JPH0514475A - Loudspeaker telephone set - Google Patents

Abstract

PURPOSE:To easily realize the loudspeaking telephone set for mobile body mount with excellent quality in hand-free talking with simple circuit configuration. CONSTITUTION:A noise canceller 60 consists of a noise signal generating means 61, an ADF 64 and an adder 65. Then the generating means 61 generates a pseudo noise signal S61 in response to the drive speed of a mobile body and gives it to the ADF 64. The ADF 64 uses a signal S61 as a reference signal on the non-input of a near end talker voice signal, the transfer characteristic from a noise source of the mobile body to a microphone 51 is estimated to output a noise cancelling signal S64 to the adder 65. The adder 65 subtracts the signal S64 from the mobile body drive noise signal from the microphone 51 to delete the actual drive noise MN mixed from the microphone 51. The signal after the delete of drive noise is subject to cancellation of an acoustic echo EC by an echo canceler 70.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a loudspeaker telephone equipped with a hands-free call function, which is mounted on a moving body such as an automobile and allows a call using a speaker and a microphone without having a handset. The present invention relates to a loudspeaker telephone having a noise canceller that suppresses a background noise component caused by traveling of a moving body that is superposed on.

[0002]

2. Description of the Related Art Conventionally, as a public address telephone of this type, there is one described in, for example, JP-A-60-140964.

In a mobile phone mounted as a loudspeaker, for example, a car phone, a driver may make a call while traveling, and in terms of safety and operability, a hands-free call is made using a microphone and a speaker. Calling functions are becoming essential. An example of this type of public address telephone is shown in FIG.

FIG. 2 is a block diagram showing an example of the structure of a loudspeaker telephone installed in an automobile. This loud phone is
A microphone 1 for inputting a voice signal of a driver (near-end speaker) A is provided, and an analog / digital converter (hereinafter referred to as A / D converter) 2 is provided on the output side thereof to detect a voice signal. A wireless device 31 is connected via a voice detector 3 that outputs a control signal S3, a noise canceller 10, an echo canceller 20, and a digital / analog converter (hereinafter referred to as a D / A converter) 30. The radio 31 has A
A speaker 41 that outputs a far-end talker voice signal is connected via the / D converter 32, the echo canceller 20, and the D / A converter 40.

The noise canceller 10 is a circuit for eliminating engine noise, running noise, etc., and has a microphone 11 installed in an engine room or the like, which is a main source of noise. On the output side of the microphone 11, the control signal S3 for the A / D converter 12 and the voice detector 3 is provided.
An adaptive digital filter (hereinafter referred to as ADF) 13 whose adaptive operation is on / off controlled by an adder 1 and an adder 1
4 is connected. The echo canceller 20 eliminates the acoustic echo EC from the speaker 41 to the microphone 1, and is composed of an ADF 21 whose adaptive operation is turned on / off by a control signal S3 of the voice detector 3 and an adder 22. There is.

Next, the operation will be described. During a hands-free call, the driver (near-end speaker) A receives the voice of the other party (far-end speaker) transmitted through the wireless device 31 as a received signal, and the A / D converter 32, the echo canceller 20, D /
It can be heard through the A converter 40 and the speaker 41. Further, the voice signal of the near-end speaker A is sent to the wireless device 31 via the microphone 1, the A / D converter 2, the noise canceller 10, the echo canceller 20, and the D / A converter 30, and the far-end talker It is transmitted to the person.

In the case of this hands-free call, the speaker 4
There is a problem that the voice from No. 1 reverberates in the vehicle and is input to the microphone 1, becomes an acoustic echo EC, and returns to the far-end speaker again to make a call. Particularly, when the level of the acoustic echo EC is high, howling occurs and the communication quality deteriorates. Therefore, the echo canceller 20 eliminates the acoustic echo EC to suppress the deterioration of the call quality.
In the echo canceller 20, the audio signal from the far-end speaker output from the A / D converter 32 is used as a reference signal.
It is input to the DF 21, the transfer characteristic of the echo path is adaptively estimated by the ADF 21, and the echo cancellation signal is output to the adder 22. In the adder 22, by subtracting the echo cancellation signal from the signal from the microphone 1,
The acoustic echo EC is erased.

Further, when a call is made while the vehicle is traveling, particularly when traveling at a high speed, the noise of the engine of the automobile, the ground noise of the tire, the wind noise, etc. are large, and the echo canceller 20 is erased due to these noises during the hands-free conversation. Poor performance causes acoustic echo, and howling occurs, resulting in very poor call quality. On the other hand, even when these acoustic echoes are generated or howling is not significant, when the voice of the near-end speaker A is input from the microphone 1 and transmitted to the far-end speaker, engine noise or running noise is generated. Since the same is also input, if these levels are large, the signal-to-noise ratio (S / N ratio) becomes worse on the far-end talker side, and it becomes difficult or difficult to talk.

