C1 = C2 = 0 and C4 = 1, the reception mode, and when C1 ≠ C2 and C3 and C4 are not 0, high-speed idle In the mode and other states, the mode is determined as the slow idle mode. When the determination result is the transmission mode, the insertion loss amount of the transmission side loss insertion unit 3 is the minimum value, and the insertion loss amount of the reception side loss insertion unit 4 is the maximum value. When the determination result is the reception mode, the insertion loss amount of the transmission side loss insertion unit 3 is set to the maximum value, and the insertion loss amount of the reception side loss insertion unit 4 is set to the minimum value. Sometimes the transmission loss insertion unit 3 and the reception loss insertion unit 4 have the same insertion loss amount in a short transition time, and when the determination result is the slow idle mode, the transmission loss insertion unit 3 and the reception loss insertion in a long transition time. Part Equal to each other in the insertion loss. Note that the configuration and operation of the voice switch VS described above are the same as those disclosed in Patent Document 1, and thus detailed description thereof is omitted.

Next, the gist of this reference example will be described. In this reference example , the attenuator 13 that attenuates the instantaneous power estimate value Ps (R) on the receiver side and the attenuation of the attenuator 13 when the background noise power estimate value Pn (R) on the receiver side exceeds a predetermined threshold value. A feature resides in that the voice switch VS is provided with an attenuation amount control unit 14 for increasing the amount.

As shown in FIG. 3, the attenuation amount control unit 14 calculates an average value E [Pn (R)] of the receiver side background noise power estimated value Pn (R), and the average value E [Pn (R)] is compared with a predetermined threshold value. When the average value E [Pn (R)] is equal to or less than the threshold value, the attenuation amount of the attenuator 13 is set to the initial value (= 0 dB), and the average value is set. When E [Pn (R)] exceeds the threshold value, the attenuation amount determination unit 14b that increases the attenuation amount to a value larger than the initial value, and the reception side speech section detection signal RSD is “1”, that is, the second speech. When a voice section is detected by the section detection unit 12, or when the insertion loss amount (gain G _T ) of the transmission side loss insertion unit 3 is less than a predetermined threshold (gain G _T is greater than the threshold). An average calculation update / stop determination unit 14c that sometimes stops the average value calculation (update) by the average value calculation unit 14a, and a decrease by the attenuation amount determination unit 14b Configured to determination of the amount in a counter 14d for intermittently performed every predetermined time.

  Thus, when the ambient noise level on the far end side is large, the receiving side instantaneous power estimate value Ps (R) also becomes large, so that the insertion loss amount control unit 10 erroneously determines the call state as the receiving state. The voice switch Vs may fall to the receiver side, but if the ambient noise level at the far end becomes large and the receiver background noise power estimate Pn (R) exceeds the threshold value, the attenuation control is performed. Since the attenuation amount of the attenuator 13 is increased from the initial value by the unit 14 and the reception side instantaneous power estimated value Ps (R) is decreased, the insertion loss amount control unit 10 is prevented from erroneously determining the call state as the reception state. Therefore, it is possible to prevent the voice switch VS from falling down. In addition, since the average calculation update / stop determination unit 14c stops the calculation of the average value by the average value calculation unit 14a when the voice segment is detected by the second voice segment detection unit 12, the voice is received in the received signal. Therefore, it is possible to prevent the insertion loss amount control unit 10 from erroneously switching between the transmission mode and the reception mode by changing the attenuation amount of the attenuator 13. Furthermore, since the average calculation update / stop determination unit 14c stops the calculation of the average value by the average value calculation unit 14a when the insertion loss amount of the transmission side loss insertion unit 3 is less than a predetermined threshold, Since the attenuation amount of the attenuator 13 is changed when the signal wraps around the transmission side signal path, the insertion loss amount control unit 10 can be prevented from erroneously switching between the transmission mode and the reception mode.

By the way, as shown in FIG. 4, the applicant inserts an attenuator 13 ′ into the path for inputting the received signal to the second speech section detecting unit 12, and the received background noise power estimate Pn (R) is a predetermined value. An application has already been filed for a loudspeaker device in which the attenuation control unit 14 ′ increases the attenuation amount of the attenuator 13 ′ when the threshold value is exceeded (see Japanese Patent Application 2003-2003). 394670). However, in the case shown in FIG. 4, since the feedback loop is formed by the attenuator 13 ′, the second voice section detector 12 and the attenuation controller 14 ′, the attenuation of the attenuator 13 ′ increases. When the received signal is attenuated, the receiving side background noise power estimated value Pn (R) decreases and falls below the threshold value, and the receiving side background noise power estimated value Pn (R) falls below the threshold value so that the attenuation amount control unit 14 ′ decreases the attenuation amount of the attenuator 13 ′, and therefore the attenuation amount control unit 14 ′ repeatedly increases and decreases the attenuation amount of the attenuator 13 ′, and the operation of the insertion loss amount control unit 10 may become unstable. was there. On the other hand, in the present reference example , since the receiving side instantaneous power estimate value Ps (R) is attenuated by the attenuator 13 and no feedback loop is formed, the attenuation control unit 14 ′ attenuates as described above. There is an advantage that the operation of the insertion loss amount control unit 10 can be stabilized without repeating the increase and decrease of the attenuation amount of the device 13 '.

