JP2012205161A - Voice communication device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice communication device making the transfer of sense of presence at a place where an opposite party is present compatible with the improvement of voice clarity of the opposite party.SOLUTION: A background sound splitter 101 separates a received speech signal output from a voice communication module into voice and background sound. There are provided a voice reproduction speaker 102 and a background sound reproduction speaker 103 for reproducing the voice and the background sound, respectively, which are output from the background sound splitter 101. Consequently, the background sound at the place where the opposite party is present is reproduced, so that the sense of presence can be transferred. Further, by the reproduction of the voice and the background sound on the opposite party from different directions, voice clarity of the opposite party can be improved as compared to the case of reproduction from the same direction.

Description

  The present invention relates to a voice call device that reproduces voice of a communication partner from a speaker when a call is made through a communication line.

  As a conventional voice communication device, there is known a communication device including background sound output means for outputting a predetermined background sound during a call (see, for example, Patent Document 1).

  Further, as a noise suppression processing method, a method is known in which stationary noise is suppressed by subtracting a noise spectrum estimated in a non-speech interval from a speech signal on which noise is superimposed (for example, Non-Patent Document 1). .

Japanese Patent Laid-Open No. 11-191796

S. F. Boll, “Suppression of Acoustic Noise in Spectral Usage Subtraction”, IEEE Transactions Acoustics, Speech and Signal Processing, vol. 27, pp. 113-120, 1979

  However, the conventional voice communication device has a problem that it cannot transmit the background sound superimposed on the actual voice of the other party to the caller only by outputting a specific background sound prepared in advance. . Further, although the clarity of speech is improved by performing the conventional noise suppression processing, there is a problem that the realism of the other party on the other side of the call included in the noise superimposed on the speech is impaired.

  The present invention has been made to solve the conventional problems, and an object of the present invention is to provide a voice call device that achieves both improvement in clarity of a call partner's voice and transmission of realism.

  In order to achieve the above object, the present invention provides a background sound separator that separates a signal sent from a call partner into a call partner's voice and a background sound, and a call partner's voice separated by the background sound separator. And a plurality of speakers installed at different positions for reproducing the background sound.

  According to the present invention, since the other party's voice and the background sound can be reproduced from the speakers installed at different positions, two effects of improving the clarity of the other party's voice and transmitting a sense of reality can be obtained. The voice call device can be provided.

The block diagram of the voice call apparatus in the 1st Embodiment of this invention The block diagram which shows an example of the background sound separator in the voice call apparatus of this invention 1 is a block diagram showing an example of an echo canceller in a first embodiment of the present invention Block diagram of a voice communication device in a second embodiment of the present invention The block diagram which shows an example of the echo canceller in the 2nd Embodiment of this invention The block diagram which shows an example of the background sound suppressor in the 2nd Embodiment of this invention

(Embodiment 1)
A voice communication device according to a first embodiment of the present invention will be described below with reference to the drawings.

  A voice communication device according to a first embodiment of the present invention is shown in FIG. In FIG. 1, a voice call device 100 is configured to input a signal sent from a call partner from a voice communication module (not shown) and output a signal sent to the call partner to the voice communication module. . Hereinafter, the signal sent from the call partner and the signal sent to the call partner are referred to as a reception signal and a transmission signal, respectively.

  The voice communication device 100 is configured to separate a reception signal input from the voice communication module into voice and background sound, and to reproduce the voice and background sound separated by the background sound separator 101, respectively. The sound reproduction speaker 102, the background sound reproduction speaker 103, the microphone 104 for collecting the voice of the caller, and the sound reproduced by the sound reproduction speaker 102 and the background sound reproduction speaker 103 circulates through the microphone 104. It is composed of an echo canceller 105 for removing sound (echo).

  An amplifier for amplifying a signal may be provided between the background sound separator 101 and the sound reproduction speaker 102 and the background sound reproduction speaker 103 and between the microphone 104 and the echo canceller 105.

  Also, a D / A converter that converts a digital signal into an analog signal between the background sound separator 101 and the sound reproduction speaker 102 and the background sound reproduction speaker 103 and between the echo canceller 105 and the sound communication module. However, an A / D converter that converts an analog signal into a digital signal may be provided between the voice communication module and the background sound separator 101 and between the microphone 104 and the echo canceller 105.

  The background sound separator 101 and the echo canceller 105 may be realized by an arithmetic unit such as a digital signal processor (DSP) that can perform digital arithmetic processing. The operation of the speech processing apparatus configured as described above will be described with reference to FIG.

