JPH10285083A - Voice communication equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the voice communication equipment that realizes an echo cancel function without providing a sense of speech incongruity to an opposite party by reducing a discontinuous sense of a speech signal and production of a click tone. SOLUTION: A center clip control circuit 14 calculates a mean level of residual signals ESS outputted from a subtractor 8 of an echo canceller 16 and attenuated by an attenuator 12 for a prescribed period in a transmission signal path. Furthermore, a clip level used by a center clip circuit 15 is controlled to be changed based on the mean level of the residual signals ESS for a prescribed period, a result of discrimination of a speech state of a single talk or a double talk by a double talk detection circuit 10, and a voice switch level VSL given from a voice switch control circuit 13 as control information representing the attenuation of the attenuator 12 inserted to the transmission signal line.

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 having a hands-free call function, and more particularly, to an improvement in a method for suppressing an acoustic echo generated during a hands-free call.

[0002]

2. Description of the Related Art Conventionally, for example, some wireless telephone devices used in automobiles have a hands-free call mode in addition to a handset call mode. In the hands-free call mode, a speaker for receiving and a microphone for transmitting are installed on a dashboard or the like, and talking is performed using these speakers and the microphone as a handset. When the hands-free call mode is used, the speaker can talk without holding the handset, and can talk without one-handed driving even during talking, which is effective in terms of safety during driving.

[0003] However, during such a hands-free call, the received voice from the speaker may be transmitted to the microphone and transmitted to the other party as an acoustic echo. In order to prevent the deterioration of call quality due to the acoustic echo, an echo canceller for canceling the wraparound received voice is required. Particularly, in a digital wireless telephone device for digitizing and transmitting voice, a signal delay caused by a codec or the like is large, and an acoustic echo inputted by wraparound is large. A canceling echo canceller is essential.

[0004] As a conventional example of an echo canceller, a relation between a received signal and an echo mixed into a transmission input signal is estimated using an adaptive filter, and the above estimation contents are used from a received signal input to the echo canceller. A method of generating a pseudo echo and subtracting it from a transmission input signal is known.

Further, when the echo cannot be eliminated even by this subtraction processing, a method of suppressing the echo by a non-linear means is adopted. A center clipper is known as this type of echo suppression function. The center clipper forcibly sets the output from its own circuit to 0 when the absolute value of the level of the input signal (residual signal) is equal to or less than the threshold (clip level), otherwise, the input signal (residual signal). Works as it is.

FIG. 7 shows a time variation characteristic of a transmission signal output level in a conventional echo canceller using a center clipper. In the figure, echo cancellation for a transmission signal, which is an input signal, is started from time t0 'at which the state switches to a single talk (reception) state, and the output signal level TS gradually decreases. afterwards,
The output signal level TS further decreases, and the center clip operates from the time t1 'at which the preset clip level CL0 is divided to the time t2' at which the double talk state is established, during which the output signal level TS is forcibly set. Becomes 0. Thereafter, a time t3 'at which the state returns to the single talk state again.
, The transmission signal output signal level TS gradually decreases, and the operation of the center clip is restarted from time t4 'when the transmission signal output signal level TS falls below the clip level CL0.

In the conventional echo canceller using such a center clipper, echo and noise can be effectively suppressed, but when the input signal level to the center clipper falls below a certain level, the transmission output is forcibly forced. Is completely removed, causing a sense of discontinuity in the call. In particular, if you are listening to a call from a partner who has in-vehicle noise such as running noise and the other party's call is interrupted, the center clip will operate, suppressing not only the echo signal but also the in-vehicle noise, and the other party's voice will suddenly be heard. The silence resulted in a sense of discomfort for the user due to the discontinuity of the call.

[0008]

As described above, in a conventional radio telephone apparatus equipped with an echo canceller, when only the other party talks, that is, in a single talk state, the center clipper is always activated and the input signal (residual error) is set. Signal) is suppressed by the center clipping operation when the level of the signal is smaller than a certain clip level, so that the transmission signal is interrupted every time the signal level of the residual signal falls below the clip level, and the other party suddenly generates voice. However, there is a problem in that the sound becomes silence and a sense of incongruity is generated due to discontinuity of the call. This discomfort was clearly recognized when the in-vehicle noise such as running noise was relatively high, resulting in a noticeable decrease in call quality. Also, when a call repeats a single talk state and a double talk state in a short time, the change in the call state causes the center clipper to work frequently, and as a result,
An unpleasant click sound was generated in the call signal, which also lowered the call quality.