In this reference example , the attenuator 13 and the attenuation amount control unit 14 are provided on the reception side. However, the attenuator for attenuating the transmission side instantaneous power estimation value Ps (T) output from the first speech section detection unit 11. If the transmission side background noise power estimated value Pn (T) exceeds a predetermined threshold value, an attenuation amount control unit for increasing the attenuation amount of the attenuator is provided. It is possible to prevent the voice switch VS from collapsing by preventing erroneous determination of the call state in the unit 10, and an attenuator and an attenuation amount control unit are provided not only on either the transmitting side or the receiving side but on both sides. It doesn't matter.

(Embodiment 1 )
FIG. 5 shows a block diagram of the present embodiment. However, since the basic configuration of the present embodiment is the same as that of the reference example , the same components are denoted by the same reference numerals, description thereof is omitted, and only the configuration that is a feature of the present embodiment will be described.

  The present embodiment is characterized in that the voice switch VS includes a high-pass filter (HPF) 15 that removes a frequency band component lower than the main component band of the voice from the received signal referred to by the insertion loss amount control unit 10. is there.

  For example, since the door phone slave unit S, which is the other party's call device, is installed outdoors, the reception signal transmitted from the door phone slave unit S may include a so-called wind noise. Since such wind noise generally includes many frequency band components that are lower than the main component band of speech, the high-pass filter 15 can remove wind noise included in the received signal to some extent.

  Thus, since the received voice signal after the wind noise has been removed to some extent by the high-pass filter 15 is referred to by the second voice segment detector 12, the second voice segment detector 12 detects the wind noise. It is possible to prevent erroneous detection of the voice section, and to reduce the instantaneous instantaneous power estimate value Ps (R) for the wind noise, and as a result, it is possible to suppress the fall of the voice switch VS. .

  However, in a situation where the level of wind noise is low or almost zero, or in a situation where the other party's call device is installed indoors, the second low frequency component of the received signal is removed by the high pass filter 15. An erroneous detection of the voice section detection unit 12 may occur. Therefore, the communication device of the other party to be called via the transmission system is specified, and the means for switching the high pass filter 15 between valid and invalid according to the specified call device, or the operation input by the operation of the user is taken in and It is desirable to provide means for switching between valid / invalid of the high-pass filter 15 according to the operation input. For example, in the loudspeaker device according to the present embodiment, a control unit (not shown) made of a microcomputer detects a call from a partner telephone device and switches the transmission line for each individual telephone device. The caller identifies the calling device of the caller and enables the high-pass filter 15 to be effective during a call with a call device installed outdoors, thereby preventing erroneous detection of a voice section by noise removal and receiving-side instantaneous power against the noise. By reducing the estimated value Ps (R), it is possible to prevent the voice switch VS from collapsing, and the voice switch VS can be used to switch the call state by disabling the high-pass filter 15 when talking to a telephone device installed indoors. Balance can be maintained. Alternatively, a push button is provided on the housing (not shown) of the loudspeaker apparatus according to the present embodiment, and when the user presses the push button, an operation input is taken into the control unit, and the control unit is set according to the operation input. If the high pass filter 15 is switched between valid / invalid, the user intentionally switches the valid / invalid of the high pass filter 15 to prevent erroneous detection of the speech section and to receive instantaneous power for the noise. By reducing the estimated value Ps (R), the voice switch VS can be prevented from falling down and the usability can be improved. In order to switch the high-pass filter 15 between valid and invalid, for example, a signal path that bypasses the high-pass filter 15 is provided between the reference point of the receiver-side signal path and the second speech section detection unit 12. What is necessary is just to open and close the contact inserted in the middle of the signal path.

(Embodiment 2 )
FIG. 6 shows a block diagram of the first and second speech section detection units 11 ′ and 12 ′ in the present embodiment. However, since the basic configuration and other configurations of the first and second speech section detection units 11 ′ and 12 ′ are the same as those in the reference example , common components are denoted by the same reference numerals and illustrated and described as appropriate. Is omitted.

