JPH08335977A - Loudspeaking device - Google Patents

Abstract

PURPOSE: To attain stable excellent speech even in an environment with much reverberation and background noise by measuring a line coupling amount and an acoustic coupling amount and setting a sum of the attenuation of a transmission signal variable attenuator and the attenuation of a reception signal variable attenuator based on the result of measurement. CONSTITUTION: When power is supplied to a speaker phone 1 and it is connected to a telephone line, a coupling amount measurement section 23 sends a signal such as white noise to a transmission signal path 9 under the control of a variable controller 22, detected by a reception signal path 15 to obtain a line coupling amount. Similarly, the measurement section 23 obtains the acoustic coupling amount based on the level of the signal such as white noise sent to the path 15 and the level of the signal detected by the path 9. The controller 22 decides the sum of attenuation of transmission/reception variable attenuators 13, 19 based on the line/acoustic coupling amount to be measured and on the known gain of all parts in the hand-free circuit to set a gain. Thus, natural speech is conducted with a simple and inexpensive configuration.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice communication device such as a hands-free communication device or a telephone conference system, and more particularly to a technique for switching voices according to the level of a signal to be transmitted / received.

[0002]

2. Description of the Related Art When making a voice call using a speaker and a microphone in a voice call device, a signal is input to a circuit from a loop gain and a line in the voice call device to prevent howling. It is necessary to reduce the amount of returns that represents the ratio of. Therefore, depending on the levels of the transmission signal and the reception signal, the transmission signal is provided in the transmission signal path through which the transmission signal is transmitted to the telephone line and the reception signal path through which the reception signal is transmitted through the speaker. A technology that complementarily controls the attenuation amounts of the variable attenuator and the reception signal variable attenuator is used. 4 to 7 as techniques conventionally proposed in order to prevent the occurrence of line echo or acoustic echo during a call in such a voice communication device.
There is. For example, FIG. 4 is proposed in German Patent PAT. 3839627. In the communication device shown in the figure, in order to cancel the acoustic echo from the speaker 251 to the microphone 253, the transmission line 55 and the reception line 57 are used.
Acoustic echo cancellation circuit 265 including an adaptive FIR (Finite Inpulse Response) filter 267 between
Is provided. In addition, the device has a microphone 253
The transmission volume from the line and the reception volume from the line are compared, and the transmission line 255 and the reception line 2 are compared according to the magnitude relationship.
A control unit 263 for controlling the amount of attenuation of the attenuators 259 and 261 in 57 is provided. As a result, the device controls the attenuators 259 and 261 by the control unit 63 by the amount of the attenuation by the acoustic echo cancellation circuit 265.
Of the transmission line 255, the reception line 257, and the acoustic echo cancellation circuit 265 and the attenuation amount due to the acoustic coupling from the speaker 251 to the microphone 253. The amount of attenuation is constant. As a result, as the adaptation of the coefficient of the adaptive FIR filter 267 progresses and the attenuation amount of the acoustic echo cancellation circuit 265 increases, the attenuation amount of the attenuator 259 or the attenuator 261 decreases, and finally it becomes almost It is designed to be a two-way simultaneous call (front duplex). The operation of suppressing the amount of attenuation of this device is as follows.

First, the attenuators 259 and 261 on the transmission line 255 and the reception line 257 are controlled by the control unit 263 so that only the normal echo suppressor functions. Next, when the other speaker starts speaking, the adaptive filter 2
67 starts adaptation and causes some attenuation R for the acoustic echo. The attenuation amount in the echo suppressor is reduced by this attenuation amount R. This reduces the one-way (half-duplex) characteristic of the echo suppressor. When the adaptation progresses sufficiently and the coefficient of the adaptive filter 267 converges, if the acoustic echo cancellation performance is sufficiently large, the attenuation amounts of the attenuators 259 and 261 become almost 0, and the two-way simultaneous call (full duplex) occurs. It will be possible. While both parties are talking (during double talk), learning of the adaptive filter 267 must be prohibited, but the voice of the other party is the acoustic echo cancellation circuit 2.
Since it is attenuated by 65 and does not mix with the voice of the person input from the microphone 253, accurate double talk detection is possible. FIG.
FIG. 3 is a block diagram of a hands-free circuit proposed in Japanese Patent Laid-Open Publication No. In the figure, a voice signal transmitted by a near-end speaker is input to a microphone 402 of a telephone 401. The audio signal input to the microphone 402 is input to the microphone amplifier 40.
Amplification by 3 and transmission signal variable attenuator 405, amplifier 40
7, and is output from the output terminal 409 to the telephone line via the hybrid transformer.

The received signal sent from the far-end speaker through the telephone line enters the telephone 401 from the input terminal 411 through the hybrid transformer, and the amplifier 413, received signal variable attenuator 415, Speaker amplifier 4
It passes through 17 and is output by the speaker 419 and reaches the near-end speaker. Here, the amount of attenuation of the transmission signal variable attenuator 405 and the amount of attenuation of the reception signal variable attenuator 415 are controlled by the control unit 420 based on the magnitude of the transmission signal and the reception signal. That is, the levels of the transmission signal and the reception signal are compared by the comparators 421 and 423. Further, the signal-to-noise detector 427 detects the presence or absence of voice in the transmitted signal, and the signal-to-noise detector 425 detects the presence or absence of voice in the received signal. Based on the output signals of the comparators 421 and 423 and the signal-to-noise detectors 425 and 427, the control unit 420 determines which of the transmission mode, the reception mode, and the high speed / slow idle mode should be performed, and the transmission mode is determined accordingly. The attenuation amount of the speech signal variable attenuator 405 and the attenuation amount of the reception signal variable attenuator 415 are determined. Attenuation amount of transmitting signal variable attenuator 405 and receiving signal variable attenuator 4
The attenuation amounts of 15 are controlled complementarily, that is, while maintaining a complementary relationship such that the sum of the attenuation amounts of both is constant. When the transmission signal level is higher than the reception signal level in both of the comparators 421 and 423 and the voice is detected by the signal-to-noise detector 427, the transmission mode is set, and the attenuation amount of the transmission signal variable attenuator 405 is set. Is set to the minimum, and the attenuation amount of the reception signal variable attenuator 415 is set to the maximum. Further, in both the comparators 421 and 423, the reception signal level is higher than the transmission signal level, and the signal-to-noise detector 4
When voice is detected by 25, the reception mode is set, and the attenuation amount of the transmission signal variable attenuator 405 is set to the maximum and the attenuation amount of the reception signal variable attenuator 415 is set to the minimum. When voice is not detected in both the transmission signal and the reception signal, the slow idle mode, the comparison results of the transmission / reception signal levels in the comparators 421 and 423 do not match, and the signal-to-noise detectors 425 and 427 detect the noise. When at least one of the voices is detected, the high speed idle mode is set, and the transmission signal variable attenuator 405 and the reception signal variable attenuator 4 are set.
Both attenuations of 15 are placed at an intermediate level. In any mode, by maintaining the sum of the attenuation amount of the transmission signal variable attenuator 405 and the attenuation amount of the reception signal variable attenuator 415 at a predetermined constant value, it is possible to make a stable call without howling. .