The present invention eliminates the above-mentioned problems, reduces the occurrence of a discontinuity in the call signal and the generation of a click sound, and realizes a stable echo canceling function while maintaining an extremely natural call without giving the other party an uncomfortable feeling. It is an object to provide a voice communication device.

[0010]

To achieve the above object, according to the present invention, there is provided a reproducing means for reproducing a received signal, a transmitting means for generating a transmission signal from a transmission input signal,
In a voice communication device having an echo canceling unit that suppresses an acoustic echo signal generated when the received signal is input to the transmitting unit and outputs a residual signal,
A call state determination unit that determines a call state based on the level of the reception signal and the level of the residual signal output from the echo cancel unit, and a clip level in which an average level of the residual signal for a predetermined period is set. Signal clipping means for removing the residual signal when the signal level is also small, and signal clipping control means for variably setting the clip level used in the signal clipping means based on the determination result of the call state determination means. It is characterized by.

Further, according to a second aspect of the present invention, in the first aspect of the present invention, there is further provided a transmission attenuator which is inserted into a transmission signal path and which attenuates the residual signal based on a communication state. The signal clip control means variably sets the clip level based on the call state determined by the call state determination means and the attenuation of the transmission attenuator.

Further, according to a third aspect of the present invention, in the second aspect of the present invention, there is provided an attenuation amount indicating means for outputting a control signal representing an attenuation amount of the transmission attenuator when attenuating the residual signal. The signal clip control unit stores a clip level corresponding to the attenuation of the transmission attenuator, and stores the clip level corresponding to the control signal output from the attenuation instruction unit in the storage unit. And a clip level selecting means for reading out and outputting.

According to a fourth aspect of the present invention, in the third aspect of the present invention, the clip level stored in the storage means becomes smaller as the attenuation amount of the transmission attenuator increases. It is characterized by being associated so that

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A voice communication device according to the present invention will be described below in detail with reference to the accompanying drawings taking a radio telephone device as an example. FIG. 1 is a block diagram showing the configuration of an echo canceller and its peripheral circuits of a wireless telephone device according to an embodiment of the present invention. FIG. 2 is a diagram showing the overall configuration of the wireless telephone device including the circuit portion shown in FIG.

Here, the digital radio telephone apparatus shown in FIG. 2 will be described first. The transmission system of this wireless telephone device is:
A microphone 111 as a transmitter, a speech encoder (S
PCOD) 112, error correction encoder (CHCOD) 1
13, digital modulator (MOD) 114, multiplier 11
5. Power amplifier (PA) 116, high frequency switch (S
W) 117, and an antenna 118.

The speech coder 112 includes the microphone 11
The transmission signal output from 1 is encoded. The error correction encoder 113 performs error correction encoding on the digitized transmission signal output from the audio encoder 112 and the digitized control signal output from the control circuit 130.

The digital modulator 114 generates a modulation signal corresponding to the digitized transmission signal output from the error correction encoder 113. Multiplier 115 mixes the modulated signal with a local oscillation signal output from frequency synthesizer 131, and thereby converts the frequency into a radio frequency signal. Power amplifier 116 amplifies the wireless transmission signal output from multiplier 115 to a predetermined transmission power.

The high frequency switch 117 includes a control circuit 130
The transmission state is turned on only during a transmission time slot designated by the wireless communication apparatus. During this period, the wireless transmission signal output from the power amplifier 116 is supplied to the antenna 118 and transmitted from the antenna 118 to a base station (not shown).

On the other hand, the receiving system includes a receiver (RX) 1
21, digital demodulator (DEMOD) 122, error correction decoder (CHDEC) 123, audio decoder (SPD)
EC) 124 and a speaker 125 as a receiver.

The receiver 121 performs a mixing operation for directly converting the frequency of a received high-frequency signal received by the antenna 118 and the high-frequency switch 117 into an intermediate frequency signal or a baseband signal. The digital demodulator 122 performs bit synchronization and frame synchronization, that is, word synchronization with respect to the received digital signal output from the receiver 121, and further performs digital demodulation. Here, the synchronization signal obtained by the word synchronization is supplied to the control circuit 130.

The error correction decoder 123 performs error correction decoding of the digital demodulated signal output from the digital demodulator 122. The signal obtained by this error correction decoding is composed of a digital reception signal and digital control information,
The digital reception signal is input to the audio decoder 124, and the digital control information is input to the control circuit 130 for various controls such as channel setting and communication. Audio decoder 124
Performs decoding processing of a digital reception signal. The received signal decoded into an analog signal is transmitted to the speaker 1.
25.