  The first and second speech section detection units 11 ′ and 12 ′ in the present embodiment include an instantaneous power estimation unit 20, a background noise power estimation unit 21, a first determination unit 22, and a reference signal (transmission signal). Or a low-pass filter (LPF) 23 that removes a frequency band component higher than the main component band of the voice from the received signal) and the instantaneous power of the reference signal after the high-frequency band component is removed by the low-pass filter 23 The second instantaneous power estimation unit 24 for estimating the instantaneous power estimation value Ps estimated by the instantaneous power estimation unit 20, the detection flag indicating the determination result by the first determination unit 22, and the second instantaneous power estimation unit 24 And a second determination unit 25 that detects a speech section based on the instantaneous power estimation value Ps_L estimated in (1). The first determination unit 22 determines that the voice section is a voice flag when the ratio (= Ps / Pn) of the instantaneous power estimated value Ps and the background noise power estimated value Pn is equal to or greater than a predetermined threshold, and a detection flag (voice section detection signal). ) Is set to “1”, and when the ratio is less than the threshold value, it is determined as a non-speech interval and the detection flag is set to “0”.

  On the other hand, as shown in the flowchart of FIG. 7, the second determination unit 25 has a detection flag output from the first determination unit 22 of “0”, that is, the first determination unit 22 determines that it is a non-speech segment. Is immediately determined as a non-speech interval, the output flag (speech interval detection signals TSD, RSD) is set to “0”, and the detection flag output from the first determination unit 22 is “1”, that is, the first determination unit 22. , The instantaneous power estimation value Ps estimated by the instantaneous power estimation unit 20 is multiplied by a predetermined coefficient γ (0 <γ <1) (= Ps · γ) and the second instantaneous time. The instantaneous power estimated value Ps_L estimated by the power estimator 24 is compared. If Ps · γ <Ps_L, the speech section is determined, and if Ps · γ ≧ Ps_L, the non-voice section is determined and the output flag is set to “1”. A determination process for switching to “0” is performed.

That is, in an environment in which noise having a frequency band component higher than a main component band of speech such as a ringing tone of a telephone or an animal (particularly a small dog) is present, Although there is a possibility that the second speech segment detection units 11 and 12 may erroneously detect the speech segment, the first or second speech segment detection units 11 ′ and 12 ′ of the present embodiment use the main component of speech. The instantaneous power estimated value Ps_L of the reference signal after the noise having a frequency band component higher than the band is removed by the low-pass filter 23 is estimated by the second instantaneous power estimating unit 24, and the estimated value Ps_L Based on the instantaneous power estimation value Ps of the reference signal before removing the frequency band component higher than the main component band and the determination result by the first determination unit 22, the second determination unit 25 comprehensively selects the voice section. Is judging As described above, it is possible to prevent the first and second voice section detection units 11 ′ and 12 ′ from erroneously detecting the voice section due to noise having a frequency band component higher than the main component band of the voice. . Note that the coefficient γ multiplied by the instantaneous power estimation value Ps estimated by the instantaneous power estimation unit 20 in the determination process of the second determination unit 25 is considered to be different from the optimum value for each counterpart telephone equipment. If γ is variable and, for example, is automatically set to an optimum value for each call device as described in the first embodiment, or a value is set by user operation, the first and first It is possible to optimize the detection operation of the two voice section detection units 11 ′ and 12 ′.

(Embodiment 3 )
The present embodiment is characterized by the determination process in the first determination unit 22 of the first and second speech section detection units 11 ′ and 12 ′, and the overall configuration is the same as that of the second embodiment, so that it is illustrated and described. Is omitted.

  In the present embodiment, the first determination unit 22 (1) the instantaneous power estimated value Ps (n) is greater than or equal to a predetermined threshold Pth, and (2) the instantaneous power estimated value Ps (n) and the background noise power. The ratio Ps (n) / Pn (n) of the estimated value Pn (n) is equal to or greater than the threshold value δ, and (3) two instantaneous power estimated values Ps (n calculated with a predetermined time interval K. ), Ps (nK), the reference signal x (n) is determined to be a voice only when all of the three conditions that the absolute value of the difference between Ps (nK) is greater than or equal to a predetermined threshold value χ are satisfied. The time interval K is set to an appropriate value in accordance with the characteristics of the target noise (for example, a baby cry described later).

  Next, specific determination processing in the first determination unit 22 will be described based on the flowchart of FIG. First, the instantaneous power estimation value Ps (n) calculated by the instantaneous power estimation unit 20 is compared with a threshold value Pth (step 1). The ratio Ps (n) / Pn (n) of the background noise power estimated value Pn (n) is compared with the threshold value δ (step 2). If the ratio Ps (n) / Pn (n) is greater than or equal to the threshold value δ, the absolute value of the difference between the two instantaneous power estimated values Ps (n) and Ps (nK) | Ps (n) −Ps ( nK) | is compared with a threshold value χ (step 3). Further, the instantaneous power estimated value Ps (n) is less than the threshold value Pth, the ratio Ps (n) / Pn (n) is less than the threshold value δ, or the absolute value of the difference | Ps (n) −Ps (nK) | Is less than the threshold value χ, it is determined as a non-voice segment (step 5).