In addition, FIG. 6 shows US Pat. No. Three
It is a structural example of the conventional hands-free circuit proposed by 952166. In the figure, the voice signal transmitted by the near-end speaker is input to the microphone 503 of the telephone 501.
The voice signal input to the microphone 503 is amplified by the microphone amplifier 505, passes through the transmission signal variable attenuator 507, and is output from the output terminal 509 to the telephone line via the hybrid transformer. In addition, the reception signal transmitted from the telephone on the far-end talker side via the telephone line is transmitted from the input terminal 511 to the telephone 50 via the hybrid transformer.
1, the received signal variable attenuator 513, the speaker amplifier 515, the speaker 517 outputs, and reaches the near-end speaker. Here, the attenuation amount of the transmission signal variable attenuator 507 and the attenuation amount of the reception signal variable attenuator 513 are complementarily controlled by the magnitudes of the transmission signal and the reception signal as in the prior art example of FIG. During the call, call voice is used to measure the amount of acoustic coupling and line coupling. That is, during the transmission of speech from the near-end speaker, the line coupling amount is measured by detecting the voice signal level in the reception signal path with respect to the voice signal level in the transmission signal path, and is set first. If it is smaller than the line coupling amount, the attenuation amount of the reception signal variable attenuator 513 is changed to be smaller by the difference. Also, while the far-end speaker is speaking, the amount of acoustic coupling is measured by detecting the voice signal level in the transmission signal path with respect to the voice signal level in the reception signal path, and the acoustic level set first If it is smaller than the coupling amount, the attenuation amount of the transmission signal variable attenuator 507 is changed to be smaller by the difference. By continuously performing this operation during a call, the attenuation amounts of the transmission signal variable attenuator 507 and the reception signal variable attenuator 513 are appropriately adjusted in accordance with the state in which the loudspeaker communication device is placed. It is possible to prevent the howling margin in the hands-free circuit from unnecessarily increasing, and to reduce the amount of change in the attenuation amount of the transmission signal variable attenuator 507 and the reception signal variable attenuator 513 as much as possible to achieve the phase-direction simultaneous call. It is possible to improve the property.

FIG. 7 is a block diagram showing an example of a conventional hands-free circuit proposed in Japanese Patent Laid-Open No. 2-67049. In the figure, the transmitted voice signal from the near-end speaker is input to the microphone 601. Microphone 601
The audio signal input to is amplified by the microphone amplifier 602 and passes through the attenuator 607. Also, the received signal input from the telephone line into the circuit is the speaker amplifier 60.
The signal is amplified by 5 and output by the speaker 603 to reach the near-end speaker. Here, the attenuator 607 does not operate at all when a normal call without howling is being performed. If some reflector is present in the vicinity of the microphone 601, the amount of acoustic coupling from the speaker 603 to the microphone 601 becomes large and howling occurs,
From the signal S2 in the transmission signal path to the band pass filter (BP
F) 611 extracts the howling signal S4 and inputs it to the comparator 613. The comparator 613 compares the input howling signal with a predetermined reference voltage V ₀ and outputs the result to the attenuator control circuit 60.
Output to 9. The attenuator control circuit 609 controls the attenuator 607 so that the attenuation rate of the attenuator 607 increases as the howling signal increases.
By controlling the attenuation factor of linearly, the loop gain in the hands-free circuit is reduced to 1 or less, and howling is suppressed.

[0007]

However, in the device shown in FIG. 4, since the amount of attenuation of the voice switch is dynamically reduced corresponding to the amount of attenuation by the acoustic echo cancellation, the device is, for example, a loudspeaker telephone. When applied, (1) When the environment in which the telephone is placed is easily reverberated or there is a lot of background noise, acoustic echo cancellation does not work well and its attenuation changes rapidly. As a result, the amount of attenuation of the attenuator controlled by the echo presser also changes abruptly, causing frequent transitions between the two-way communication state and the one-way communication state, leading to head loss and volume fluctuations. It tends to occur, making it difficult to hear the voice. (2) Since the attenuation amount of the echo presser is dynamically controlled, the calculation amount of signal processing is increased, and the relatively expensive D
SP is required. There is a problem such as. Further, in the configuration shown in FIG.
The amount of line coupling changes depending on the line to which the loudspeaker is connected, and the amount of acoustic coupling changes depending on the environment in which the speaker and microphone are installed and the conditions of use.
Therefore, in the former cases of the above two conventional examples, the line coupling amount and the acoustic coupling amount used in determining the sum of the attenuation amount of the transmission signal variable attenuator and the attenuation amount of the reception signal variable attenuator. Since it is necessary to assume in advance that the respective coupling amounts are close to the maximum, the sum of the attenuation amount of the transmission signal variable attenuator and the attenuation amount of the reception signal variable attenuator is determined to be larger than necessary. It often happens. In this way, if the sum of the attenuation of the transmission signal variable attenuator and the attenuation of the reception signal variable attenuator is set larger than necessary, the change in the attenuation of the variable attenuator when switching between transmission and reception Becomes larger, leading to disconnection at the beginning and ending of the ending, and even if one tries to talk while the other is talking, the amount of attenuation of the received signal variable attenuator inserted in the received signal path is large. Therefore, the other voice is hardly or not heard by the one voice. That is, the two-way simultaneous callability is low and the naturalness of the call is impaired.

Further, in the configuration shown in FIG. 6, the acoustic coupling amount and the line coupling amount are constantly measured during a call, and the attenuation amount of the transmitting or receiving signal variable attenuator is changed accordingly. ， While talking, adjust the attenuation of the transmitting or receiving signal variable attenuator to the optimum state for the state that the loudspeaker is placed, but always measure the acoustic coupling amount and line coupling amount during the call. In order to adjust the attenuation amount of the transmission signal variable attenuator or the reception signal variable attenuator accordingly, the circuit configuration becomes complicated and the amount of arithmetic processing increases, which increases the cost of the device. Moreover, when the amount of acoustic coupling changes due to the movement of the near-end speaker during reception, etc., the amount of attenuation given to the transmitted signal changes accordingly, especially when the ambient noise of the near-end speaker is large. However, the noise level heard by the far-end speaker from the handset may fluctuate, giving an unstable and unnatural call feeling. Further, in the configuration shown in FIG. 7, howling does not always occur at a constant frequency because it occurs when the condition that the loop gain becomes 1 or more is satisfied at a certain frequency. Therefore, the occurrence of howling cannot always be detected even if the level of the signal that has passed through a single bandpass filter is observed. However, providing a large number of bandpass filters in consideration of the possibility of howling at various frequencies complicates the circuit configuration and raises the cost. Moreover, even if it is detected that the signal level included in a certain frequency band exceeds the reference voltage, it is difficult to distinguish the case where the near-end speaker's voice or background noise is extremely large from howling. The risk of detection is high.

Furthermore, when the technique shown in FIGS. 4 to 7 is applied to a speakerphone with a handset, there are the following problems. When talking in a handset using a handset, the sound transmitted by the speaker returns to the receiving signal path to some extent, and it is easier for the speaker to speak while listening to his own voice. The sound coupling amount is set to a large state. In handset calls, the ear and speaker of the caller,
Since the speaker's mouth is close to the microphone, the volume of the speaker and the gain of the amplifier may be small, and howling does not occur even if the sidetone coupling amount is large. On the other hand, even in the hands-free state, the sidetone coupling amount remains the same as in the handset state, but in this case, since the speaker talks away from the microphone and the speaker, the speaker volume and the gain of the amplifier are large. Is set. Also, since there is no shielding effect by the speaker's face, the acoustic coupling between the speaker and the microphone becomes large.