Therefore, a conventional loudspeaker telephone is provided with a noise canceller 10, and a noise canceller microphone 11 is provided in an engine room or the like which is a main source of noise.
Is installed, the engine sound is picked up as a reference signal, the A / D converter 12 converts it into a digital signal and AD
Input to F13. The ADF 13 outputs a noise canceling signal to the adder 14 by estimating the transfer characteristic from the noise source to the microphone 1. The adder 14 eliminates noise by subtracting the noise cancellation signal from the noise signal input from the microphone 1.

As described above, the noise canceller 10 removes the influence of noise to prevent the canceling ability of the echo canceller 20 from being deteriorated, and at the same time, does not transmit unnecessary noise to the far-end talker, and the call quality of the hands-free call. Can be kept good.

[0011]

However, in the conventional loudspeaker telephone, the microphone 11 and the A / D converter 12 for adding noise are newly required, and the circuit configuration becomes complicated. Moreover, there are problems such as the location of these installations and the number of installation steps,
It was an obstacle to practical use.

The present invention provides a loudspeaker telephone for a mobile body, which solves the problems of the above-mentioned prior art in that the circuit structure is complicated, and that the installation location and the number of installation steps are required. .

[0013]

In order to solve the above-mentioned problems, a first aspect of the present invention provides a microphone for inputting a near-end talker voice signal, a speaker for outputting a far-end talker voice signal, and a microphone from the microphone. A loudspeaker including a noise canceller for canceling a moving object traveling noise signal, and an echo canceller having an ADF for canceling an acoustic echo from the microphone output from the noise canceller using the far-end talker voice signal as a reference signal. In the telephone, the noise canceller is configured as follows.

That is, the noise canceller according to the first aspect of the present invention comprises a noise signal sequence generating means for generating a pseudo noise signal according to the traveling speed of the moving body, and the pseudo noise when the near-end talker voice signal is not input. ADF for a noise canceller that estimates a transfer characteristic from the noise source of the moving body to the microphone by using a signal as a reference signal and outputs a noise canceling signal, and an adder that subtracts the noise canceling signal from the moving body running noise signal And are equipped with.

According to a second aspect of the present invention, the noise signal sequence generating means of the first aspect of the invention includes a memory for storing data of a pseudo noise signal corresponding to the traveling speed of the moving body, and a traveling speed of the moving body. Switching means for switching the output of the memory and giving it to the noise canceller ADF.

[0016]

According to the first aspect of the invention, since the loudspeaker telephone is constructed as described above, the noise signal sequence generating means generates a pseudo noise signal very similar to the characteristic of the actual running noise in the moving body to generate the ADF. Give to. When the near-end talker voice signal is not input, the ADF estimates the transfer characteristic from the noise source to the microphone using the pseudo noise signal as a reference signal and outputs the noise cancellation signal to the adder. The adder subtracts the noise canceling signal from the moving object running noise signal to appropriately suppress the background noise component caused by the running of the moving object acoustically superimposed on the acoustic signal. Due to the signal in which the background noise component is suppressed, the acoustic echo from the microphone is canceled by the echo canceller in the subsequent stage. As a result, the call quality can be improved easily and easily.

According to the second aspect of the present invention, the noise of the speed of the moving body and the road condition is measured, the sampled value is stored in the memory in advance, and the switching means is used to change the speed according to the traveling speed of the moving body. The contents of the memory are switched and given to the ADF. Then, the ADF and the adder use the pseudo noise signal provided by the switching means as a reference signal, and cancel the acoustic echo by adaptive control. As a result, the quality of hands-free call can be improved easily and easily. Therefore, the above problem can be solved.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an embodiment of the present invention and is a block diagram showing the construction of a loudspeaker telephone installed in an automobile, for example.

This loudspeaker telephone has a microphone 51 for inputting the voice signal of the near-end speaker A, and its output side is
An A / D converter 52 for converting an analog voice signal into a digital signal, a voice detector 53 for detecting an input of a near-end talker voice signal and outputting a control signal S53, a noise canceller 6
0, the echo canceller 70, and a wireless device 81 are connected via a D / A converter 80 that converts a digital signal into an analog signal. This radio 81 has an A / D for converting an analog signal from a far-end speaker into a digital signal.
A speaker 91 that outputs a far-end talker voice signal is connected via a converter 82, an echo canceller 70, and a D / A converter 90 that converts a digital signal into an analog signal.