The control system includes a control circuit (CONT) 130, a frequency synthesizer (SYN) 131, and the like.
The frequency synthesizer 131 generates a local oscillation frequency corresponding to each of the control, speech, and synchronization channel frequencies specified by the control circuit 130 and outputs the generated local oscillation frequency to each unit. The control circuit 130 performs a series of controls related to the initial setting of the device, outgoing / incoming calls, and calls. The power supply circuit 140 is a circuit that supplies a stable power supply of the digital wireless telephone device from a DC power supply of a vehicle.

Next, the circuit of FIG. 1 will be described. This circuit includes the microphone 111 side of the voice encoder 112 and the voice decoder 12 of the digital wireless telephone apparatus shown in FIG.
4 corresponds to the speaker 125 side. In the figure, 1 is D
/ A converter, 2 is a receiving amplifier, 3 is a speaker, 4 is a microphone, 5 is a transmitting amplifier, 6 is an A / D converter, 7 is an adaptive filter (ADF), 8 is a subtractor, 9 is a tap coefficient. An update unit, 10 is a double talk detection circuit, 11 is a reception attenuator, 12 is a transmission attenuator, 13 is a voice switch control circuit, 14 is a center clip control circuit, and 15 is a center clip circuit. In this circuit, the adaptive filter 7 and the subtractor 8 constitute an echo canceller body 16.

Next, the operation of this circuit will be described.
In this circuit, in the hands-free communication mode, the reception signal RS is applied to the voice switch control circuit 13 and passes through the reception attenuator 11 which is turned on / off by the voice switch control circuit 13, and then the double talk detection circuit 10 and the tap. Coefficient updating unit 9, adaptive filter 7, D /
It is supplied to the A converter 1.

The reception signal supplied to the D / A converter 1 is converted into an analog signal, and after being amplified by the reception amplifier 2,
The sound is output by the speaker 3. Here, a part of the received voice output from the speaker 3 is echo path EP.
Through the microphone 4 and input to the transmission system as an audio signal in some cases. In this case, the input signal is amplified by the transmission amplifier 5, converted into a digital signal by the A / D converter 6, and an acoustic echo signal ES is generated. This acoustic echo signal ES is supplied to the + input terminal of the subtractor 8 of the echo canceller main body 16.

On the other hand, the adaptive filter 7 simulates the acoustic characteristics of the echo path from the speaker 3 to the microphone 4 to generate a pseudo echo signal ESS, and supplies this to the minus input terminal of the subtractor 8. The generation of the pseudo echo ESS in the adaptive filter 7 is performed by the following method. First, the adaptive filter 7 adds a plurality of delay circuits connected in series, a plurality of amplifiers for amplifying the respective outputs of the delay circuits, and the outputs of the respective amplifiers, and adds the result of the addition to the pseudo echo signal ESS. And an adder that outputs as

In the adaptive filter 7, the received speech signal after passing through the reception attenuator 11 is fetched by the first delay circuit in the series connection as described above.
The signals are sequentially sent to the next-stage delay circuit. At this time, the input audio signal to the first delay circuit and the output signal of each delay circuit including the first delay circuit are supplied to the adder through the amplifier corresponding to each of the amplifiers. At this time, the acoustic characteristic of the echo path EP to be simulated can be changed by changing the amplification coefficient of each amplifier.

The change of the amplification coefficient of each of these amplifiers is controlled by the control of the tap coefficient updating unit 9. That is, the tap coefficient updating unit 9 includes at least a coefficient calculating unit and a coefficient memory, and the coefficient memory holds an amplification coefficient to be supplied to each of the amplifiers. In the tap coefficient updating unit 9, the coefficient calculating unit calculates a value based on the received voice signal RS and the output of the subtractor 8 (residual signal).
The amplification coefficient for the pseudo echo signal generation operation is adaptively changed on the coefficient memory so that the residual signal always becomes 0, and the amplification coefficient is read from the coefficient memory and supplied to the amplifiers. I do. By this adaptive operation, the acoustic characteristics of the used space, that is, the echo path is estimated, and the pseudo echo signal ESS is generated based on the estimation result and the received voice signal RS.

