JPH11355450A - Voice detecting device and speech communication device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To surely detect the generation of voice by a speaker without time delay to a voice signal by delaying the voice signal only for time that corresponds to processing of a level detecting means which detects a mean level of the voice signal and outputting it. SOLUTION: A background noise detection circuit 35 detects and outputs a background noise level included in voice data DA by detecting bottom value from power of the data DA. A comparator circuit 49 compares the background noise level detected by the circuit 35 with means power of the data DA detected by a transmitting level detection circuit 36 and outputs an on-off control signal of a switch circuit 28 according to the comparison result. And, a delay circuit 37 delays the data DA for time that is needed to make power means value in the circuit 36 and outputs it. Thus, a voice signal is delayed only for time that corresponds to processing in the circuit 36 and is outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice detection device and a communication device, and can be applied to, for example, a video conference device that can be connected to a multipoint terminal. The present invention determines the average level of an audio signal based on the background noise level detected from the audio signal, and delays the audio signal to ensure the generation of a voice by a speaker and time for the audio signal. A voice detection device capable of detecting without delay and a communication device such as a video conference device using the voice detection device are proposed.

[0002]

2. Description of the Related Art Conventionally, in a video conference apparatus, a voice switch is used to suppress a voice signal input from a microphone when the speaker is not generating voice and transmit the voice signal to a communication partner, thereby making a voice call possible. It is made to ensure quality.

FIG. 5 is a block diagram showing a configuration of a voice switch in a video conference apparatus together with its peripheral circuits. In the video conference apparatus 1, the audio signal SA obtained from the microphone is amplified by a predetermined amplifier circuit and input to the audio switch 2. In the audio switch 2, the multiplier 3 calculates and outputs the power of the audio signal SA by squaring the audio signal SA.

[0004] The peak detection circuit 4 is composed of an envelope detection circuit comprising a diode 5, a capacitor 6 and a resistor 7, and detects and outputs the peak value of the power value of the audio signal SA from the output signal of the multiplier 3.

[0005] The comparator 8 compares the threshold value set by the reference voltage generation circuit 9 with the output voltage of the peak detection circuit 4, and based on the peak value of the power in the audio signal SA as a reference, It is determined whether or not the speaker has generated a voice toward.

[0006] The video conferencing apparatus 1 selects a selection circuit 11 based on the determination result of the audio switch 2 detected in this manner.
The switching of the contacts is controlled. Here, the selection circuit 11
The audio signal SA is directly input to the selected input terminal, and the audio signal SA whose signal level is suppressed by the attenuator 10 is input to the other selected input terminals.

When the peak value of the power value in the audio signal SA falls below the threshold value, the video conference apparatus 1 determines that the speaker has stopped generating voice and switches the contact of the selection circuit 11. The audio signal SA input from the microphone is suppressed and transmitted.

In some conventional voice switches, the occurrence of a speaker is detected based on the average value of the power instead of the peak value of the power.

[0009]

By the way, in the voice switch 2 having the configuration shown in FIG. 5, since the generation of the voice in the speaker is detected based on the peak value of the power in the voice signal, the voice switch 2 generates the impulse noise. They will also react sensitively. As a result, this voice switch has a problem of malfunction due to noise.

When the background noise level changes variously depending on the use environment, there is also a problem of malfunction. In other words, when background noise considerably larger than the microphone is picked up, the voice switch 2 having the configuration shown in FIG. Will do. Incidentally, in this case, by adjusting the threshold value in the comparator 8,
Although erroneous detection can be reduced to some extent, there is a disadvantage that adjustment of the threshold value is complicated, and erroneous detection cannot always be prevented by the adjustment.

On the other hand, in the case of a system in which the occurrence of a speaker is detected based on the average value of power instead of the peak value of power, malfunction due to impulse noise can be avoided. However, due to a time delay with respect to an audio signal required for detecting an average value, for example, when the present invention is applied to a video conference apparatus, the beginning of the audio may be lost and transmitted. In addition, there is a problem that a malfunction may occur with respect to the background noise that changes depending on the use environment, similarly to the voice switch of FIG.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and a voice detection device capable of reliably detecting the generation of a voice by a speaker without a time delay with respect to a voice signal. It is intended to propose a communication device such as a used video conference device.