  Here, the above two conditions (1) and (2) have been generally used, and the present inventors have added non-steady ambient noise other than speech by adding the condition (3). Has been confirmed by experiments to prevent false detection as a voice. That is, assuming the baby's cry as non-stationary ambient noise, the instantaneous power estimate for the reference signal x (n) including the caller's voice (male voice and female voice) and the baby's cry Since Ps and the absolute difference value | Ps (n) −Ps (nK) | of the instantaneous power estimation value are obtained, the results are shown in FIGS. FIGS. 9 (a), 10 (a) and 11 (a) show the instantaneous power estimate Ps when the reference signal x (n) includes baby cry, male voice, and female voice, respectively. (B) of each figure has shown the absolute value of the difference of an instantaneous power estimated value, respectively. It should be noted that the time interval K was 4 ms, and the level of the reference signal x (n) was equal to the average sound pressure of male and female voices, which was about 4 dB greater than the baby cry.

  Thus, comparing FIG. 9 (a), FIG. 10 (a) and FIG. 11 (a), the voice of the caller has a large temporal fluctuation of the estimated power Ps (n) relative to the baby's cry. I understand. Therefore, as shown in FIGS. 9 (b), 10 (b) and 11 (b), there is a significant difference in the absolute difference value | Ps (n) −Ps (nK) | . Therefore, by adding the absolute difference value | Ps (n) −Ps (nK) | to the determination condition, the baby's cry can be determined as noise (non-speech), in other words, erroneous determination as voice can be prevented. . However, it is considered that non-stationary ambient noises that have a small time variation compared to the call voice as in the case of the baby's cry, such as classical music and howling dogs can be determined as non-voice according to this embodiment. .

(Embodiment 4 )
FIG. 12 shows a block diagram of this embodiment. However, the basic configuration of the present embodiment because it is common to Embodiment 1, the same components will not be described with the same reference numerals, a description will be given only of the elements having the corresponding functions in this embodiment.

This embodiment is in the configuration of Embodiment 1 is characterized in that the application of the first speech section detecting unit 11 of the embodiment 2 '. That is, since it is considered that noises such as telephone ringing sounds and dog calls are often generated indoors, a measure for preventing erroneous detection of a voice segment due to this type of noise is applied only to the first voice segment detector 11 '. By doing so, the total cost can be reduced. In addition, it is desirable to prevent the baby's cry from being erroneously determined as speech by performing the same determination process as in the third embodiment in the first determination unit 22 of the first speech section detection unit 11 ′.

(Embodiment 5 )
FIG. 13 shows a block diagram of this embodiment. However, the basic configuration of the present embodiment because it is common to Embodiment 1, the same components will not be described with the same reference numerals, a description will be given only of the elements having the corresponding functions in this embodiment.

This embodiment is in the configuration of Embodiment 1, the low-pass filter for removing a high frequency band component than the main component band of the audio from the reference signal to be referred to in the first speech section detecting unit 11 (transmission signal) ( LPF) 16 is provided in the voice switch VS.

That is, as described in the second embodiment, there is an environment where there is noise that has a non-stationary frequency band component higher than the main component band of voice, such as a telephone ringing tone or an animal (particularly a small dog) cry. The first speech segment detection unit 11 may erroneously detect the speech segment due to these noises, and therefore the frequency band higher than the main component band of the speech by the low-pass filter 16. If the component is removed from the transmission signal, the influence of noise such as the ringing tone of the telephone and the crying of the animal is reduced by a simple configuration compared to the first voice section detecting unit 11 ′ in the second embodiment. It is possible to suppress the erroneous detection of the speech section in the speech section detection unit 11.

  By the way, when setting the cut-off frequency of the low-pass filter 16, it is determined mainly considering the frequency characteristics of an adult voice. Generally, the pitch of an infant's voice is relative to that of an adult voice. Therefore, part of the infant's voice may be removed by the low-pass filter 16. Therefore, when the loudspeaker device of this embodiment is used in a home where such an infant is present, the first voice interval detection unit 11 must be set in the infant unless the cutoff frequency of the low-pass filter 16 is set to an appropriate value. There is a possibility that the voice may not be determined to be voice, and the voice section may be erroneously detected and one-sided fall may occur. Therefore, the operation input by the user's operation is captured and the cut-off frequency of the low-pass filter 16 is switched according to the operation input, or the pitch of the voice included in the reference signal is detected and the detected voice pitch is detected. Accordingly, it is desirable to provide means for changing the cutoff frequency of the low-pass filter 16. For example, a changeover switch is provided in a housing (not shown) of the loudspeaker device according to the present embodiment, and an operation input is taken into the control unit when a user operates the changeover switch. If the cut-off frequency of the low-pass filter 16 is switched in several ways, the user switches the cut-off frequency of the low-pass filter 16 depending on the presence / absence of the infant, and the voice of the infant is the first voice. It is possible to prevent a voice section from being erroneously detected without being determined as voice by the section detection unit 11. Alternatively, the pitch of the voice included in the reference signal (transmission signal) is detected, and it is determined from the detected voice pitch whether the voice included in the reference signal is an infant voice. The frequency is switched to a high value and the cut-off frequency is switched to a low value if the voice is not an infant voice, and the voice section is erroneously detected without being judged as a voice by the first voice section detection unit 11. Can be prevented. Note that the method for detecting the voice pitch is well known in the art and will not be described in detail.