The subtractor 8 performs an arithmetic process of subtracting the pseudo echo signal ESS from the acoustic echo signal ES, thereby canceling the echo signal ES and realizing an echo canceling function.

The double talk detection circuit 10 receives the reception signal R
S and the level of the transmission signal (residual signal output from the subtractor 8 of the echo canceller main body 16 and passed through the transmission attenuator 12) are compared, and based on the comparison result, the voice input is performed on both the transmission side and the reception side. A single talk state in which there is a voice input to only one of the transmitting side and the receiving side, and a call state detection result DTS
Is output to the tap coefficient updating unit 9.

When the tap coefficient updating unit 9 recognizes that the state has been switched to the double talk state on the basis of the result DTS of the call state, the tap coefficient updating unit 9 keeps the amplification coefficient immediately before it in the coefficient memory, and thereafter, At the moment when the state returns to the single talk state, the amplification coefficient stored in the coefficient memory is read and supplied to the adaptive filter 7. In addition to the double talk state, for example, when the mode is switched from the hands-free talk mode to the handset talk mode according to an instruction from an operation unit (not shown), the echo path estimation operation of the adaptive filter 7 is stopped. Also at this time, the tap coefficient update unit 9 stores the amplification coefficient immediately before the stop of the echo path estimation operation on the coefficient memory.

The voice switch control circuit 13 detects and compares the levels of the reception signal RS and the transmission signal (residual signal ES output from the subtractor 8 of the echo canceller main unit 16), and compares the levels. Control is performed to turn on the attenuator 11 or 12 and turn off the attenuator 11 or 12 on the high-level signal side. With this control, for example, in the receiving state where only the receiving signal is present in the single talk state, the transmission attenuator 12 is inserted, and in the double talking state where both the transmitting state, the receiving signal and the transmitting signal are present, the receiving state is set. The attenuator 11 is inserted. As described above, the voice switch control circuit 13 performs automatic level control for providing attenuation on the signal path side where there is no signal to prevent phenomena such as howling and sound cracking.

Further, the voice switch control circuit 13 has a function of controlling the inserted loss amount according to a line state or the like. This function is provided, for example, by the receiver 121 (FIG. 2).
) Is detected on the basis of the received signal of the attenuator 11 or 1 according to the received electric field level.
2 is realized by increasing or decreasing the loss amount. The voice switch control circuit 13 supplies a control signal (voice switch level VSL) indicating a variable attenuation amount of the attenuator 11 or 12 to be controlled to the center clip control circuit 14 when performing the variable control of the insertion loss amount.

The center clip control circuit 14 includes a voice switch level VSL supplied from the voice switch control circuit 13, a communication state detection result DTS supplied from the double talk detection circuit 10, and a residual signal ES output from the transmission attenuator 12. The center clip circuit 15 is driven on the basis of the control signal, and when the transmission signal (residual signal ES) is lower than the clip level, the signal is removed, and when the transmission signal is higher than the clip level, the signal is passed as it is.

FIG. 3 is a block diagram showing a functional configuration of the center clip control circuit 14. The residual signal level calculator 141, the clip level selector 142, and the comparator 14
3. It is provided with an AND circuit 144.

The residual signal level calculator 141 receives the residual signal ES output from the echo canceling subtractor 8 and passed through the transmission attenuator 12. Here, the residual signal level calculator 141 calculates the average level of the input residual signal ES for a certain period of time (for example, several tens to several hundreds of ms), and inputs the calculation result to the comparator 143.

The voice switch level VSL output from the voice switch control circuit 13 is input to the clip level selector 142. This clip level selector 1
A plurality of clip levels (1 to N) respectively corresponding to the voice switch level VSL used for variable control of the insertion loss amount for the transmission attenuator 12 in the voice switch control circuit 13 described above are stored in advance as coefficient values in 42. Have been. This coefficient value has a characteristic as shown in FIG. 4, for example, and uses the fixed attenuation amount of the transmission attenuator 12 as a parameter,
A value corresponding to the voice switch level VSL from the voice switch control circuit 13 representing the variable attenuation of the transmission attenuator 12 is taken.

When the voice switch level VSL is input from the voice switch control circuit 13, the clip level selecting section 142 selects a clip level corresponding to the input voice switch level VSL from the stored data and compares the clip level. Is input to the device 143.

As described above, the clip level selector 142
Has a function of variably setting the clip level according to the voice switch level VSL input from the voice switch control circuit 13. Here, referring to the characteristic diagram shown in FIG. 4, it can be seen that the clip level becomes lower as the attenuation of the transmission attenuator 12 increases.