  First, the reception signal output from the voice communication module is input to the background sound separator 101. The background sound separator 101 separates a received signal that is an input signal into sound and background sound.

  Here, the background sound refers to a signal other than the voice, but the signal other than the voice includes an ambient sound in a place where the other party is located.

  For example, when the other party is at the platform of the station, it may be a noise of a person at the platform or a train sound.

  If this background sound is reproduced together with the sound, it is possible to reproduce the presence of the place where the other party is.

The sound and background sound separated by the background sound separator 101 are reproduced from the sound reproduction speaker 102 and the background sound reproduction speaker 103 installed at different positions.

  Sound is played from a speaker installed at a different position from the background sound because the sound and background sound can be heard from different directions, which improves the clarity of the sound compared to sound coming from the same direction. is there.

  On the other hand, the voice of the caller is input to the microphone 104 and sent as a transmission signal from the communication module to the other party through the communication line.

  The microphone 104 receives not only the voice of the caller but also the voice reproduced from the voice reproduction speaker 102 and the background sound reproduced from the background sound reproduction speaker 103.

  Since the sound reproduced from the sound reproduction speaker 102 and the background sound reproduced from the background sound reproduction speaker 103 are the sound and background sound included in the reception signal separated by the background sound separator 101, these Is sent to the other party through the voice communication module, it will be heard as an echo from the other party.

  Therefore, an echo canceller 105 is provided to remove this echo. The echo canceller 105 receives a signal picked up by the microphone 104 and a sound and a background sound that are output signals of the background sound separator 101.

  The echo canceller 105 operates to remove the voice separated by the background sound separator 101 and the signal highly correlated with the background sound from the signal collected by the microphone 104 and output only the voice of the caller.

  The signal from which the echo is removed by the echo canceller 105 is sent as a transmission signal from the voice communication module to the other party through the communication line.

  Here, the operation of the background sound separator 101 will be described with reference to FIG. FIG. 2 is a block diagram illustrating an example of the background sound separator 101.

  In FIG. 2, reference numeral 200 denotes a background sound separator, 201 and 204 denote multipliers that change the gain of the input signal, 202 denotes a background sound level estimator that estimates the level of the background sound included in the input signal, and 203 denotes a multiplier 201. It is a coefficient updater for updating 204 coefficients.

  Next, the operation of the background sound separator 200 configured as described above will be described. The received signals input to the background sound separator 200 have their gains changed by the multipliers 201 and 204, respectively, and are output as a sound output signal and a background sound output signal, respectively.

  The received signal input to the background sound separator 200 is also input to the background sound level estimator 202 to estimate the background sound level.

  When the received signal is a signal in which background sound is superimposed on speech, the background sound level estimator 202 estimates the background sound level by a process such as detection of a minimum value, for example.

  By performing such processing, it is possible to detect a steady background sound level in a time section where there is no sound.

The background sound level estimation method in the background sound level estimator 202 may be a method of estimating by averaging the background sound levels in the sections other than the speech section. Also in this case, a steady background sound level can be detected.

  The background sound level estimated by the background sound level estimator 202 becomes one input of the coefficient updater 203.

  The other input of the coefficient updater 203 is a reception signal input from the voice communication module. The coefficient updater 203 updates the coefficients set in the multipliers 201 and 204.

The coefficient can be calculated as follows, for example. If the amplitude value of the input signal is X and the amplitude value of the background sound estimated by the background sound level estimator 202 is N, the amplitude value Y of the output signal is
Y = X-N
In this way, the amplitude value of the output signal can be obtained by subtracting the amplitude value of the background sound from the amplitude value of the input signal.
Dividing both sides of the above equation by X,
Y / X = (X−N) / X
Therefore,
Y = H · X
It can be expressed as However, H is
H = (X−N) / X
It is.
When the input signal is multiplied by H as a coefficient of the multiplier 201, an audio signal obtained by subtracting the background sound is obtained as the output signal. However, since these expressions are expressed as amplitude values, the phase component of the audio output signal is the same as the phase component of the input signal.

Moreover, if Y = X−N is substituted into Y = H · X,
X−N = H · X
N = X−H · X
= (1-H) · X
It is.

  By multiplying the input signal by (1-H) as a coefficient of the multiplier 204, a background sound can be obtained as an output signal. However, as in the case of the audio signal described above, these equations are expressed as amplitude values, and therefore the phase component of the background sound output signal is the same as the phase component of the input signal.

  In the above description, the entire signal is described as one multiplication. However, the input signal is divided into a plurality of frequency bands, the background sound level is estimated for each frequency band, and the multiplication process is performed. Also good.