(Embodiment 6 )
FIG. 14 shows a block diagram of the present embodiment. However, since the basic configuration of the present embodiment is the same as that of the fifth embodiment, the same components are denoted by the same reference numerals and description thereof is omitted, and only the configuration that characterizes the present embodiment will be described.

The present embodiment is characterized in that, in the configuration of the fifth embodiment, a high-pass filter (HPF) 17 that removes frequency band components lower than the main component band of speech is provided in series with the low-pass filter 16. . That is, the cutoff frequency of the high-pass filter 17 is set to the lower limit value of the main component band of the voice, and the cutoff frequency of the low-pass filter 16 is set to the upper limit value of the main component band of the voice.

  Thus, according to the present embodiment, the low-pass filter 16 removes noise having a frequency band component higher than the main component band of speech, such as a telephone ringing tone and an animal call, By removing noise having a frequency band component lower than the main component band of the voice with the high-pass filter 17, it is possible to prevent the first voice section detection unit 11 from erroneously detecting the voice section due to noise, and noise. The transmission side instantaneous power estimate value Ps (T) can be reduced, and as a result, the fall of the voice switch VS can be suppressed.

(Embodiment 7 )
FIG. 15 shows a block diagram of the present embodiment. However, since the basic configuration of the present embodiment is the same as that of the sixth embodiment, the same components are denoted by the same reference numerals, description thereof is omitted, and only the configuration that is a feature of the present embodiment will be described.

  In the present embodiment, as shown in FIG. 15, among the constituent elements of the voice switch VS, the constituent elements other than the transmission side and receiving side loss insertion sections 3 and 4 are realized by a digital circuit A such as a DSP. There are features. That is, the voice switch VS in the present embodiment performs A / D conversion on the transmission side loss insertion unit 3 and the reception side loss insertion unit 4 made of analog circuits and analog reference signals (transmission signal and reception signal). A / D converters 18 and 19 to be output to the digital circuit A, insertion loss amount control unit 10, first and second speech section detection units 11 and 12, attenuator 13, attenuation amount control unit 14, digital reference It comprises a digital circuit A that implements the functions of the first to third filter units 31 to 33 for filtering signals by software.

  The first and second speech section detection units 11 and 12 have a common configuration, and as shown in FIG. 16, a reference signal (a transmission signal input to the transmission side loss insertion unit 3 or a reception side loss insertion unit 4). Of the reference signal based on the ratio of the instantaneous power estimation value and the background noise power estimation value, and the background noise power estimation unit 21 for estimating the background noise power of the reference signal. A first determination unit 22 that determines a section and a time constant update unit 26 are provided. When the instantaneous power estimation unit 20 is calculated by the absolute average value calculation unit 201 for obtaining the time average value (absolute average value) Pz (n) of the absolute value of the reference signal x (n) and the absolute average value calculation unit 201 The absolute average value smoothing unit 202 smoothes the absolute average value Pz (n) of the series.

  The absolute average value calculation unit 201 calculates an absolute value calculation unit 201a that calculates the absolute value of the reference signal x (n) sampled at a predetermined sampling time, and calculates the sum of absolute values in a predetermined time frame (sampling number M). The sum calculation unit 201b and the division unit 201c that calculates the absolute average value Px (n) by dividing the calculated sum by the sampling number M. As a result, the absolute average value calculation unit 201 uses the following formula (1 ).

  The absolute average value smoothing unit 202 multiplies the absolute average value Pz (n) by a positive constant α (<1), a delay shift register 202b, and an instantaneous power estimation delayed by the delay shift register 202b. It consists of a multiplier 202c that multiplies the value Ps (n-1) by a positive constant (1-α), and an adder 202d that adds the outputs of the two multipliers 202a and 202c. The smoothing unit 202 performs the following equation (2).