As a result, the loop gain formed by the reception signal path and the acoustic coupling from the transmission signal path through the sidetone coupling increases, and howling easily occurs. Therefore,
In the hands-free state, the attenuation amount of the transmission signal variable attenuator and the reception signal variable attenuator is large, that is, the transmission volume and the reception volume are compared with each other, and in the transmission state, the reception volume is reduced. It may be possible to control the so-called half-duplex operation that maximizes the attenuation of the attenuator and maximizes the attenuation of the transmitting attenuator in the receiving state. Since it is dealt with by increasing the attenuation amount of the reception attenuator 9, by using the half-duplex operation, as a result, there is a problem that the beginning of speech and the ending of the ending of speech when the transmission and the reception overlap are inevitable. Yes, the two-way simultaneous callability is greatly impaired. If the amount of attenuation of the transmission attenuator and the reception attenuator is reduced, howling is likely to occur. The present invention has been conceived in view of the above circumstances, and is a loudspeaker communication device provided with an echo canceller and an echo suppressor, which has a relatively simple and inexpensive structure, and is easily reverberated and has a lot of background noise. An object of the present invention is to provide a loudspeaker communication device capable of performing stable and good calls even when installed in a circumstance.

[0011]

In order to achieve the above-mentioned object, a first aspect of the present invention transmits a transmission signal input from a microphone to a line section via a transmission signal path, and
A loudspeaker communication device for outputting a reception signal output from the line section to a speaker via a reception signal path, the acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone, and / or Alternatively, a line echo cancel circuit having a filter for canceling a line echo signal output from the reception signal path to the line section and sneaking into the transmission signal path, and input from the microphone to the transmission signal path A transmission signal variable attenuator for attenuating the transmission signal, a reception signal variable attenuator for attenuating the reception signal input from the line section to the reception signal path,
In a loudspeaker system including a control means for controlling the transmission signal variable attenuator and the reception signal variable attenuator while maintaining a predetermined complementary relationship, the speech communication device is obtained by using a predetermined pseudo signal. Line coupling amount measuring means for measuring the line coupling amount from the transmission signal path to the reception signal route via the line section, and the acoustic coupling amount from the speaker to the microphone obtained by using a predetermined pseudo signal Based on the measurement output of the line coupling amount measuring means and the measurement output of the acoustic coupling amount measuring means, and the attenuation amount of the transmission signal variable attenuator and the reception signal variable attenuation. A loudspeaker communication device, comprising: a gain setting means for setting a sum with an attenuation amount of a device and sending the set amount to the control means when a call is not started. A second aspect of the present invention is a loudspeaker communication device that transmits a transmission signal input from a microphone to a line section via a transmission signal path and outputs a reception signal output from the line section to a speaker via a reception signal path. And an acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone and / or a line echo output from the reception signal path to the line section and sneaking into the transmission signal path. A line echo cancellation circuit having a filter for canceling a signal, a transmission signal variable attenuator for attenuating a transmission signal input from the microphone to the transmission signal path, and the line section to the reception signal path A reception signal variable attenuator for attenuating an input reception signal, and the transmission signal variable attenuator In a loudspeaker communication device having a control means for controlling a voice-receiver signal variable attenuator while maintaining a predetermined complementary relationship, a constant voice signal input to the microphone is connected to the voice signal path. A level difference between the maximum value and the effective value in the time range is detected, the level difference becomes smaller than a predetermined second threshold value, and the level of the transmission signal in the transmission signal path is A howling detection means for outputting a howling detection signal in response to a predetermined second threshold value or more, and the control means outputs the howling detection signal from the howling detection means. When output, the voice communication device is controlled so that the sum of the attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator increases.

A third aspect of the present invention is a signal path switching for switching between a loudspeaker communication unit and a handset communication unit, each having a microphone and a speaker, and a transmission signal route and a reception signal route for connecting both the communication units and the line unit. A loudspeaker communication device with a handset capable of switching between a loudspeaker call and a handset call, the acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone, and / or Alternatively, a line echo cancel circuit having a filter for canceling a line echo signal output from the reception signal path to the line section and sneaking into the transmission signal path, and input from the microphone to the transmission signal path A transmission signal variable attenuator for attenuating a transmission signal; Control unit for controlling the reception signal variable attenuator for attenuating the reception signal input to the reception signal path from the control unit, and for controlling the transmission signal variable attenuator and the reception signal variable attenuator while maintaining the complementary relationship therebetween. And a sidetone coupling amount changing means for changing the sidetone coupling amount from the transmission signal path to the reception signal path according to the switching state of the signal circuit switching means. It is a voice communication device characterized by the above.

[0013]

According to the first aspect of the invention, based on the measurement output of the line coupling amount measuring means and the measurement output of the acoustic coupling amount measuring means, the attenuation amount of the transmission signal variable attenuator and the reception signal variable can be changed. A sum with the attenuation amount of the attenuator is set, and the control means controls the transmission signal variable attenuator and the reception signal variable attenuator while maintaining a predetermined complementary relationship based on the set value. The line coupling amount is obtained by outputting a predetermined pseudo signal to the transmission signal path, detecting the reception signal when the signal is sent to the line portion and spills into the reception signal path and is input, It is measured based on the relationship between the transmitted signal and the received signal. The acoustic coupling amount is obtained by outputting a predetermined pseudo signal to the reception signal path, detecting the transmission signal when the signal is output from the speaker and sneaked into the microphone, and the reception signal is received. It is measured based on the relationship between the signal and the transmitted signal. The measured acoustic coupling is
It is a value that reflects the canceling effect of the acoustic echo canceling circuit, and the line coupling amount is also measured when the device is equipped with a line echo canceling circuit. It
Therefore, the effect of acoustic echo cancellation is reflected in the set value of the sum of the attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator, which is determined by the gain setting means at the initial setting, for example. The attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator during a call are controlled by the set values, and the attenuation amount does not change from time to time due to changes in the environment during a call, and a stable call is performed. Can be done.

According to the second aspect of the invention, the howling detecting means detects the level difference of the maximum value of the transmission signal with respect to the effective value within the fixed time input to the microphone, that is, the peak coefficient. For general audio signals, the peak coefficient is about 12d
B, 3 dB for pure tone. Since the howling sound is close to a pure tone, the above coefficient when the howling is generated and the circuit is in an oscillating state at a certain frequency is smaller than the peak coefficient in a normal communication state. Therefore, when the transmission signal level is higher than the predetermined first threshold value and the level difference (peak coefficient) becomes equal to or lower than the predetermined second threshold value, the howling detection means Judge that howling has occurred. When howling is detected in this way, the howling control means reduces the loop gain by changing the amount of attenuation in the transmission signal path or the reception signal path, and suppresses howling. According to the third invention, the sidetone coupling amount is switched to an appropriate value according to the switching state of the signal circuit switching means for switching between the handsfree state and the handset state. In this way, it is possible to make calls without feeling uncomfortable, and in the hands-free call state, the half-duplex operation does not cause annoying states such as head and tail breaks and inconveniences such as howling. Good callability is secured.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the drawings. [First Embodiment] FIG. 1 is a block diagram of a speakerphone (hands-free telephone) according to a first embodiment of the present invention.
In the figure, a speakerphone 1 includes a microphone 2 for inputting a transmission voice and a speaker 3 for outputting a reception voice. Each of the microphone 2 and the speaker 3 is connected to a telephone line via a hybrid transformer 5. In a transmission signal path 9 connecting the microphone 2 and the hybrid transformer 5, a microphone amplifier 11, a transmission signal variable attenuator 13 (“speech attenuator” in the figure), a subtracter 12, a transmission amplifier 14, and an adder 16 are provided. And are provided,
A speaker amplifier 17, an adder 18, a receiver signal variable attenuator 19 (“receiver attenuator” in the figure), a receiver amplifier 20, and a subtractor 21 are provided in the receiver signal path 15 connecting the speaker 3 and the hybrid transformer 7. It is provided. Speakerphone 1
Further, an attenuation controller 22 for preventing howling in a loud voice call, a coupling amount measuring unit 23 for measuring an acoustic coupling amount and a line coupling amount, comparators 27 and 29, and a howling detector 41, Adaptive filters 43 and 47,
Coefficient setting devices 45 and 49 for acoustic echo and line echo
Are provided. As shown in FIG. 2, the coupling amount measuring unit 23 includes a signal generator 31, 33 and an acoustic coupling amount measuring device 35.
And a line coupling amount measuring device 37. Adaptive filter 4
3 and acoustic echo coefficient setting device 45, and adaptive filter 4
7 and the line echo coefficient setter 49 respectively form a conventionally known "acoustic echo cancel circuit" and "line echo cancel circuit". In addition, by combining the transmission signal variable attenuator 13 and the reception signal variable attenuator 19, the comparators 27 and 29, and the attenuation controller 22, "the attenuation amount of the transmission signal input from the microphone and the telephone line The function of complementarily controlling the amount of attenuation of the received signal input from ", that is, a conventionally known" echo suppressor "is configured. The part surrounded by the one-dot chain line in FIG.
Say. Further, the comparators 27 and 29 detect whether the signal on the transmitting side or the signal on the receiving side is larger, that is, the transmitting state or the receiving state, and send it to the attenuation controller 22. Attenuation controller 2
2 performs stable echo suppressor control with few malfunctions by determining which of the transmission mode and the reception mode should be performed based on these information. As another control method of the echo suppressor, for example,