  In this case, more detailed control is possible, and the effect of separating the sound and the background sound is improved.

  Next, the operation of the echo canceller 105 will be described with reference to FIG. FIG. 3 is a block diagram illustrating an example of the echo canceller 105.

In FIG. 3, 300 is an echo canceller, 301 is a background sound canceller adaptive filter, 302 is a speech canceller adaptive filter, and 303 is a background sound canceller adaptive filter updater for updating the coefficient of the background sound canceller adaptive filter 301. , 304 is a voice canceller adaptive filter updater for updating the coefficient of the voice canceller adaptive filter 302, and 305 subtracts the output signal of the background sound canceller adaptive filter 301 from the microphone input signal collected by the microphone 104. A subtractor 306 is a subtracter that subtracts the output signal of the speech canceller adaptive filter 302 from the output signal of the subtractor 305.

  An operation of the echo canceller 300 configured as described above will be described. As described above, in the microphone 104, in addition to the voice of the caller, the sound and the background sound separated by the background sound separator 101 are reproduced from the sound reproduction speaker 102 and the background sound reproduction speaker 103, respectively. Is also entered.

  The signal picked up by the microphone 104 is input to the echo canceller 300 as a microphone input signal, and becomes one input of the subtractor 305.

  On the other hand, the background sound output from the background sound separator 101 is input to the background sound canceller adaptive filter 301 as a background sound input signal.

  The background sound canceller adaptive filter 301 is composed of, for example, a FIR (Finite Impulse Response) filter, and the coefficient of the FIR filter is updated by the background sound canceller adaptive filter updater 303, whereby the characteristics of the filter are set to desired amplitude and phase characteristics. It can be.

  The background sound canceller adaptive filter 301 changes the amplitude characteristic and phase characteristic of the input signal and outputs the result.

  The output signal of the adaptive filter 301 for background sound canceller is the other input of the subtractor 305. The subtractor 305 subtracts the output signal of the background sound canceller adaptive filter 301 from the microphone input signal and outputs the result.

The background sound canceller adaptive filter updater 303 uses, for example, an LMS (Least Mean) from the background sound input signal output from the background sound separator 101 and the output signal of the subtractor 305.
The coefficient of the adaptive filter 301 for background sound canceller is updated using a (Square) algorithm.

  The LMS algorithm is an algorithm that sequentially updates the coefficients of the FIR filter so that the sum of squares of errors is minimized. If the coefficient of the background sound canceller adaptive filter 301 is updated as the output signal of the subtractor 305 that subtracts the output signal of the background sound canceller adaptive filter 301 from the microphone input signal, the background sound input from the microphone input signal The coefficient of the background sound canceller adaptive filter 301 can be set so as to remove only the signal highly correlated with the signal.

  That is, the output of the subtractor 305 is such that only the signal highly correlated with the background sound separated from the microphone input signal by the background sound separator 101 is removed, and the voice of the caller and the sound reproduced from the sound reproduction speaker 102 are removed. It becomes the signal which consists of.

  The output of the subtracter 305 becomes one input of the subtracter 306. On the other hand, the sound output from the background sound separator 101 is input to the sound canceller adaptive filter 302 as a sound input signal.

Similar to the background sound canceller adaptive filter 301, the speech canceller adaptive filter 302 is formed of, for example, an FIR filter, and the coefficient of the FIR filter is updated by the voice canceller adaptive filter updater 304, so that the filter characteristics can be set to a desired value. It can be amplitude and phase characteristics. The speech canceller adaptive filter 302 changes the amplitude characteristic and phase characteristic of the input signal and outputs the result. The output signal of the speech canceller adaptive filter 302 is the other input of the subtractor 306. The subtracter 306 subtracts the output signal of the speech canceller adaptive filter 302 from the microphone input signal and outputs the result.

  The speech canceller adaptive filter updater 304 updates the coefficient of the speech canceller adaptive filter 302 from the speech input signal output from the background sound separator 101 and the output signal of the subtractor 306 using, for example, an LMS algorithm.

  Similar to the coefficient update of the adaptive filter updater 303 for the background sound canceller, the voice canceller is used as the output signal of the subtractor 306 that subtracts the error signal in the LMS algorithm from the output signal of the subtractor 305 from the output signal of the voice canceller adaptive filter 302. If the coefficient of the adaptive filter 302 is updated, the coefficient of the speech canceller adaptive filter 302 is set so that only the signal having a high correlation with the audio signal output from the background sound separator 101 is removed from the output signal of the subtractor 305. can do.

  That is, the output of the subtracter 306 is only the voice of the caller by removing only the signal highly correlated with the voice separated by the background sound separator 101 from the output signal of the subtractor 305.