  On the other hand, the background noise power estimation unit 21 delays the instantaneous power estimate value Pn (n), the instantaneous power estimate value Ps (n), and the instantaneous power estimate value Pn (delayed by the delay shift register 211). n-1), first and second counters 213 and 214 for incrementing the count values Cu and Cd according to the comparison result by the comparator 212, and the first and second counters, respectively. One of three correction values β (n), 0, −β (n) (where β (n)> 0) depending on the magnitude relationship between the count values Cu and Cd of 213 and 214 and the threshold values Us and Ud. Is selected and output, and an adder 216 that adds the delayed instantaneous power estimation value Pn (n-1) to the correction value output from the selector 215. Here, the first and second counters 213 and 214 update the count values Cu and Cd according to the following rules for each sampling time of the reference signal x.

If Ps (n) ≧ Pn (n−1), Cu = Cu + 1, Cd = 0
If Ps (n) <Pn (n-1), Cu = 0, Cd = Cd + 1
The selector 215 selects and outputs one of the three correction values according to the following rules.

If Cu = Us, β (n) (at the same time reset to Cu = 0)
If Cd = Ds, -β (n) (at the same time reset to Cd = 0)
0 if Cu ≠ Us and Cd ≠ Ds
Therefore, if the thresholds Us and Ds to be compared with the count values Cu and Cd of the first and second counters 213 and 214 are set such that Us >> Ds, the rising time constant is large and the falling time constant is A filter having a response characteristic with a small constant can be realized. The rising time constant is determined by the positive correction value β (n) and its threshold value Us, and becomes shorter as the correction value β (n) is larger or the threshold value Us is smaller.

  Further, the first determination unit 22 determines that the voice section is a speech flag when the ratio (= Ps / Pn) of the instantaneous power estimated value Ps and the background noise power estimated value Pn is equal to or greater than a predetermined threshold, and a detection flag (voice section Detection signal) is set to “1”, and when the ratio is less than the threshold value, it is determined as a non-speech interval and the detection flag is set to “0”.

  FIG. 17 is a block diagram showing the first and second filter units 31 and 32, and FIG. 18 is a block diagram showing the third filter unit 33, both of which are constituted by digital filters. The first filter unit 31 is a secondary filter of order 2, and is realized by five parameters b0 to b5, two delay processes D, and four addition processes, and the second filter unit 32 is 1 of order 1. A next filter, which is realized by three parameters a0 to a2, one delay process D, and two addition processes. Then, by appropriately setting the values of the parameters a0 to a2 and b0 to b5, a high-pass filter and a low-pass filter having desired characteristics can be easily realized without changing the circuit configuration. Further, the third filter unit 32 is a first-order filter of order 1, which is realized by two parameters c1 and c2, one delay process D, and two addition processes, and appropriately sets the values of the parameters c1 and c2. By setting, a high-pass filter having desired characteristics can be easily realized without changing the circuit configuration.

It is a block diagram which shows the reference example of this invention. It is a block diagram which shows the 1st and 2nd audio | voice area detection part in the same as the above. It is a block diagram which shows the attenuation amount control part in the same as the above. It is a block diagram which shows the comparative example same as the above. It is a block diagram which shows Embodiment 1 of this invention. It is a block diagram which shows the 1st and 2nd audio | voice area detection part in Embodiment 2 of this invention. It is a flowchart for operation | movement description of the 2nd determination part in the same as the above. It is a flowchart for operation | movement description of the 1st determination part in Embodiment 3 of this invention. It is a wave form diagram which shows the experimental result with respect to the baby's cry in the same as the above. It is a wave form diagram which shows the experimental result with respect to the male voice in the same as the above. It is a wave form diagram which shows the experimental result with respect to the female voice in the same as the above. It is a block diagram which shows Embodiment 4 of this invention. It is a block diagram which shows Embodiment 5 of this invention. It is a block diagram which shows Embodiment 6 of this invention. It is a block diagram which shows Embodiment 7 of this invention. It is a block diagram which shows the 1st and 2nd audio | voice area detection part in the same as the above. It is a block diagram which shows the 1st and 2nd filter part in the same as the above. It is a block diagram which shows the 3rd filter part in the same as the above. It is a block diagram which shows a prior art example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Microphone 2 Speaker 3 Transmission side loss insertion part 4 Reception side loss insertion part 10 Insertion loss amount control part 11 1st audio | voice area detection part 12 2nd audio | voice area detection part 13 Attenuator 14 Attenuation amount control part

Claims (13)