1) Whether the signal is transmitted or received only by comparing the levels of signals at two points such as point A and point D, point A and point C, point B and point D, point B and point C in FIG. Judgment 2 points detection 2) In addition to the level comparison between 2 points shown in 1) above, point A
Alternatively, three-point detection in which the sound of the signal at the point C is detected can be considered. [Echo suppressor and echo canceller] As described above, in "complementarily controlling the attenuation amounts of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19,""contents of complementary control" is important. is there. For example, when maintaining a complementary relationship in which the sum of the attenuation amounts of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19 is constant, it is important how much the sum is set. Therefore, it is important how to determine the necessary and sufficient value according to the actual situation, that is, to set it so as to avoid the disadvantage due to the unnaturalness of the call. In the case where the following features, which are the features of this embodiment, are not provided, the two variable attenuators 13 and 13 are designed on the safe side so that howling does not occur in consideration of variations in the environment in which the telephone or the like is placed. The sum of the attenuation amounts of 19 becomes unnecessarily small, and there is a great risk that the call becomes unnatural. Therefore, in the present embodiment, the sum of the attenuation amounts of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19 is properly initialized according to the environment in which the speakerphone 1 is placed, so that the necessary minimum The unnaturalness (if conditions are good, it is also possible to realize a call that does not feel completely unnatural), it enables a more natural call.

Next, a method for calculating the sum of the attenuation amounts of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19 will be described. When the speakerphone 1 is powered on and connected to the telephone line, the signal generator 33 shown in FIG. 2 sends a signal for measuring the line coupling amount, for example, white noise as a speech signal according to a command from the attenuation controller 22. Call point E on route 9. This signal is sent to the telephone line and enters the receiving signal path 15 as a side tone by the line coupling. The line coupling amount is obtained by the line coupling amount measuring device 37 from the difference between the level of the signal transmitted in the transmission signal path 9 and the level of the signal detected in the reception signal path. Similarly, in response to a command from the attenuation controller 22, the signal generator 31 transmits a signal for measuring the acoustic coupling amount, for example, white noise, to the point F in the reception signal path 15. The signal is radiated as a sound from the speaker 3, a part of which is incident on the microphone 2 and is transmitted to the transmission signal path 9. The acoustic coupling amount is determined by the acoustic coupling amount measuring device 35 from the level of the signal transmitted in the reception signal path 15 and the level of the signal detected in the transmission signal path 9. The attenuation controller 22 is the acoustic coupling amount measuring device 3
5 and the measured output from the line coupling amount measuring device 37 and the known gains of all parts in the hands-free circuit, the attenuation amount of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19 are obtained.
The sum G and the attenuation amount are determined, that is, the gain is set. In this case, the value to be considered in the calculation is the gain GMA of the microphone amplifier 11 of the handsfree circuit.
(Known), line coupling amount GST (unknown), gain GSA of speaker amplifier 17 (known), acoustic coupling amount GAC (unknown), other gain GT existing in the transmission signal path (known), reception signal path Other gains GR (known), gains GHT (known) given to the transmission signal by the hybrid transformer 5, gains GHR (known) given to the reception signals by the hybrid transformer 5, the return amount RL to the line
(Known) There is a desired howling margin HM (Known). In the gain setting, first, there is a demand for HM such as "I want to set it to 6 dB". To meet this,
Determine G using the above values. And, of these values, the line coupling amount GST and the acoustic coupling amount GAC are
One of the features of this embodiment is that the attenuation due to the effects of the acoustic echo canceller and the line echo canceller is added. That is, the adaptive filters 43 and 47 functioning as the acoustic echo canceller and the line echo canceller receive the coefficients corresponding to the signals detected at the points G and H from the acoustic echo coefficient setter 45 and the line echo coefficient setter 49, respectively. A cancellation signal is generated from this signal and the signals detected at points G and H, and points I and J (that is, subtracters 12 and 2) are generated.
1), but in this case, the coefficient setter 45,
Reference numeral 49 indicates a so-called converging state in which an optimum coefficient can be set for the detected signal after detecting the signal in order to gradually learn and determine the output coefficient each time the signal is detected at points G and H. It will take some time to get there. Therefore, in the present embodiment, when the pseudo signal is generated from the signal generators 31 and 33 at the time of initial setting to measure the coupling amount of the sound and the line (described above), the signal for measuring the coupling amount is described above. The measurement value of the coupling amount is settled at the time when the learning of the coefficient setters 45 and 49 converges.

Using the above values, the transmission signal variable attenuator 3
The sum G of the attenuation amount GTX (dB) and the attenuation amount GRX (dB) of the reception signal variable attenuator 6 is G = GTX + GRX> HM + GMA + GT + GST + GR + GSA + GAC G = GTX + GRX> -RL + GHR + GR + GSA + GAC + GAM + GMA + GMA + GMA + GMA + GAM + (However, howling allowance H
In the gain setting, the M and the return amount RL are first determined in accordance with the requirement that the howling margin HM be 6 dB and the return amount RL be −2 dB, and are treated as known values. On the right side of the above equation, the desired howling margin HM and the known GM
A, GSA, GT, GR, and G obtained by measurement
By substituting ST and GAC, an appropriate sum of the attenuation amount GTX in the transmission signal path and the attenuation amount GRX in the reception signal path can be obtained. In the present embodiment, as described above, the transmission signal variable attenuator at the time of a call is set so that the sum GTX + GRX of the attenuation amounts of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19 becomes constant. Since the attenuation amount of 13 and the attenuation amount of the reception signal variable attenuator 19 are complementarily controlled, the change of the attenuation amount of the transmission or reception variable attenuator does not increase more than necessary, and the head of the speech is reduced as much as possible. ， Can realize a natural call with little tail cut. The concrete measuring methods of the line coupling amount and the acoustic coupling amount are as follows. In this embodiment, specifically, the line / acoustic coupling amount is calculated from the ratio of "signal magnitude" at two points in the circuit.