  The output of the subtracter 306 becomes a transmission output signal as an output signal of the echo canceller 300. As a whole, the echo canceller 300 operates so as to remove only the voice of the caller by removing the voice output from the background sound separator 101 and the signal highly correlated with the background sound from the microphone input signal.

  However, there are restrictions on the period in which the coefficients of the background sound canceller adaptive filter 301 and the speech canceller adaptive filter 302 are updated.

  When the coefficient of the background sound canceller adaptive filter 301 is updated, the background sound canceller adaptive filter updater 303 uses only the background sound reproduced from the background sound reproduction speaker 103 as the microphone input signal input to the subtractor 305. When there is an update, it is necessary to perform update processing.

  This is to prevent degradation of the coefficient update accuracy due to the presence of other signals. Therefore, the background sound canceller adaptive filter updater 303 detects the background sound input signal output from the background sound separator 101.

  Furthermore, the speech canceller adaptive filter updater 304 performs signal detection of the speech input signal output from the background sound separator 101 and notifies the background sound canceller adaptive filter updater 303 of the detection result.

  From the two signal detection results, the background sound canceller adaptive filter updater 303 detects a period in which only the background sound exists, and performs update processing in the period in which only the background sound exists.

  Similarly, when the coefficient of the speech canceller adaptive filter 302 is updated, the speech canceller adaptive filter updater 304 uses the output signal of the subtractor 305 input to the subtractor 306 to reproduce the speech reproduced from the speech playback speaker 102. If only exists, it is necessary to perform update processing.

Therefore, the speech canceller adaptive filter updater 304 performs signal detection of the speech input signal output from the background sound separator 101.

  Further, the background sound canceller adaptive filter updater 303 performs signal detection of the background sound input signal output from the background sound separator 101 and signal detection of the output signal of the subtractor 305, and the detection result is applied to the sound canceller adaptive filter. Notify the filter updater 304.

  If no signal is detected from the output of the subtractor 305, the background sound canceller adaptive filter updater 303 removes the background sound by the background sound canceller adaptive filter 301 regardless of the signal detection result of the background sound input signal. The result that no background sound is detected is output.

  The background sound canceller adaptive filter updater 303 also outputs the signal detection result of the background sound input signal output from the background sound separator 101 as usual when a signal is detected from the output of the subtractor 305.

  From the sound and background sound signal detection results output from these background sound separators 101, a period in which only sound is present is detected, and the sound canceller adaptive filter updater 304 performs update processing.

  According to the voice communication device of the first embodiment of the present invention, the background sound separator for separating the sound included in the received signal and the background sound, the sound separated by the background sound separator, By providing an audio reproduction speaker for reproducing background sound from different positions and a background sound reproduction speaker, the voice of the received signal and the background sound can be reproduced from different positions. This makes it possible to improve the clarity of the sound while reproducing the realistic sensation.

In addition, by providing a microphone that collects the voice of the caller, and an echo canceller that removes a signal reproduced from the sound reproduction speaker and the background sound reproduction speaker included in the input signal of the microphone, The echo component contained in the input signal of the microphone can be removed, and the voice of the caller without echo can be sent to the other party.
(Embodiment 2)
Next, FIG. 4 shows a voice communication apparatus according to the second embodiment of the present invention. In FIG. 4, the voice call device 400 is configured to input a reception signal from a voice communication module (not shown) and output a transmission signal to the voice communication module, as in the first embodiment. Yes.

  The voice call device 400 separates the received signal input from the voice communication module into voice and background sound, and reproduces the voice and background sound separated by the background sound separator 401, respectively. The sound reproduction speaker 402, the background sound reproduction speaker 403, the microphone 404 for collecting the voice of the caller, and the sound (echo) around which the sound reproduced by the sound reproduction speaker 402 passes through the microphone 404 are removed. An echo canceller 405 for the background sound, and a background sound suppressor 406 for suppressing the background sound reproduced from the background sound reproducing speaker 403 from the sound input to the microphone 404.

  The operation of the voice communication device configured as described above will be described with reference to FIG. In FIG. 4, the operations of the background sound separator 401, the sound reproduction speaker 402, the background sound reproduction speaker 403, and the microphone 404 are the same as those in the first embodiment.

  In the first embodiment, the echo canceller 105 operates to remove the sound reproduced from the sound reproduction speaker 102 and the background sound reproduced from the background sound reproduction speaker 103. Then, the echo canceller 405 operates to remove the sound reproduced from the sound reproduction speaker 402, and the background sound suppressor 406 operates to suppress the background sound reproduced from the background sound reproduction speaker 403.