Next, the comparator 143 calculates the clip level selected by the clip level selector 142 in accordance with the voice switch level VSL and the residual signal level calculator 141.
Is compared with the residual signal level (average level within a certain period) calculated in step (1), and the result of the comparison is input to the AND circuit 144.

The AND circuit 144 includes the comparator 143
Along with the comparison result, the call state detection result DTS from the double talk detection circuit 10 is input. Here, when the communication state detection result DTS input from the double talk detection circuit 10 is in the single talk state and the comparison result input from the comparator 143 has a residual level smaller than the clip level, the logical product circuit 144 performs the center operation. The center clip control signal CS is output to the clip circuit 15.

The center clip circuit 15 clips the residual signal based on the center clip control signal CS. FIG. 5 shows a general example of the input / output characteristics of the residual signal due to the function of the center clip in the center clip circuit 15. According to this characteristic diagram, in the center clip region to which the center clip control signal CS is applied, even if the residual signal is input to the center clip circuit 15, the output is suppressed to 0, and the center clip control signal CS is applied. It can be seen that the input residual signal is output at the same level outside the center clip region where it is not available.

FIG. 6 shows an example of a temporal variation characteristic of the transmission signal output level accompanying the signal clipping operation in the center clipping circuit 15 based on the above control of the center clipping control circuit 14. In this example, the transmission attenuator 12 is controlled by the variable control of the insertion attenuation amount in the voice switch control circuit 13.
Is set to a value larger than 0. In the center clip control circuit 14, the clip level selecting unit 142 sets the attenuation in accordance with the voice switch level VSL with respect to the set attenuation of the transmission attenuator 12. A clip level CL smaller than the clip level CL0 corresponding to the voice switch level VSL when the amount is 0
Is assumed to be set.

In FIG. 6, at time t0, a single talk (reception) state is established, the transmission attenuator 12 has already been activated, echo cancellation is performed, and the output signal level TS of the transmission signal gradually decreases. Even if the output signal level TS becomes lower than the clip level CL0 (corresponding to the clip level in the conventional example) when the transmission attenuation is 0, the clip level CL already corresponds to the attenuation in the transmission attenuator 12.
, The center clip does not work here, and the output signal level TS changes to the clip level C at time t1.
The center clip operates only when L is reached.

From time t1 to time t2, the center clip is operating. At time t2, the call enters the double talk state, the center clip is stopped, and the transmission attenuator 1
The attenuation of 2 also becomes 0. The double talk state lasts from time t2 to time t3, and at time t3 single talk (receiving)
Enter the state. Here, the transmission attenuator 12 operates again, and echo cancellation is performed. Thereafter, the output signal level TS of the transmission signal gradually decreases due to the continuation of the echo cancellation, and the output signal level TS is divided by the clip level CL0 (corresponding to the clip level in the conventional example) when the transmission attenuation is 0. And a time t at which the clip level CL, which is lower by an amount corresponding to the attenuation in the transmission attenuator 12 at this time, is divided.
At 4, the center clip operates again.

When the function of the center clip in the center clip circuit 15 of the present invention (see FIG. 6) is compared with the function of the center clip in the conventional example (see FIG. 7), the clip level is CL0 in the conventional example. In contrast to the fixed level, in the present invention, the clip level can be varied according to the attenuation of the transmission attenuator 12, and in this example, the attenuation of the transmission attenuator 12 with respect to the clip level CL0. Is set to the clip level CL that is lower by the amount corresponding to. As described above, operating the clip at the clip level CL that is lower by the amount corresponding to the amount of attenuation in the transmission attenuator 12 means that the far end talker has lower voice (received voice) of the other talker. This means that the sound is interrupted after the level is reached, thereby reducing the sense of discomfort that the received voice is suddenly interrupted.

As described above, in the center clip control circuit 14 of the present invention, since the clip level is made variable in accordance with the amount of attenuation of the transmission attenuator 12, the center level is controlled with respect to the residual signal level ES of the echo canceller output. The point where the clip works is constant.

When a call is made in a state where the ambient noise is large, the average level of the residual signal level ES is constantly increased. As a result, the residual signal level ES becomes higher than the clip level, and the center clip control is not performed in spite of the single talk state. No more naturalness.