(1) At the time of line coupling measurement A signal is sent to point E in the circuit, the "signal magnitude" at points E and D is obtained, and (signal magnitude at point D) / (point E) Signal magnitude at) = R '. Assuming that the attenuation amounts of the paths from the point E to the hybrid transformer 5 and from the hybrid transformer 5 to the point D are negligible, the line coupling amount R is the above-mentioned attenuation rate R in the circuit and line in the hybrid transformer 5, that is, the line portion. It is synonymous with ′. still,
In the level display, the line coupling amount R is R = 20log ₁₀ R ¹
Since it is expressed in (unit dB), using the level display,
The line coupling amount R = (signal level at point D)-(signal level at point E). (2) At the time of acoustic coupling measurement The signal is sent to the point F in the circuit, the “signal magnitude” at the points F and A is obtained, and the speaker amplifier 17 → the acoustic space from the speaker 3 to the microphone 2 → the microphone amplifier 11 The attenuation rate r'in the path of (the magnitude of the signal at point A) / (the magnitude of the signal at point F)
= R '. The acoustic coupling amount r is the attenuation rate r ′ in the acoustic space from the speaker 3 to the microphone 2. The acoustic coupling amount r is 20 log ₁₀ r _{1 in the} level display.
Because represented by (unit dB), the use of level display,
Acoustic coupling amount r = (signal level at point A)-(signal level at point F)-(amplification factor level of speaker amplifier)-
It can be measured by (amplification factor level of microphone amplifier).

The above-mentioned automatic measurement of the line coupling amount and the acoustic coupling amount and the transmission signal variable attenuator 13 to be subsequently performed.
The sum of the amount of attenuation and the amount of attenuation of the reception signal variable attenuator 19 is determined when the power is turned on. In the present embodiment, in order to recognize such a state, the control unit 21 controls the points A and C.
(See FIG. 1) Detect and judge. In addition, as the timing, in addition to this, for example, when the user presses the initial setting button to instruct, the power is on, and the call is not yet performed, that is, It may be any time as long as the voice signal by the conversation of the speaker is not transmitted to either the transmission signal path 9 or the reception signal path 15. As a method of measuring the signal magnitude in the above method, for example, the following means can be considered. 1) RMS value at a certain time interval (for example, 100 ms) 2) Output value when a signal is input to a certain number of filters Noise and a signal with large fluctuation such as voice signal are used. Such processing is effective. Regarding the timing of measurement, a) Measure at the same timing at the two points to be observed. b) Consider the time delay of propagation in the circuit or space. a) has an advantage that the processing is simple, and b) has a merit that the processing is complicated, but has an advantage that a more accurate binding amount can be measured. However, the receiving signal path 1 by line coupling
The delay time of the signal entering 5 is extremely small and constant, but the delay time of the signal entering the transmission signal path 9 due to acoustic coupling always fluctuates depending on the surrounding conditions such as a person approaching, so it is accurate. In order to perform b), it is necessary to constantly detect the delay time. However, when considering the application of such time detection to an actual speakerphone configuration, it is difficult to realize it when using noise or pure tones, and when using a voice signal, use the peak of the signal as a landmark. It is important to keep in mind that it is necessary to provide extremely fine and complicated circuits such as.

In addition to the white noise described above, the signal used for measuring the amount of coupling may be, for example, a signal containing a frequency component necessary for a speech transmission band, band noise, or a pure tone. By determining the types of signals generated from the above, the following unique effects are achieved for each. For example, if the signal contains the frequency components required for the telephone transmission band, the following advantages are obtained. originally,
Acoustic coupling can occur in the frequency band that is transmitted in the telephone line, output by the speaker, and receivable by the microphone 2. This is set according to the transmission frequency band of the telephone and It is the frequency component transmitted by the telephone that the signal arrives via the telephone line, so it is necessary to measure the acoustic coupling in the signal of such frequency component. For example, by using a signal containing frequency components that are basically transmitted by telephone, such as telephone ringing tone, ringing confirmation tone, synthetic voice (guidance sound such as "starting a voice call", etc.) Accurate measurement is possible. Also, for example, the use of band noise for measurement has the following advantages. For example, when the loudspeaker 1 is expected to be used in an environment with large background noise, the frequency band expected to have a large S / N ratio (signal-to-noise ratio) (generally an environment such as fan noise of an air conditioner) Noise has a large low-frequency component, and voice is 1 KHz to 3 KHz.
Since the frequency component around is large, the S / N ratio is small in the low frequency range and often increases around 1 KHz to 3 KHz.) By using band noise (for example, center frequency of 1 kHz or 2 kHz) ， It is possible to reduce the magnitude of the generated signal. That is, the sound emitted from the speaker 3 on the near-end speaker side can be reduced during the acoustic coupling measurement, and the sound emitted from the receiver of the far-end speaker can be reduced during the line coupling measurement.

Further, by using the white noise for the measurement, the configuration of the signal generators 31 and 33 can be simplified, and the white noise includes all frequency components necessary for measuring the line coupling and the acoustic coupling. Therefore, the above-mentioned attenuation control can be properly performed without impairing the measurement accuracy. Also, by using pure tones for measurement, the configuration of the signal generator can be made much simpler than the above signals. In particular, when an LSI is used for the hybrid transformer 5, the line coupling amount is almost constant regardless of the frequency, so the line coupling amount measured using a pure tone (single frequency sine wave) that is easily generated. In practice, a sufficiently accurate value can be obtained. The specific operation of the speakerphone 1 configured as described above is as follows. (1) At the time of initial setting (when the power is turned on or the line is connected), the acoustic coupling amount measuring instrument 35 and the line coupling measuring instrument 37 automatically measure the acoustic coupling GAC and the line coupling GST. Based on this measured value, the attenuation controller 22 recognizes the acoustic environment and line environment in which the telephone is installed, and the maximum value of the attenuation amount allowed for the attenuators 13 and 19 from the requirement of howling margin. , Sum of minimum values is set. Further, the minimum value of the attenuation amount is determined by the requirements of the reception gain from the line to the speaker 3 and the transmission gain from the microphone 2 to the line. The maximum value and the minimum value of this attenuation amount are determined based on "a method for determining the sum of the attenuation amount in the transmission signal variable attenuator 13 and the attenuation amount in the reception signal variable attenuator 19" which will be described later. , It is decided as described below according to the call state such as the listening mode. In this embodiment, each attenuation amount is determined on the basis of a complementary relationship that keeps the "sum" of the above attenuation amounts constant, but the complementary relation is only the above "case where the two sums are constant". Otherwise, control is performed so that the sum of both attenuations does not exceed a certain value. For example, in the transmission state, the attenuation amount of the transmission signal variable attenuator 13 is maximized, the attenuation amount of the reception signal variable attenuator is minimized, and the attenuation amount of the transmission signal variable attenuator 13 is minimized in the reception state. It is conceivable that the amount of attenuation of the received signal variable attenuator is maximized, and the amount of attenuation of both variable attenuators is maximized in other states (such as when both callers are silent). (2) When a call is started, the above-mentioned normal echo suppressor operation (so-called "shallow voice switch") is performed, and at the same time, processing for acoustic echo cancellation and line echo cancellation is started as follows. To be done.