The noise canceller 60 includes a microphone 5
The running noise MN from No. 1 is deleted by using the ADF, and the pseudo noise signal S6 corresponding to the running speed of the automobile is used.
It is composed of a noise signal sequence generating means 61 for generating 1, a transversal type ADF 64, and an adder 65, for example. The noise signal sequence generation means 61 includes a memory 62 for storing data of a pseudo noise signal according to the traveling speed of the vehicle, and a switching means 63 for switching the output of the memory 62 according to the traveling speed of the vehicle and giving it to the ADF 64.
It is configured. The memory 62 is composed of, for example, a plurality of read-only memories (ROM) 62a to 62c.
The switching means 63 detects a vehicle speed sensitive sensor 63a for detecting the speed of the vehicle, and outputs R from the vehicle speed sensitive sensor 63a.
It is composed of a switch 63b that switches the outputs of the OMs 62a to 62c.

ADF connected to the output side of the switch 63b
64 is a control signal S53 output from the voice detector 53.
The adaptive operation is controlled to be turned on and off by the pseudo-noise signal S61 from the switch 63b when the near-end speaker voice signal is not input, and the transfer characteristic from the noise source of the automobile to the microphone 51 is set. And outputs the noise canceling signal S64 to the adder 65. The adder 65 has a function of subtracting the noise cancellation signal S64 from the signal from the A / D converter 52.

The echo canceller 70 has a function of eliminating the acoustic echo EC from the microphone 51, and is composed of, for example, a transversal ADF 71 and an adder 72. The ADF 71 uses the far-end speaker voice signal from the A / D converter 82 as a reference signal, estimates the acoustic transfer characteristic from the speaker 91 to the microphone 51, and outputs the echo cancellation signal S71 to the adder 72. Have The adder 72 is a circuit that subtracts the echo cancellation signal S71 from the output signal of the A / D converter 52 and cancels the acoustic echo EC.

Next, the operation will be described. For example, since the running sound of an automobile is considered to be almost constant depending on the vehicle type, it is possible to know the noise generation sequence of the running sound in advance. These are very similar to the characteristics of actual running noise. Therefore, noise is measured at typical speeds and road conditions. That is, the running noise is measured at another place closer to the noise source. Then, the sampled values are stored in the ROMs 62a to 62c. When the near-end speaker A makes a call, the voice signal of the near-end speaker A is input from the microphone 51, the output of the microphone 52 is converted into a digital signal by the A / D converter 52, and the noise canceller 60 and D via the echo canceller 70
The signal is again converted into an analog signal by the A / A converter 80 and sent to the far-end talker side by the wireless device 81. On the other hand, the wireless device 81
When a call is received from the far-end talker side via the, the voice signal from the far-end talker side is converted into a digital signal by the A / D converter 82, and is passed through the echo canceller 70.
The analog signal is again converted by the D / A converter 90, and is output from the speaker 91 to be heard by the near-end speaker A.

The output signal of the A / D converter 52 can be regarded as an acoustic echo signal when the voice of the near-end speaker A is not present. Therefore, the voice detector 53 detects that the voice signal is gone, and the control signal S53 output from the voice detector 53 causes the ADF 71 in the echo canceller 70 to perform an adaptive operation. That is, ADF71
Uses the far-end speaker voice signal from the A / D converter 82 as a reference signal, estimates the acoustic transfer characteristic from the speaker 91 to the microphone 51, and minimizes the residual signal S72 output from the adder 72. Such an echo cancellation signal S71 is output to the adder 72. The number of taps of the ADF 71 is determined according to the impulse response length between the speaker 91 in the vehicle and the microphone 51. The adaptive control of the ADF 71 is performed by a learning identification method, an LMS (least mean square) method, or the like so that the root mean square value of the residual signal S72 is minimized. In the adder 72, the A / D converter 5
The echo canceling signal S71 is subtracted from the output of 2 to cancel the acoustic echo EC from the microphone 51.

When the voice signal of the near-end speaker A is not present, the voice signal is detected by the voice detector 53, and the control signal S5 is output.
3, the adaptive operation of the ADF 64 in the noise canceller 60 is performed. In this noise canceller 60, the traveling speed of the vehicle is detected by the vehicle speed sensitive sensor 63a, and the switch 63b is switched by the detection signal, and the ROM 6
The noise sample data stored in 2a to 62c is
The pseudo noise signal S61 is provided to the ADF 64.

Like the ADF 71 in the echo canceller 70, the ADF 64 uses the pseudo noise signal S61 as a reference signal, estimates the transfer characteristic from the noise source to the microphone 51, and outputs the noise canceling signal S64 to the adder 65. . The adder 65 subtracts the noise canceling signal S64 from the traveling noise signal input from the microphone 51 to eliminate the actual traveling noise MN mixed from the microphone 51. Then, if the signal after the cancellation is sent to the adder 72 of the echo canceller 70 and the acoustic echo EC is canceled by the echo canceller 70, an accurate echo cancellation operation becomes possible.