Further, when the double talk state and the single talk state are repeated within a relatively short time, the average level of the residual signal level ES is hardly reduced, and the residual signal level ES is larger than the clip level. Continues, the center clip control is not performed, and the residual signal level E
S becomes a transmission signal as it is, and generation of a click sound is prevented.

[0050]

As described above, according to the present invention,
The above-mentioned clip level is variably set in accordance with the call state with respect to a signal clipping means for clipping the residual signal after echo cancellation according to the clip level, and the variably set clip level and the residual signal within a certain period are set. The clip is operated in accordance with the result of comparison with the average level, so that the call is suddenly interrupted compared to the case where the clip level is fixed and the clip is operated by comparison with the absolute value of the residual signal. Discomfort can be reduced.

Further, according to the present invention, a transmission attenuator inserted into a transmission signal path to attenuate the residual signal based on a communication state is further provided, and when the clip level is variably set, the transmission attenuator is used. The amount of attenuation may be added to the parameter. As described above, by setting the clip level variably based on the average level of the residual signal for a certain period, the result of the call state determination, and the amount of attenuation of the transmission attenuator, when the vehicle interior noise is relatively high, such as when driving, Can reduce click noise in situations where the call signal is discontinuous or when single talk and double talk are frequently switched, realizing a stable echo canceling function while maintaining an extremely natural call that does not cause discomfort to the other party .

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an echo canceller of a wireless telephone device according to the present invention and peripheral circuits thereof.

FIG. 2 is an overall configuration diagram of a wireless telephone device including a circuit portion shown in FIG. 1;

FIG. 3 is a configuration diagram of a center clip control circuit of the wireless telephone device according to the present invention.

FIG. 4 is a diagram showing a relationship between a clip level used in a center clip control circuit and a voice switch level.

FIG. 5 is a diagram showing input / output characteristics of a residual signal in a center clip circuit.

FIG. 6 is a diagram showing a time variation characteristic of a transmission output level in the wireless telephone device according to the present invention.

FIG. 7 is a diagram showing a time variation characteristic of a transmission output level in a conventional wireless telephone device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 D / A converter 2 Receiving amplifier 3 Speaker 4 Microphone 5 Transmitting amplifier 6 A / D converter 7 Adaptive filter (ADF) 8 Subtractor 9 Tap coefficient update part 10 Double talk detection circuit 11 Receiving attenuator 12 Transmission Attenuator 13 Voice switch control circuit 14 Center clip control circuit 141 Residual signal level calculation unit 142 Clip level selection unit 143 Comparator 144 Logical product circuit 15 Center clip circuit 16 Echo canceller 111 Microphone 112 Audio encoder (SPCOD) 113 Error Correction encoder (CHCOD) 114 Digital modulator (MOD) 115 Multiplier 116 Power amplifier (PA) 117 High frequency switch (SW) 118 Antenna 121 Receiver (RX) 122 Digital demodulator (DEMOD) 123 Error correction decoder ( CHDEC) 124 Audio decoder (SPDEC) 125 Speaker 130 Control circuit (CONT) 131 Frequency synthesizer (SYN) 140 Power supply circuit

Claims (4)

[Claims] 1. A reproducing means for reproducing a received signal, a transmitting means for generating a transmission signal from a transmission input signal, and an acoustic echo signal generated by inputting the received signal to the transmitting means for suppressing the remaining echo signal. A voice communication device having an echo canceling unit that outputs a difference signal, wherein a call state determining unit that determines a call state based on a level of the received signal and a level of a residual signal output from the echo canceling unit; When the average level of the residual signal for a certain period is smaller than the set clip level, the signal clipping means for removing the residual signal is used by the signal clipping means based on the determination result of the call state determination means. And a signal clip control unit for variably setting the clip level. 2. A speech attenuator inserted into a transmission signal path for attenuating the residual signal based on a speech state, wherein the signal clip control unit determines the speech state determined by the speech state determination unit. 2. The voice communication device according to claim 1, wherein the clip level is variably set based on a state and an amount of attenuation of the transmission attenuator. 3. An attenuator for outputting a control signal representing an attenuation of the transmission attenuator when the residual signal is attenuated. Storage means for storing a clip level corresponding to the amount of attenuation, and clip level selection means for reading and outputting a clip level corresponding to the control signal output from the attenuation amount instruction means from the storage means. 3. The voice communication device according to claim 2, wherein: 4. The clip level stored in the storage means is associated with the attenuation of the transmission attenuator such that the greater the attenuation, the smaller the attenuation. Item 4. The voice communication device according to item 3.