A) When in the transmitting state, a line echo coefficient setting device is used by using a learning identification method or the like based on the correlation between the transmitting signal output to the line and the line echo signal sneaking into the receiving line 15 by line coupling. The coefficient of the adaptive filter 47 is updated by 49. The line echo cancellation by the applied filter 47 works based on the updated coefficient. b) When in the receiving state, a learning identification method or the like is used from the correlation between the receiving signal supplied to the speaker 3 and the acoustic echo signal generated by the microphone 2 due to the acoustic coupling between the speaker 3 and the microphone 2. Then, the coefficient of the adaptive filter 43 is updated by the acoustic echo coefficient setting unit 45. Acoustic echo cancellation by the adaptive filter 43 operates based on the updated coefficient. c) In the silent state and the double talk state, the coefficients of the adaptive filters 43 and 47 are not updated, and the line and the acoustic echo cancel function with the same coefficient as before. <Howling detection and suppression> During a call, if background noise increases and acoustic echo cancellation does not work well, or if line noise increases and line echo cancellation does not work well, the total loop gain becomes positive. ， May cause howling. As a howling loop, Near Howling Loop in the telephone
And Far caused by reflections from the exchange or the other party's telephone.
Howling Loop can be considered. The former becomes unable to talk. In the latter case, in addition to the inability to call, the whole line is annoying. Therefore, howling must be eliminated immediately. Howling detection and suppression in this embodiment will be described below.

In FIG. 1, a receiving signal generated by a far-end speaker at the time of receiving a call passes through a hybrid transformer 5 from a telephone line, passes through a receiving signal variable attenuator 19, is amplified by a speaker amplifier 17, and then is output from a speaker 3. Is output. A part of the signal is input from the microphone 2 into the transmission signal path 9. At this time, if the amount of acoustic coupling from the speaker 3 to the microphone 2 fluctuates significantly due to the movement of the near-end speaker, the loop gain in the hands-free circuit increases from the initially set value, and the howling is scheduled. It is not possible to keep a margin and howling occurs. In order to detect the occurrence of howling, the howling detector detects a level difference (hereinafter referred to as a peak coefficient) of the maximum value (peak value) of the signal level input from the microphone 2 into the transmission signal path 9 within a certain period of time. Detected at 41. When the peak coefficient becomes smaller than a predetermined first threshold value, for example, 6 dB, it can be said that not only the audio signal but also the pure tone component is large, that is, howling may occur. However, for example, when the handset is picked up before making a call, or when the other party sends a tone signal to operate a telephone terminal or the like during a call, a signal close to a pure tone is spilled from the speaker 3. Since it is input to the microphone 2, if howling is determined only by the pure tone of the signal, it may be erroneously detected. Therefore, in the present embodiment, the howling detector 41 determines that the peak coefficient has the above-mentioned first threshold value when the condition that the signal level in the transmission signal path 9 is higher than the predetermined second threshold value is satisfied. When the condition that the value is smaller than the threshold value is satisfied, howling is determined to occur. This first
The threshold value may be set according to the characteristics of the speakerphone 1 and the installation environment. The howling detector 31 is
When the occurrence of howling is detected, a signal indicating that is sent to the attenuation controller 22. When the occurrence of howling is transmitted from the howling detector 31 to the attenuation controller 22, the attenuation controller 22 inserts a loss in the transmission signal path 9 or the reception signal path 15 to reduce the loop gain in the handsfree circuit. Lower and suppress howling. Specifically, the sum of the attenuation amount of the transmission variable attenuator 13 and the attenuation amount of the reception signal variable attenuator 19 is reduced. That is, the amount of attenuation of the reception signal variable attenuator 6 is increased during transmission, and the amount of attenuation of the transmission signal variable attenuator 13 is increased during reception. In this embodiment, each attenuation amount is increased by 10 dB. In this way, by temporarily approaching the so-called half-duplex operation state (deep voice switch), the loop gain becomes 1
Howling is suppressed as follows. When the suppression of howling is completed, the attenuation controller 22 executes the same procedure as in the initial setting described above to newly set the suitable attenuation amounts of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19. It is also possible to set to. In this embodiment, the number of comparators is 2
When a digital circuit is used, since the attenuation amounts of the transmission / reception attenuators 13 and 19 are known in advance, a comparator for comparing the respective inputs of the transmission attenuator 13 and the reception attenuator 19 is used. You may prepare only one. Further, although the attenuation amounts of the attenuators 13 and 19 are set at the time of initial setting, for example, when the external volume of the speaker amplifier may be changed during use, a predetermined timing can be set within a range permitted by the calculation amount. Alternatively, the attenuation amount may be set intermittently.

[Second Embodiment] Next, a second embodiment of the present invention will be described with reference to FIG. This embodiment applies the functions of the first embodiment to a speakerphone with a handset. Note that, in FIG. 3, a portion surrounded by a one-dot chain line shows that the configuration is the same as that of the attenuation adjusting circuit 50 of FIG. 1, and the illustration of the inside thereof is omitted. The speakerphone 101 includes a main body 105 and the main body 105.
And a handset 103 connected to the. The main body 105 is provided with a speaker 109b and a microphone 111b, and a transmission signal captured by the microphone 111b is amplified by an amplifier 121b, and then a switch 115 inserted in a transmission signal path 113,
The transmission attenuator 113 and the amplifier 114 of the attenuation adjustment circuit 50
It goes through to the hybrid transformer 107 which is a line section and is sent out to the line. The received signal from the other speaker is sent from the outside line to the hybrid transformer 107, and is inserted in the received signal path 117 to the attenuation adjustment circuit 50.
Is attenuated by the amplifier 120 and the reception attenuator 119 (see FIG. 1) of FIG.
It is transmitted to the speaker 109b via 3b. The speaker 109b and the microphone 111b on the main body side constitute a loudspeaker communication section. The handset 10
3 side speaker 109a and microphone 111a
And switches 119 and 115 which are switched and controlled by a switch 149 (signal circuit switching means).
It is connected to the reception signal path 117 and the transmission signal path 113 via. The speaker 109a and the microphone 111a form a handset communication unit. Therefore, the speaker can switch the switches 119 and 115 by the switch 149 to switch between the handset call using the microphone 111a and the speaker 109a of the handset 103 and the voice call using the microphone 111b and the speaker 109b on the main body side. it can. By the way, when talking in a handset state using the handset 3, the sound transmitted by the speaker returns to the reception signal path to some extent, and it is easier for the speaker to speak while listening to his own voice. The quantity is set to a large state. In a handset call, since the speaker's ear and the speaker 109a are close to each other and the speaker's mouth and the microphone 111a are close to each other, the volume of the speaker 109a and the gain of the amplifier 121a may be small, and howling may occur even if the sidetone coupling amount is large. Absent.

On the other hand, even in the hands-free state, the sidetone coupling amount remains the same as that in the handset state, but in this case, the speaker separates from the microphone 111b and the speaker 109b, and therefore the speaker 109b.
And the gain of the amplifier 121b are set to be large. Further, since there is no shielding effect by the face of the speaker, acoustic coupling between the speaker 109b and the microphone 111b becomes large. As a result, the loop gain formed by the reception signal path 17 and the acoustic coupling from the transmission signal path 13 via the sidetone coupling increases, and howling is likely to occur. Therefore, in the hands-free state, the attenuation amount of the transmission signal variable attenuator 13 and the reception signal variable attenuator 19 is large, that is, the transmission volume and the reception volume are compared by the comparators 27 and 29, and the transmission volume is changed. In the state, it is considered that the attenuation amount of the reception attenuator 19 is maximized, and in the reception state, the attenuation amount of the transmission attenuator 33 is maximized, that is, so-called half-duplex operation is performed. . However, in order to reduce the loop gain in the hands-free state, simply by increasing the attenuation amount of the transmission attenuator 13 and the reception attenuator 19, the half-duplex operation is used. ，
As a result, there is an unavoidable problem that the beginning of a talk and the ending of a tail are inevitable when the transmission and reception overlap, and the two-way simultaneous callability is greatly impaired. If the attenuation amounts of the transmission attenuator 33 and the reception attenuator 37 are reduced, howling is likely to occur.