A microphone and a speaker, and a voice switch that switches between a transmission state that sends a transmission signal to the transmission system and attenuates the reception signal and a reception state that sends the reception signal to the speaker and attenuates the transmission signal,
The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side loss insertion unit that inserts a loss into the reception side signal path from the transmission system to the speaker. A first voice section detector for detecting a voice section of a transmitted signal, a second voice section detector for detecting a voice section of a received signal, and a transmitter instantaneous power estimation for estimating the instantaneous power of the transmitted signal A transmission side background noise power estimation unit for estimating the background noise power of the transmission signal, a reception side instantaneous power estimation unit for estimating the instantaneous power of the reception signal, and a reception side for estimating the background noise power of the reception signal A background noise power estimation unit, and an insertion loss amount control unit for controlling an insertion loss amount in each loss insertion unit on the transmission side and reception side,
The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result ,
A first attenuator for attenuating the instantaneous power estimation value on the transmission side and a first attenuation amount for increasing the attenuation amount of the first attenuator when the transmission side background noise power estimation value exceeds a predetermined threshold value A second attenuator for attenuating the instantaneous power estimate value on the control unit or the receiver side and a second attenuator for increasing the attenuation amount of the second attenuator when the receiver side background noise power estimate value exceeds a predetermined threshold value at least one comprises one Rutotomoni attenuation control unit, comprising a high-pass filter for removing low frequency band component than the main component band of a speech from the receiving signal to be referenced in the second speech section detecting unit in the voice switch hands-free communication device, characterized in that the. 2. The loudspeaker apparatus according to claim 1, further comprising means for specifying a call device of a partner with which a call is made via a transmission system and switching between valid / invalid of a high-pass filter in accordance with the specified call device. The loudspeaker apparatus according to claim 1, further comprising means for taking in an operation input by a user's operation and switching valid / invalid of a high-pass filter in accordance with the operation input . A microphone and a speaker, and a voice switch that switches between a transmission state that sends a transmission signal to the transmission system and attenuates the reception signal and a reception state that sends the reception signal to the speaker and attenuates the transmission signal,
The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side loss insertion unit that inserts a loss into the reception side signal path from the transmission system to the speaker. A first voice section detector for detecting a voice section of a transmitted signal, a second voice section detector for detecting a voice section of a received signal, and a transmitter instantaneous power estimation for estimating the instantaneous power of the transmitted signal A transmission side background noise power estimation unit for estimating the background noise power of the transmission signal, a reception side instantaneous power estimation unit for estimating the instantaneous power of the reception signal, and a reception side for estimating the background noise power of the reception signal A background noise power estimation unit, and an insertion loss amount control unit for controlling an insertion loss amount in each loss insertion unit on the transmission side and reception side,
The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result ,
A first attenuator for attenuating the instantaneous power estimation value on the transmission side and a first attenuation amount for increasing the attenuation amount of the first attenuator when the transmission side background noise power estimation value exceeds a predetermined threshold value A second attenuator for attenuating the instantaneous power estimate value on the control unit or the receiver side and a second attenuator for increasing the attenuation amount of the second attenuator when the receiver side background noise power estimate value exceeds a predetermined threshold value At least one of the attenuation amount control units is provided, and the first and second speech section detection units are constantly present in the reference signal and a first instantaneous power estimation unit that estimates the instantaneous power of the reference signal. A background noise power estimator that estimates the power of the background noise, a first determination unit that determines the speech section of the reference signal based on the ratio of the instantaneous power estimate value and the background noise power estimate value, Frequency band component higher than the component band A low-pass filter to be removed, a second instantaneous power estimation unit that estimates the instantaneous power of the reference signal after the high-frequency band component has been removed by the low-pass filter, and a non-voice section in the first determination unit When it is determined, it is determined as a non-speech interval, and when it is determined as a speech interval by the first determination unit, the instantaneous power estimation value estimated by the first instantaneous power estimation unit is multiplied by a positive coefficient less than 1. expanding voice call, characterized by comprising a second determining unit that determines whether the speech section based value and the magnitude relationship between the instantaneous power estimate estimated by the second instantaneous power estimator apparatus. The loudspeaker apparatus according to claim 4, wherein the coefficient in the second determination unit is variable . The first determination unit included in the first speech section detection unit obtains an absolute value of a difference between the two instantaneous power estimation values estimated by the first instantaneous power estimation unit at a predetermined time interval, absolute value with a predetermined hands-free communication device according to claim 4 or 5, wherein the determining by referring to the result of comparison between the threshold value of the difference. A microphone and a speaker, and a voice switch that switches between a transmission state that sends a transmission signal to the transmission system and attenuates the reception signal and a reception state that sends the reception signal to the speaker and attenuates the transmission signal,