[0013]

According to the present invention, there is provided a voice detecting apparatus for comparing a background noise level with an average level to output a voice detection result, and a transmission level detecting apparatus. Delay means for delaying and outputting the audio signal for a time corresponding to the processing in the means.

In a communication device for transmitting a speaker's voice signal input from a microphone to a call target, the level of the voice signal output from the voice detection means is switched and transmitted based on the voice detection result by the voice detection means. Switching means, wherein the sound detection means compares the background noise level with the average level to output a sound detection result, and outputs the sound signal only for a time corresponding to the processing in the transmission level detection means. And delay means for outputting with a delay.

Further, in a communication device for transmitting a voice signal of a speaker input from a microphone to a communication target together with a video signal of the speaker obtained by the imaging means, the voice detection means outputs the voice signal based on the voice detection result by the voice detection means. Switching means for switching the signal level of the output audio signal, wherein the audio detection means compares the background noise level with the average level and outputs an audio detection result; Delay means for delaying and outputting an audio signal for a time corresponding to the processing, and a delay time in the first delay means for directly or indirectly delaying encoded data by audio; The timing of the coded data by audio is made coincident with the coded data by video according to the delay time of the delay means.

According to the comparing means for comparing the background noise level with the average level and outputting the voice detection result in the voice detection device, even if the background noise level changes due to the environment, the voice detection is performed in accordance with the change. The result can be obtained, and error detection is reduced accordingly. In addition, erroneous detection due to impulse noise is reduced. At this time, if the audio signal is delayed and output by the time corresponding to the processing in the transmission level detection means, the audio signal can be delayed so as to correspond to the time delay required for detecting the average level, and the time delay with respect to the audio signal can be reduced. Can also be prevented.

In a communication apparatus for transmitting a voice signal of a speaker input from a microphone to a communication target, if the voice detection means is similarly configured, generation of voice by the speaker is ensured, and there is no time delay with respect to the voice signal. Can be detected. In this way, if the sound signal output from the sound detecting means is switched and transmitted based on the sound detection result by the sound detecting means, the sound of the speaker can be transmitted with high clarity by preventing sound interruption.

Further, in a communication device for transmitting a speaker's voice signal input from a microphone to a communication target together with the speaker's video signal acquired by the imaging means, if the voice detection means is similarly configured, the voice of the speaker can be detected. The occurrence can be detected reliably and without a time delay with respect to the audio signal. Thereby, based on the voice detection result by the voice detection means,
If the sound signal output from the sound detecting means is switched and sent, the sound of the speaker can be sent with high intelligibility by preventing sound interruption. Further, at this time, the audio encoded data is encoded by the delay time of the first delay means for directly or indirectly delaying the encoded data of the audio and the delay time of the delay means of the audio detection means. By correcting the timing of the digitized data,
The function of the lip sync can be obtained by effectively utilizing the delay means in the voice detection means.

[0019]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(1) First Embodiment FIG. 2 is a block diagram showing an audio system of a video conference apparatus according to a first embodiment of the present invention. In this video conference device 21, the demultiplexer (DMX)
The audio data is separated and output from the input data D1 transmitted from the communication target via an SDN (Integrated Services Digital Network) line or the like, and the decoder 23 decodes the audio data output from the demultiplexer 22. After doing
Data is decompressed and output.

The video conference device 21 converts the audio data output from the decoder 23 into an echo canceller (EC) 2.
The signal is input to a digital-to-analog conversion circuit (not shown) via an analog-to-digital converter 4, where it is converted to an analog audio signal and output from a speaker 25. Thus, the video conference device 21 can listen to the voice of the speaker to be called.

On the other hand, the video conference device 21 obtains the voice of the speaker using the microphone 26 and generates voice data MA by analog-to-digital conversion processing. The echo canceller 24 removes, from the audio data MA, an audio signal component of the communication target, which is obtained by filtering the voice D1 of the communication target output from the speaker 25 and is crowded from the speaker 25 to the microphone 26, and outputs the signal. I do.

The voice detection circuit 27 detects the voice of the speaker from the voice data DA output from the echo canceller 24 and outputs an on / off control signal SC. The switch circuit 28 receives the audio data DA1 via the audio detection circuit 27, and outputs the audio data DA1 under the control of the on / off control signal SC. The video conference device 21
Thus, only when the speaker utters, the audio data DA1
To the call target.