In order to solve the above problems in the prior art, the present invention changes the degree of sidetone coupling depending on whether it is a handset state or a handsfree state, and a transmission attenuator, By reducing the amount of attenuation of the receiving attenuator, the purpose of the present invention is to reduce the beginning and ending of a word without causing howling. Therefore, in the present embodiment, the transmission signal path 113 and the reception signal path 1
A bypass circuit 143 connecting 17 is provided in parallel with the hybrid transformer 107. A switch 145 and a bypass impedance 147 are inserted in series in the bypass circuit 143. Switch 1 above
45 is turned on by the switching device (switching means) 149.
Controlled off. Therefore, as in the prior art, when the switches 115 and 119 for switching the hands-free state and the handset state are switched by the switch 149, the switch 145 is also switched at the same time. In this embodiment, as shown in FIG. 3, a handset state (a state in which signals are input / output to / from the handset 103)
, The switch 145 is turned on, the transmission signal path 113 and the reception signal path 117 are connected via the impedance of the hybrid transformer 107 and the impedance of the bypass impedance 147 in parallel therewith, and the transmission signal path 113 The sidetone coupling amount returning to the reception signal path 117 is determined by the two impedances in parallel. In this way, since the sidetone coupling amount increases in the handset state, the speaker can speak while listening to his / her own voice that circulates through the bypass impedance 147 and the hybrid transformer 107, and can maintain a state in which it is easy to speak. Bypass impedance 1 in this state
The impedance (sidetone coupling amount) calculated from both 47 and the hybrid transformer 107 is, for example, -7 to -15.
dB is considered appropriate. On the other hand, hands-free state (state to input signal from microphone 111b)
, The switch 145 is turned off, and the transmission signal path 113 and the reception signal path 117 are connected only via the impedance of the hybrid transformer 107.
Howling is prevented because the side sound coupling amount returning from the transmission signal path 113 to the reception signal path 117 is small. In this state, the sidetone coupling amount based on the impedance of the hybrid transformer 107 is appropriately set to, for example, -30 to -20 dB.

In the above embodiments, the sidetone coupling amount changing means is composed of the bypass impedance 147 and the switch 145 connected in parallel to the hybrid transformer 107, and the hybrid transformer 107 and the bypass impedance 147 are connected in series. Although the case has been described, this is an example, and the bypass impedance 147 and the switch 145 can be connected in parallel. Further, in the above embodiment, the present invention has been described based on a loudspeaker telephone device with a handset. However, the present invention is a system that enables a hands-free call and a handset call, such as a conference call system. Of course, the present invention can be applied to the system, and in that case, the gist of the present invention can be applied to the system without any change. The embodiments of the present invention have been described above. The characteristic points of the configuration of the above embodiment will be described below. Speakerphone 1 of the first embodiment
Corresponds to a loudspeaker communication device that transmits a transmission signal input from a microphone to a line section via a transmission signal path and outputs a reception signal output from the line section to a speaker via a reception signal path. The adaptive filter 43 and the acoustic echo coefficient setting device 45 constitute an acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone, and the adaptive filter 47 and the line echo coefficient setting device. 49, a line echo cancel circuit having a filter for canceling a line echo signal output from the reception signal path to the line section and sneaking into the transmission signal path. The transmission signal variable attenuator 13 constitutes a transmission signal variable attenuator for attenuating the transmission signal input from the microphone to the transmission signal path, and the reception signal variable attenuator 19
A reception signal variable attenuator configured to attenuate a reception signal input from the line unit to the reception signal path is configured. A loudspeaker communication device comprising control means for controlling the transmission signal variable attenuator and the reception signal variable attenuator while maintaining a predetermined complementary relationship by the attenuation controller 22, wherein the howling detection is performed. When a howling detection signal is output from the means, control is performed so that the sum of the attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator becomes larger than the set value of the gain setting means. Means, and the sum of the attenuation amount of the transmission signal variable attenuator and the reception signal variable attenuation amount is set based on the measurement outputs of the line coupling amount measuring device and the acoustic coupling amount measuring device. , Gain setting means for sending the set value to the control means when a call is not started. A line coupling amount for measuring the line coupling amount from the transmission signal path obtained from the predetermined pseudo transmission signal and reception signal by the line coupling amount measuring device 37 to the reception signal route via the line section. Measuring means is configured, and the acoustic coupling amount measuring device 3
5 constitutes the acoustic coupling amount measuring means for measuring the acoustic coupling amount from the speaker to the microphone in the environment where the speaker and the microphone are installed, which is obtained by using a predetermined pseudo signal. The howling detector 41 connects to the transmission signal path,
The level difference between the maximum value of the transmission signal input to the microphone and the effective value within a certain time range is detected, and the level difference becomes smaller than a predetermined first threshold value. In response to the level of the transmission signal in the speech signal path being equal to or higher than the second predetermined threshold value,
A howling detection unit that outputs a howling detection signal is configured. Further, the speakerphone 101 of the second embodiment has a loudspeaker communication unit and a handset communication unit each having a microphone and a speaker, and a transmission signal path and an reception signal path for connecting both the communication units and the line unit. It corresponds to a voice communication device with a handset, which is equipped with a signal path switching means for switching and can switch between a voice call and a handset call. The bypass circuit 143 constitutes sidetone coupling amount changing means for changing the sidetone coupling amount from the transmission signal path to the reception signal path according to the switching state of the signal circuit switching means. .

[0029]

As described above, according to the voice communication device of the present invention, the line coupling amount and the acoustic coupling amount are improved by outputting the pseudo signal in the transmission signal path and the reception signal path. The sum of the attenuation amount of the transmission signal variable attenuator and the attenuation amount of the reception signal variable attenuator is automatically measured using the line coupling amount and the acoustic coupling amount measured in the actual use environment of the loudspeaker. Since the amount of attenuation is set according to the usage environment to obtain the desired howling margin, the sum of the amounts of attenuation of the transmission variable attenuator or the reception variable attenuator is set to a larger value than necessary. There is no end. Therefore, the change in the attenuation amount at the time of switching between transmitting and receiving is relatively small, and head disconnection and tail disconnection are less likely to occur, and natural communication can be performed. Moreover, the measurement of the acoustic coupling amount and the line coupling amount and the setting of the sum of the attenuation amounts of the transmission variable attenuator and the reception variable attenuator corresponding thereto are performed at a suitable predetermined timing, for example, at the time of initial setting. Therefore, it is not necessary to measure the acoustic coupling amount and the line coupling amount at all times during a call and to change the attenuation amount during the call accordingly, thereby simplifying the line configuration and The amount of calculation processing can also be reduced, and the assembly cost of the device can be reduced. Moreover, the measured acoustic coupling amount is a value that reflects the canceling effect of this circuit when the acoustic echo cancellation circuit is provided, and the line coupling amount is also provided with a line echo cancellation circuit in the device. If so, the value that reflects the cancellation effect is measured. Therefore, the effect of acoustic echo cancellation is reflected in the set value of the sum of the attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator, which is determined by the gain setting means at the initial setting, for example. Attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator during a call are controlled by the set values, and a stable call can be performed without the attenuation amount changing from time to time due to changes in the environment during a call. You can do it. Further, according to the present invention, the occurrence of howling is detected by detecting both the level difference between the maximum value and the effective value of the signal in the transmission signal path and the effective value of the signal, so the risk of false detection is small. Also, since no complicated algorithm is required,
It is possible to realize a stable howling detection and suppression function at a relatively low cost. Therefore, when the voice communication device has the echo suppressor function by the variable attenuators provided in each of the transmission signal path and the reception signal path, the acoustic coupling amount from the speaker to the microphone may be changed by, for example, the movement of the near-end speaker. Due to the large fluctuation and the change in the loop gain, the setting value of the attenuation amount of the variable attenuator becomes inappropriate,
Even when the howling margin cannot be maintained, the howling can be detected by the function described above, and the howling suppression function can reduce the loop gain to 1 or less to suppress the howling. Furthermore, when the voice communication device has an acoustic echo canceller function, for example, the amount of acoustic coupling fluctuates greatly due to the movement of the near-end speaker,
Even when howling occurs in the circuit due to the fact that the coefficient updating process of the adaptive filter for the echo canceller fails to catch up and howling occurs in the circuit, the howling can be detected and suppressed in a similar manner. further,
When the voice communication device is equipped with a handset, the hands-free state switches to the handset state, and at the same time, the sidetone coupling amount switches to a value suitable for the handset state. On the contrary, the handset state is switched to the handsfree state, and at the same time, the value is switched to a value suitable for the handsfree state. Therefore, in either state, communication is improved and howling is prevented from occurring due to switching.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a speakerphone according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of a binding amount measurement unit.