The present embodiment has the following advantages.

(A) Running noises during running and road conditions are measured, sample values thereof are stored in advance in the ROMs 62a to 62c, and these are input as reference signals to the ADF 64 via the switch 63b. And adder 6
Since the noise is canceled by 5, the actual running noise MN mixed from the microphone 51 can be considerably removed.

It is sufficient that the noise sample data stored in the ROMs 62a to 62c be taken for several seconds when the speed is changed in an actual traveling vehicle. It is desirable to detect the sample at the center of the windshield surface as much as possible.

(B) Since the operation of the echo canceller 70 is performed after the traveling noise MN is eliminated by the noise canceller 60, the operation of the echo canceller 70 is stabilized, and there is no acoustic echo feeling or howling, and the traveling noise MN is also generated. Since it is not transmitted to the far-end talker side, a natural and high quality hands-free call becomes possible.

(C) Apart from the original microphone 51,
It is not necessary to provide a microphone and an A / D converter for collecting the running noise MN of the automobile. Therefore, the noise canceller 60 can be configured at low cost by omitting the number of mounting steps and simplifying the circuit configuration without specially manufacturing the vehicle body.

The present invention is not limited to the above embodiment,
Various modifications are possible. The following are examples of such modifications.

(I) Noise signal sequence generating means 61 of FIG.
In the above, the memory 62 may be configured by one relatively large ROM, or a readable / writable semiconductor memory (RAM) or the like may be used instead of the ROM. In addition, the switch 63b
Instead of the predetermined pseudo noise signal S61 from the memory 62 according to the address based on the output of the vehicle speed sensitive sensor 63a.
May be read. (Ii) On the output side of the microphone 51 of FIG. 1, the outputs of the D / A converters 80 and 90, and the input side of the A / D converter 82,
Various circuits for improving the call function may be added to the loudspeaker telephone of FIG. (iii) Although the loudspeaker telephone installed in a car has been described in the above embodiments, the present invention can be applied to various moving bodies such as ships and aircraft other than cars.

[0034]

As described in detail above, according to the first invention, the noise canceller is composed of at least the noise signal sequence generating means, the ADF, and the adder. Separately, a microphone and A for collecting traveling noise of a moving body as in the conventional case
There is no need to provide a / D converter. Therefore, without making a special feature on the moving body, with a simple circuit configuration, at low cost,
A noise canceller can be configured. Moreover, the operation of the echo canceller is stable, there is no acoustic echo feeling and howling, and running noise is not transmitted to the far-end speaker, so that a natural and high quality hands-free call becomes possible.

According to the second aspect of the present invention, the noise signal sequence generating means is composed of the memory and the switching means. Therefore, the pseudo noise signal which is very similar to the characteristic of the actual running noise can be accurately ADF by a simple circuit structure. Can be given to. Therefore, it is possible to accurately eliminate the actual running noise mixed in from the transmission microphone, and it can be expected that the quality of hands-free communication will be improved.

[Brief description of drawings]

FIG. 1 is a configuration block diagram of a public address telephone according to an embodiment of the present invention.

FIG. 2 is a configuration block diagram of a conventional public address telephone.

[Explanation of symbols]

51 microphone 53 voice detector 60 noise canceller 61 Noise Signal Sequence Generating Means 62 memory 63 switching means 63a Vehicle speed sensitive sensor 63b Switch 64 ADF 65 adder 70 Echo Canceller 71 ADF 72 adder 91 speaker

Claims (2)

[Claims] 1. A microphone for inputting a near-end talker voice signal, a speaker for outputting a far-end talker voice signal, a noise canceller for canceling a moving object running noise signal from the microphone, and an adaptive digital filter. A loudspeaker telephone including an echo canceller for canceling an acoustic echo from the microphone output from the noise canceller using the far-end talker voice signal as a reference signal, wherein the noise canceller corresponds to a traveling speed of a moving body. Noise signal sequence generating means for generating a pseudo noise signal, and estimating transfer characteristics from the noise source of the moving body to the microphone using the pseudo noise signal as a reference signal when the near-end talker voice signal is not input. A noise canceller adaptive digital filter for outputting a noise canceling signal, and the moving body Loudspeaker telephone is characterized by an adder for subtracting the noise cancellation signal from the line noise signal comprises a. 2. The loudspeaker telephone set according to claim 1, wherein the noise signal sequence generating means stores a memory of pseudo noise signal data according to a traveling speed of the moving body,
A loudspeaker telephone constituted by switching means for switching the output of the memory according to the traveling speed of the moving body and giving it to the noise canceller adaptive digital filter.