The delay circuit 29 delays the audio data DA1 output from the switch circuit 28 for a predetermined time, so that a change in the video of the
The timing of the audio data DA1 is corrected and output so that the audio by the audio data DA1 corresponds. The delay circuit 29 executes a so-called lip sync process.

The encoder (ENC) 30 includes a delay circuit 29
After compressing the audio data DA1 output from the
Encode and output. The multiplexer (MUX) 31
The audio data DA1 output from the encoder 30 is multiplexed with image data of a speaker that has been similarly encoded and transmitted to the transmission target. The video conference device 21
Has been set up so that a video conference can be performed with a communication target.

FIG. 1 is a block diagram showing the voice detection circuit 27 together with peripheral components. This voice detection circuit 27
Is the audio data D output from the echo canceller 24
A is input to the background noise detection circuit 35, the transmission level detection circuit 36, and the delay circuit 37.

Here, the background noise detection circuit 35 detects and outputs the background noise level included in the audio data DA by detecting the bottom value from the power of the audio data DA.
That is, in the background noise detection circuit 35, the multiplier 41
The power of the audio data DA is calculated and output by squaring the audio data DA. The diode 42 includes a resistor 43
The anode is connected to the output terminal of the multiplier 41 and the cathode is grounded by a capacitor 44. As a result, the diode 42, the resistor 43, and the capacitor 44
A bottom hold circuit is configured to detect and output a background noise level having a bottom value from the power of the audio data DA.

The transmission level detection circuit 36 outputs the voice data D
The average power of A is detected and output. That is, in the transmission level detection circuit 36, the multiplier 46 outputs the audio data DA
Is squared to calculate and output the power of the audio data DA. The resistor 47 constitutes a low-pass filter together with the capacitor 48, and averages and outputs the power of the audio data DA detected by the multiplier 46.

The comparison circuit 49 is provided for the background noise detection circuit 3
5 and the transmission level detection circuit 3
6. Compare with the average power of the audio data DA detected in 6,
Based on the result of this comparison, an on / off control signal for the switch circuit 28 is output. That is, the comparison circuit 49 switches the switch circuit 28 to the ON state when the average power of the audio data DA rises above a first addition signal level obtained by adding the first reference level to the background noise level. Accordingly, the comparison circuit 49 determines the average power of the audio data DA based on the background noise level, and starts outputting the audio data DA1.

On the other hand, the comparison circuit 49 sets the audio data DA equal to or lower than a second addition signal level obtained by adding a second reference level smaller than the first reference level to the background noise level.
When the average power falls, the switch circuit 28 is turned off. Thus, the comparison circuit 49 determines the average power of the audio data DA based on the background noise level, and stops outputting the audio data DA1. The comparison circuit 49
Outputs an on / off control signal SC based on the hysteresis characteristic to switch the operation of the switch circuit 28, thereby preventing frequent switching of the switch circuit 28.

The delay circuit 37 includes a transmission level detection circuit 36
The audio data D for the time required for averaging the power in
A is delayed and output.

In the above configuration, the audio data D1 transmitted from the communication target (FIG. 2) is separated from other data by the demultiplexer 22, decoded by the decoder 23, and converted into an analog signal. The video signal is converted and output from the speaker 25, so that the speaker operating the video conference device 21 can view the voice of the call target.

On the other hand, the voice of this speaker is
6, the voice detection circuit 27 and the switch circuit 2
8. The signal is input to the encoder 30 via the delay circuit 29, where the data is encoded, multiplexed with other data by the subsequent multiplexer 31, and transmitted to the communication target. As a result, the voice of the speaker using the video conference device 21 can be viewed even in the call target.

When transmitting and receiving voices in this way, the voice of the call target output from the speaker 25 is partially acquired by the microphone 26. In this manner, the voice of the call target acquired by the microphone 26 is removed from the voice data MA by subtracting the original voice data D1 from the voice data MA by the filtering process in the subsequent echo canceller 24. Thus, the echo of the speech of the call target once output from the speaker 25 is transmitted to the call target via the microphone 26, thereby preventing the echo.

In the video conference apparatus 21, the microphone 26 acquires background noise, which is background noise, in addition to the voice of the call target output from the speaker 25 as described above. This background noise generally continues at a constant intensity in the background of the speaker, and varies in intensity depending on the environment, whereas the voice of the speaker rapidly changes according to the utterance of the speaker. There are features.