FIG. 3 is a block diagram showing a configuration of a speakerphone with a handset according to a second embodiment of the present invention.

FIG. 4 is a circuit diagram showing a configuration of a conventional voice amplification device.

FIG. 5 is a circuit diagram showing a configuration of a conventional hands-free circuit.

FIG. 6 is a circuit diagram showing a configuration of a conventional hands-free circuit.

FIG. 7 is a circuit diagram showing a configuration of a conventional hands-free circuit.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Speakerphone 2 ... Microphone 3 ... Speaker 5 ... Hybrid transformer (line part) 9 ... Received signal path 11 ... Microphone amplifier 13 ... Transmission signal attenuator 15 ... Transmission signal path 17 ... Speaker amplifier 19 ... Received signal variable attenuation Device 21 ... Control unit 23 ... Coupling amount measuring unit 25 ... Receiving side voice-to-noise detector 27, 29 ... Level comparator 31, 33 ... Signal generator 35 ... Acoustic coupling amount measuring device 37 ... Line coupling amount measuring device 41 ... Howling detector 43, 47 ... Adaptive filter 45 ... Acoustic echo coefficient setting device 49 ... Line echo 50 ... Attenuation adjusting circuit 101 ... Loudspeaker / telephone communication device with handset 103 ... Handset 107 ... Hybrid transformer (line part) 109a ... Microphone for handset (handset) Call part) 109b ... Hands-free microphone Hands-free call unit 111a ... Handset speaker (handset call unit) 111b ... Hands-free speaker (hands-free call unit) 113 ... Transmission signal path 117 ... Reception signal path 149 ... Signal path switching means 145 ... Switch (side tone) Coupling amount changing means) 147 ... Bypass impedance (side tone coupling amount changing means)

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Toru Sakata 1-5-5 Takatsukadai, Nishi-ku, Kobe-shi, Hyogo Kobe Steel Co., Ltd. Kobe Research Institute (72) Inventor Toshiaki Shimoda Takatsuka, Nishi-ku, Kobe-shi, Hyogo 1-5-5 Taiwan Kobe Works, Kobe Steel Co., Ltd. (72) Hideo Utsuno 1-5-5 Takatsukadai, Nishi-ku, Kobe City, Hyogo Prefecture Kobe Steel Works, Kobe Steel Co., Ltd.

Claims (3)

[Claims] 1. A loudspeaker communication device for transmitting a transmission signal input from a microphone to a line section via a transmission signal path and outputting a reception signal output from the line section to a speaker via a reception signal path. , And an acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone and / or a line echo signal output from the reception signal path to the line section and sneaking into the transmission signal path. A line echo canceling circuit having a filter for canceling a voice signal, a transmission signal variable attenuator for attenuating a transmission signal input from the microphone to the transmission signal path, and an input from the line section to the reception signal path A reception signal variable attenuator for attenuating the received reception signal, the transmission signal variable attenuator, and In a loudspeaker communication apparatus including a control means for controlling a voice-receiver signal variable attenuator while maintaining a predetermined complementary relationship, a line section is provided from the voice-transmitting signal path obtained by using a predetermined pseudo signal. Line coupling amount measuring means for measuring the line coupling amount reaching the reception signal path through the acoustic coupling amount measuring means for measuring the acoustic coupling amount from the speaker to the microphone obtained by using a predetermined pseudo signal. And the sum of the attenuation amount of the transmission signal variable attenuator and the reception signal variable attenuator based on the measurement output of the line coupling amount measuring means and the measurement output of the acoustic coupling amount measuring means. A loudspeaker communication device comprising: a gain setting unit that sets the set amount and sends the set amount to the control unit when a call is not started. 2. A loudspeaker communication device for transmitting a transmission signal input from a microphone to a line section via a transmission signal path and outputting a reception signal output from the line section to a speaker via a reception signal path. , And an acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone and / or a line echo signal output from the reception signal path to the line section and sneaking into the transmission signal path. A line echo canceling circuit having a filter for canceling a voice signal, a transmission signal variable attenuator for attenuating a transmission signal input from the microphone to the transmission signal path, and an input from the line section to the reception signal path A reception signal variable attenuator for attenuating the received reception signal, the transmission signal variable attenuator, and In a loudspeaker communication device having a control means for controlling a voice-receiver signal variable attenuator while maintaining a predetermined complementary relationship, a constant voice signal input to the microphone is connected to the voice signal path. A level difference between the maximum value and the effective value in the time range is detected, the level difference becomes smaller than a predetermined second threshold value, and the level of the transmission signal in the transmission signal path is A howling detection means for outputting a howling detection signal in response to a predetermined second threshold value or more, and the control means outputs the howling detection signal from the howling detection means. A loudspeaker communication device, wherein when output, the sum of the attenuation amounts of the transmission signal variable attenuator and the reception signal variable attenuator is increased. 3. A loudspeaker communication unit and a handset communication unit each having a microphone and a speaker, and signal path switching means for switching between a transmission signal path and a reception signal path for connecting both the communication units and the line unit. A loudspeaker communication device with a handset capable of switching between a loudspeaker call and a handset call, the acoustic echo cancellation circuit having a filter for canceling an acoustic echo signal from the speaker to the microphone, and / or
Alternatively, a line echo cancel circuit having a filter for canceling a line echo signal output from the reception signal path to the line section and sneaking into the transmission signal path, and input from the microphone to the transmission signal path Variable transmission signal attenuator for attenuating the transmission signal, variable reception signal attenuator for attenuating the reception signal input from the line section to the reception signal path, variable transmission signal attenuator and variable reception signal In a loudspeaker communication apparatus including a control means for controlling an attenuator and a complementary relationship between the two, a transmission signal path from the transmission signal path to the reception signal path depending on a switching state of the signal circuit switching means. A loudspeaker communication apparatus comprising sidetone coupling amount changing means for changing the sidetone coupling amount.