The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side loss insertion unit that inserts a loss into the reception side signal path from the transmission system to the speaker. A first voice section detector for detecting a voice section of a transmitted signal, a second voice section detector for detecting a voice section of a received signal, and a transmitter instantaneous power estimation for estimating the instantaneous power of the transmitted signal A transmission side background noise power estimation unit for estimating the background noise power of the transmission signal, a reception side instantaneous power estimation unit for estimating the instantaneous power of the reception signal, and a reception side for estimating the background noise power of the reception signal A background noise power estimation unit, and an insertion loss amount control unit for controlling an insertion loss amount in each loss insertion unit on the transmission side and reception side,
The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result ,
A first attenuator for attenuating the instantaneous power estimation value on the transmission side and a first attenuation amount for increasing the attenuation amount of the first attenuator when the transmission side background noise power estimation value exceeds a predetermined threshold value A second attenuator for attenuating the instantaneous power estimate value on the control unit or the receiver side and a second attenuator for increasing the attenuation amount of the second attenuator when the receiver side background noise power estimate value exceeds a predetermined threshold value The voice switch includes at least one of the attenuation amount control units, and includes a high-pass filter that removes a frequency band component lower than the main component band of the voice from the reception signal referred to by the second voice section detection unit. The first speech section detector includes a first instantaneous power estimator that estimates the instantaneous power of the reference signal, a background noise power estimator that estimates the power of background noise that is constantly present in the reference signal, Instantaneous power estimates and background A first determination unit that determines a speech section of a reference signal based on a ratio of estimated sound power values, a low-pass filter that removes a frequency band component higher than a main component band of speech from the reference signal, and a low-pass A second instantaneous power estimation unit that estimates the instantaneous power of the reference signal after the high frequency band component is removed by the filter, and a non-speech interval when the first determination unit determines that it is a non-speech segment When the first determination unit determines that the voice section is used, the second instantaneous power estimation unit uses a value obtained by multiplying the instantaneous power estimation value estimated by the first instantaneous power estimation unit by a positive coefficient less than 1. expansion voice communication apparatus being characterized in that and a second determination unit that determines whether the speech segment based on the magnitude relationship between the estimated instantaneous power estimate. The first determination unit obtains an absolute value of a difference between the two instantaneous power estimation values estimated by the first instantaneous power estimation unit at a predetermined time interval, and calculates the absolute value of the difference and a predetermined threshold value. The loudspeaker apparatus according to claim 7, wherein the determination is made with reference to a comparison result with . A microphone and a speaker, and a voice switch that switches between a transmission state that sends a transmission signal to the transmission system and attenuates the reception signal and a reception state that sends the reception signal to the speaker and attenuates the transmission signal,
The voice switch includes a transmission side loss insertion unit that inserts a loss into the transmission side signal path from the microphone to the transmission system, and a reception side loss insertion unit that inserts a loss into the reception side signal path from the transmission system to the speaker. A first voice section detector for detecting a voice section of a transmitted signal, a second voice section detector for detecting a voice section of a received signal, and a transmitter instantaneous power estimation for estimating the instantaneous power of the transmitted signal A transmission side background noise power estimation unit for estimating the background noise power of the transmission signal, a reception side instantaneous power estimation unit for estimating the instantaneous power of the reception signal, and a reception side for estimating the background noise power of the reception signal A background noise power estimation unit, and an insertion loss amount control unit for controlling an insertion loss amount in each loss insertion unit on the transmission side and reception side,
The insertion loss amount control unit determines the call state with reference to the comparison result of the instantaneous power estimation values on the transmission side and the reception side and the detection results of the first and second voice section detection units, and In a loudspeaker device that is switched to at least one of a transmission mode in which the insertion loss amount on the receiving side is relatively large or a reception mode in which the insertion loss amount on the transmission side is relatively large according to the determination result ,
A first attenuator for attenuating the instantaneous power estimation value on the transmission side and a first attenuation amount for increasing the attenuation amount of the first attenuator when the transmission side background noise power estimation value exceeds a predetermined threshold value A second attenuator for attenuating the instantaneous power estimate value on the control unit or the receiver side and a second attenuator for increasing the attenuation amount of the second attenuator when the receiver side background noise power estimate value exceeds a predetermined threshold value A low-pass filter that includes at least one of the attenuation amount control unit and removes a frequency band component higher than the main component band of the speech from the transmission signal referred to by the first speech section detection unit; expansion voice communication apparatus being characterized in that provided in the voice switch from the received signal and a high pass filter for removing low frequency band component than the main component band of sound to be referenced in section detection unit. The voice switch includes the low-pass filter having a variable cut-off frequency, and includes means for capturing an operation input by a user operation and changing the cut-off frequency of the low-pass filter in accordance with the operation input. The loudspeaker apparatus according to claim 9 . The loudspeaker apparatus according to claim 9, further comprising means for detecting a pitch of voice included in the reference signal and changing a cutoff frequency of the low-pass filter in accordance with the detected voice pitch . The loudspeaker apparatus according to claim 9, wherein a high-pass filter for removing a frequency band component lower than a main component band of the voice is provided in series with the low-pass filter . 13. The loudspeaker apparatus according to claim 12 , wherein the filter that removes a frequency band component higher or lower than the main component region of the voice from the transmission signal and the reception signal is constituted by a digital filter .

Images