  First, the operation of the echo canceller 405 will be described with reference to FIG. In FIG. 5, the echo canceller 500 includes a speech canceller adaptive filter 501, a speech canceller adaptive filter updater 502, and a subtracter 503.

  The operations of the speech canceller adaptive filter 501, speech canceller adaptive filter updater 502, and subtractor 503 are the same as those of the speech canceller adaptive filter 302, speech canceller adaptive filter updater 304, and subtractor 306 in the first embodiment, respectively. The operation is the same.

  Since only the speech canceller adaptive filter 501 is configured as the adaptive filter, the echo canceller 405 removes only the signal having a high correlation with the speech separated by the background sound separator 401 from the signal input to the microphone 404. The operation of outputting the voice of the caller and the background sound reproduced from the background sound reproducing speaker 403 is performed.

  Next, the operation of the background sound suppressor 406 will be described with reference to FIG. In FIG. 6, the background sound suppressor 600 includes a multiplier 601, a background sound level estimator 602, and a coefficient updater 603.

  The operations of the multiplier 601, the background sound level estimator 602, and the coefficient updater 603 are the same as the operations of the multiplier 201, the background sound level estimator 202, and the coefficient updater 203 in the first embodiment, respectively.

  However, the background sound separated by the background sound separator 401 is also input to the background sound level estimator 602. As the input signal of the background sound level estimator 602, not only the signal from which the echo is canceled by the echo canceller 405 from the signal input to the microphone 404 but also the background sound separated by the background sound separator 401 is input. The background sound level can be estimated with higher accuracy.

  According to the voice communication device of the second embodiment of the present invention, the background sound separator for separating the sound included in the received signal and the background sound, and the sound separated by the background sound separator. And background sound playback speakers and background sound playback speakers, so that the voice of the received signal and the background sound can be played from different positions. While reproducing the realistic sensation of the place, the clarity of the voice can be improved.

  A microphone that picks up the voice of the caller; an echo canceller that removes a signal reproduced from the sound reproduction speaker included in the input signal of the microphone; and the background sound reproduction included in the input signal of the microphone. By providing a background sound suppressor that suppresses a signal reproduced from a loudspeaker, an echo component contained in the input signal of the microphone can be removed or suppressed, and the voice of the caller without echo is sent to the call partner be able to. By performing background sound suppression using the background sound suppressor, the amount of calculation can be reduced as compared with the case of using an echo canceller, and the system configuration can be simplified.

  Since the voice call device of the present invention plays back the other party's voice separated from the background sound separator and the background sound from the speakers installed at different positions when the other party's voice is played from the speaker, As a voice communication device that has the effect of transmitting the presence of the other party and the clarity of the other party's voice, and that reproduces the other party's voice from the speaker when calling through the communication line Useful.

100, 400 Voice communication device 101, 401 Background sound separator 102, 402 Sound reproduction speaker 103, 403 Background sound reproduction speaker 104, 404 Microphone 105, 300, 405, 500 Echo canceller 200 Background sound separator 201, 204, 601 Multipliers 202, 602 Background sound level estimator 203, 603 Coefficient updater 301 Background sound canceller adaptive filter 302, 501 Voice canceller adaptive filter 303 Background sound canceller adaptive filter updater 304, 502 Voice canceller adaptive filter update 305, 306, 503 Subtractor 406, 600 Background sound suppressor

Claims (3)

  A background sound separator that separates the voice of the other party into a sound and a background sound, a sound reproduction speaker that reproduces the sound separated by the background sound separator, and a position different from the sound reproduction speaker. A speaker for reproducing a background sound that reproduces the background sound separated by the background sound separator, and the speaker for reproducing the sound provided at a position different from the sound output from the background sound separator and the background sound. And a voice call device that plays back from the background sound playback speaker.   A microphone that picks up the voice of a caller; and an echo canceller that removes echoes input to the microphone; the echo canceller is configured to reproduce the voice separated by the background sound separator and the background sound for the voice reproduction. 2. The voice communication device according to claim 1, wherein a signal reproduced by the speaker and the background sound reproducing speaker is removed.   A microphone that picks up the voice of a caller, an echo canceller that removes echoes input to the microphone, and a background sound suppressor that suppresses the background sound input to the microphone, the echo canceller, The sound separated by the background sound separator is removed from the signal reproduced by the sound reproduction speaker, and the background sound suppressor reproduces the background sound separated by the background sound separator by the background sound reproduction speaker. 2. The voice communication device according to claim 1, wherein the transmitted signal is suppressed.