The voice of the speaker acquired together with the background noise having such a characteristic (FIG. 1) is detected by the background noise detection circuit 35 of the voice detection circuit 27 after the power of the voice data DA is detected. Is detected, the level of the background noise is detected based on the bottom value.
In the transmission level detection circuit 36, the audio data DA
By detecting the average value of the power of the speaker, a total voice level of the background noise and the voice of the speaker is detected.

Thus, the comparison circuit 49 can reliably detect the utterance of the speaker by judging the total sound level based on the noise level, even if the background noise level is variously different. Becomes possible. In addition, the background noise is detected by detecting the bottom value of the power, and the total voice level is detected by the average value of the power, so that erroneous detection due to impulse noise can be effectively avoided.

The voice data DA is output to the encoder 30 via the switch circuit 28 when the speaker's utterance is detected in this way. When the speaker's utterance stops, the voice data DA is output to the encoder 30. Output is stopped. As a result, in the video conference apparatus 21, when the speaker does not emit any sound, the transmission of the audio data DA is stopped, and even when the communication partner views and synthesizes the audio from the multi-point terminal, the background noise is reduced. A decrease in clarity due to the increase is prevented.

The audio data DA transmitted in this manner is equal to the time required for detecting the average value of the electric power,
7, the signal is input to the switch circuit 28 at a timing corresponding to the on / off control of the switch circuit 28 by the comparison circuit 49, thereby preventing the sound interruption at the beginning of the utterance. Further, in the comparison circuit 49, the on / off control of the switch circuit 28 based on the hysteresis characteristic allows the off control of the switch circuit 28 to be performed once the transmission of the audio data DA is started, based on a gradual basis. Cuts and cuts at the end of the sound are prevented.

The audio data DA is supplied to the delay circuit 37.
After being delayed by the delay circuit 29, the delay circuit 29 following the switch circuit 28 receives a delay for lip-sync, so that the delay time required for the delay circuit 29 can be shortened by the provision of the delay circuit 37. .

According to the above arrangement, the average value of the power of the audio signal is determined on the basis of the background noise level detected from the bottom value of the power of the audio signal. Even if the background noise level changes depending on the environment, it is possible to reliably detect the generation of the voice by the speaker in response to the change. In addition, by delaying the audio signal by the delay circuit, it is possible to cope with a time delay required for detecting the average value, and thereby it is possible to detect the generation of the voice by the speaker without a time delay with respect to the audio signal. .

Further, the delay time in the delay circuit 29 can be shortened by the delay of the audio signal by the delay circuit 37, and the ripsing process can be executed.

Further, at this time, when the utterance of the speaker cannot be detected, the transmission of the voice data DA is stopped,
Even when a communication partner views and synthesizes voices from terminals at multiple points, a decrease in intelligibility due to an increase in background noise is prevented.

(2) Second Embodiment FIG. 3 is a block diagram showing a video conference system according to a second embodiment of the present invention. In the video conference system 50, a video conference device 21A having the same configuration as the video conference device 21 described in the first embodiment is used.
To 21N connected to the multipoint connection device 51.

Here, the multipoint connection device 51 is connected to each of the video conference devices 21A to 21N via a public line or the like, and the audio transmitted from each of the video conference devices 21A to 21N by the demultiplexers 52A to 52N. After separating the data, the data is decoded by the following decoders 53A to 53N. The multipoint connection device 51 is connected to each of these video conference devices 21A to 21A.
In the adders 54A to 54N corresponding to 21N, video conference devices (21B to 21N) other than the corresponding video conference devices 21A to 21N, (21A, 21C to 21N),
... (21A to 21M) are added together. Thereby, the multipoint connection device 51 synthesizes a plurality of voices transmitted from the other video conference devices 21A to 21N with respect to each of the video conference devices 21A to 21N.

The multipoint connection device 51 encodes the audio data thus synthesized for each of the video conference devices 21A to 21N by the corresponding encoders 55A to 55N, and then encodes the data by the multiplexers 56A to 56N. The data is multiplexed with other data by 56N and transmitted to each of the video conference devices 21A to 21N.

According to the configuration shown in FIG. 3, a video conference device is connected to a multipoint connection device, and an audio signal transmitted from another video conference device is synthesized and transmitted to one video conference device. In such a case, the transmission of the audio data is stopped when each of the video conference devices is configured by the video conference device according to the first embodiment and the speaker is not speaking.
As described above, even when voices from other video conference devices are synthesized, an increase in background noise can be prevented, and thereby a decrease in intelligibility can be prevented.

(3) Third Embodiment FIG. 4 is a block diagram showing a video conference system according to a third embodiment of the present invention. In this video conference system 60, video conference devices having the same configuration..., 6
1N-1, 61N, 61N + 1,... Are connected in cascade. In the configuration shown in FIG. 4, the same components as those described above with reference to FIG. 2 are denoted by the corresponding reference numerals, and redundant description will be omitted.

Here, these teleconference devices..., 61N
.., 61N, 61N + 1,... Each have an input port and an output port for uplink, an input port and an output port for downlink, and these input and output ports for uplink and downlink are sequentially connected and cascaded. Connected. .., 61N−1, 61N, 61N + 1,... Have the same configuration, so only the video conference device N will be described below, and redundant description will be omitted.

The video conference apparatus N separates audio data by processing the data streams input from the upstream and downstream input ports by the demultiplexers 62A and 62B, respectively. Decode by 63B. Further, the video conference device N includes adders 64A and 6
In 4B, the audio data output from the delay circuit 29 is added to the audio data output from each of the decoders 63A and 63B, and then encoded by the encoders 65A and 65B, respectively. Further, the video conference device N includes a multiplexer 66A.
And 66B multiplex the output data of the encoders 65A and 65B with other data, and output the multiplexed data from the output port for uplink and the output port for downlink.

Thus, the video conference device..., 61N-
1, 61N, 61N + 1,... Can acquire audio data transmitted from another video conference device via input ports for uplink and downlink. In the video conference device at the end of the cascade connection,
The audio data transmitted from another video conference device can be obtained only with the input port for the up or the down.

In the video conference apparatus N, the audio data output from the decoders 63A and 63B are added by the adder 67 and output to the echo canceller 24. By these means, these video conference devices..., 61N-1, 61N, 61N
+1... Can be viewed via the speaker 25 through audio data transmitted from another video conference apparatus which is input through the input ports for up and down.

According to the configuration shown in FIG. 4, a plurality of video conference devices are connected by cascade connection, and an audio signal transmitted from another video conference device to one video conference device is synthesized and viewed. In this case, the same effect as in the first embodiment can be obtained.

(4) Other Embodiments In the above embodiment, the case where the background noise level and the average level of the audio signal are detected based on the bottom value and the average value of the power has been described. For example, when detecting the background noise level and the average level of the audio signal based on the bottom value and the average value of the envelope,
Various detection methods can be widely applied.

Further, in the above-described embodiment, the case where the output of the voice data is stopped when the speaker does not utter the voice has been described. However, the present invention is not limited to this.
If practically sufficient, the audio data may be attenuated and output. In this case, in the second and third embodiments, an increase in the background noise can be prevented by varying the attenuation rate according to the number of video conference devices to be connected.

In the above embodiment, the case where the audio data output from the switch circuit is lip-synced after being delayed by the delay circuit has been described. However, the present invention is not limited to this. The present invention can be widely applied to a case where delay circuits are inserted at various places such as a stage after an encoder.

Further, in the above-described embodiment, the case where the present invention is applied to the video conference apparatus has been described. However, the present invention is not limited to this, and the present invention is not limited to this. Can be widely applied to various devices that operate by detecting the

[0058]

As described above, according to the present invention, the average level of a voice signal is determined based on the background noise level detected from the voice signal, and the voice signal is delayed by delaying the voice signal. Can be detected reliably and without a time delay with respect to the audio signal.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a voice detection circuit according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a video conference device using the audio detection circuit of FIG.

FIG. 3 is a block diagram showing a video conference system according to a second embodiment of the present invention.

FIG. 4 is a block diagram showing a video conference system according to a third embodiment of the present invention.

FIG. 5 is a block diagram showing a conventional video conference device.

[Explanation of symbols]

1, 21, 21A to 21N, 61N-1 to 61N + 1 ...
... Teleconference device, 2... Voice switch, 27... Voice detection circuit, 28... Switch circuit, 29, 37... Delay circuit, 35. , 60 ... Videoconferencing system, 51 ...
Multipoint connection device

Claims (14)

[Claims] 1. A background noise detection means for detecting a background noise level of an audio signal, a transmission level detection means for detecting an average level of the audio signal, and a sound by comparing the background noise level with the average level. A voice detection device comprising: a comparison unit that outputs a detection result; and a delay unit that delays and outputs the voice signal by a time corresponding to a process performed by the transmission level detection unit. 2. The apparatus according to claim 1, wherein the background noise detection unit includes: a power detection unit that detects power of the audio signal; and a bottom value detection unit that detects a bottom value of the power and outputs the background noise level. The voice detection device according to claim 1, wherein 3. The transmission level detection means includes: power detection means for detecting power of the audio signal; and average value detection means for detecting an average value of the power and outputting the average level. The voice detection device according to claim 1, wherein 4. The sound detection result rises or falls when the average level rises from a first addition signal level obtained by adding a first threshold level to the background noise level. When the average level falls below a second addition signal level obtained by adding a second signal level smaller than the first threshold level to the background noise level, the voice detection result falls or rises. The voice detection device according to claim 1. 5. A communication device for transmitting a voice signal of a speaker input from a microphone to a communication target, wherein the voice signal output from the voice detection means is switched and transmitted based on a voice detection result by voice detection means. The voice detection means includes: a background noise detection means for detecting a background noise level of the voice signal; a transmission level detection means for detecting an average level of the voice signal; and the background noise level. Comparing means for comparing the average level and outputting a voice detection result; and delaying means for delaying and outputting the voice signal for a time corresponding to processing in the transmission level detecting means. Intercom equipment. 6. The communication apparatus according to claim 5, wherein the switching unit stops outputting the voice signal when the speaker does not utter a voice according to the voice detection result. 7. The background noise detection unit includes: a power detection unit that detects power of the audio signal; and a bottom value detection unit that detects a bottom value of the power and outputs the background noise level. The communication device according to claim 5, characterized in that: 8. The transmission level detection means includes: power detection means for detecting power of the audio signal; and average value detection means for detecting an average value of the power and outputting the average level. The communication device according to claim 5, characterized in that: 9. The sound detection result rises or falls when the average level rises from a first addition signal level obtained by adding a first threshold level to the background noise level, When the average level falls below a second addition signal level obtained by adding a second signal level smaller than the first threshold level to the background noise level, the voice detection result falls or rises. The communication device according to claim 5. 10. A communication device for transmitting a voice signal of a speaker input from a microphone to a communication target together with a video signal of the speaker obtained by an imaging means, wherein the voice is detected based on a voice detection result by a voice detection means. A switching unit for switching the audio signal output from the detection unit; an audio encoding unit for encoding the audio signal input from the switching unit and outputting encoded data by audio; and encoding the video signal. A video encoding unit that outputs encoded video data, and a first delay unit that directly or indirectly delays encoded data by audio output from the audio encoding unit, Voice detection means for detecting a background noise level of a voice signal input from the microphone, and transmission for detecting an average level of the voice signal; Bell detecting means, comparing means for comparing the background noise level with the average level and outputting a sound detection result, and outputting the sound signal delayed by a time corresponding to processing in the transmission level detecting means. And a delay time in the first delay means, and a delay time in the delay means of the audio detection means, the timing of the encoded data by the audio with respect to the encoded data by the video. A communication device characterized by correcting. 11. The communication device according to claim 10, wherein the switching unit stops transmitting the voice signal when the speaker does not utter a voice according to the voice detection result. 12. The apparatus according to claim 12, wherein the background noise detection means includes: power detection means for detecting power of the audio signal; and bottom value detection means for detecting a bottom value of the power and outputting the background noise level. The communication device according to claim 10, wherein: 13. The transmission level detection means includes: power detection means for detecting power of the audio signal; and average value detection means for detecting an average value of the power and outputting the average level. Claim 10
A telephone device according to claim 1. 14. The sound detection result rises or falls when the average level rises from a first addition signal level obtained by adding a first threshold level to the background noise level, When the average level falls below a second addition signal level obtained by adding a second signal level smaller than the first threshold level to the background noise level, the voice detection result falls or rises. The communication device according to claim 10.