JP2010263401A - Handsfree speech communication device, and voice correcting method of the device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a handsfree speech communication device for specifying a speaker by acquiring a voice feature of the speaker, and to provide a voice correcting method of the device. <P>SOLUTION: The handsfree speech communication device includes a sampling section (44) for extracting beforehand a voice volume feature and a voice quality feature of the speaker from the voice of the speaker, and a correcting section (48) for setting a correction condition corresponding to the speaker beforehand from the extracted voice volume feature and voice quality feature, and on the other hand, specifying a caller by comparing the extracted voice volume feature and voice quality feature with a voice of the caller to output a corrected voice of the caller on the basis of the correction condition. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a hands-free communication device particularly suitable for use in a vehicle and a sound correction method for the device.

This type of hands-free communication device is electrically connected to a sound generator, for example, a driver's mobile phone, so that the driver can talk to the outside without holding the mobile phone of his / her own Is done. This is because the use of a mobile phone while driving a vehicle is regulated by law.
In addition, the call device has a structure in which the hook is hooked on, for example, an ear of a speaker, and is installed at an appropriate position in the vehicle.

  In this latter structure, the communication device is separated from the sound generator, and it is difficult to pick up the voice. Therefore, it is conceivable to use a technique in which a hands-free communication device includes a plurality of amplifier circuits (see, for example, Patent Document 1).

JP 2002-51142 A

By the way, a voice of a speaker has a unique waveform, and if the call device can obtain the characteristics of this waveform, the caller at the time of running should be identified and the voice should be clearly transmitted to the outside.
However, the above-described conventional technique has a problem that it is difficult to obtain the characteristics of the voice of the sound generator. This is because, by simply amplifying the voice of the sound generator, all the large and small waveforms of this voice are increased to the maximum value and there is no inflection.

As described above, in the above-described conventional technique, there is still a problem with respect to obtaining the characteristics of the voice of the sound generator.
SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a hands-free call device capable of solving the above-described problems, obtaining the voice characteristics of a sounder, and identifying the sounder, and a sound correction method for the device.

  According to a first aspect of the present invention for achieving the above object, there is provided a sampling unit for preliminarily extracting a sounder's volume characteristic and sound quality characteristic from a sounder's voice, and a correction condition corresponding to the sounder from the extracted volume characteristic and sound quality characteristic. On the other hand, there is provided a correction unit that compares the extracted volume feature and sound quality feature with the voice of the caller to identify the caller and outputs the corrected voice of the caller based on the correction condition.

  According to the first invention, the hands-free communication device includes the sampling unit and the correction unit. This sampling unit can extract the volume characteristic and the sound quality characteristic from the voice of the sound generator in advance.For example, the voice characteristic with the inflection is similar to the voice in a quiet environment where no noise related to the vehicle is generated. Obtainable. This cannot be obtained by simply amplifying the sound as in the prior art.

In addition, the correction unit presets correction conditions according to the sound generator from the extracted volume characteristics and sound quality characteristics, and the correction conditions can also be set from the characteristics of the speech with inflection.
The correction unit compares the extracted volume feature and sound quality feature with the voice of the caller during an actual call, so that the caller can be reliably identified. Further, since the correction unit outputs the corrected voice of the caller based on the preset correction condition described above, the voice that draws the characteristics of the caller is provided outside, that is, the other party of the call. Will be transmitted. As a result, the reliability of the hands-free call device is greatly improved.

According to a second aspect of the invention, in the configuration of the first aspect of the invention, the correction condition for the sound quality feature is set according to the main band frequency of the voice of the sound generator.
According to the second invention, in addition to the operation of the first invention, if the correction condition for the sound quality feature is classified according to the level of the main band frequency, such as male and female, noise suppression suitable for a speaker is achieved. Can be done.

According to a third invention, in the configuration of the second invention, the correction unit has a filter for cutting outside the main band.
According to the third invention, in addition to the operation of the second invention, if the outside of the main band is cut by a filter, only the main band remains, so that the other party's sound quality characteristics are further improved on the other side of the call. The voice that is pulled out further is transmitted.

According to a fourth invention, in the configurations of the first to third inventions, a storage unit for storing a finite number of voices of a sound generator is further provided.
According to the fourth invention, in addition to the operations of the first to third inventions, if the storage unit stores the voice of the speaker, the confirmation of the speaker can be facilitated, and the number of storage is limited. Therefore, the correction unit only needs to compare the finitely extracted volume feature and sound quality feature with the voice of the caller, and contributes to improving the accuracy of the caller.

According to a fifth aspect of the present invention, in the configuration of the first to third aspects of the present invention, the apparatus further includes a storage unit that stores a finite number of the extracted volume characteristics and sound quality characteristics.
According to the fifth aspect of the invention, in addition to the effects of the first to third aspects of the invention, if the volume feature and the sound quality feature extracted by the storage unit are stored in a finite amount, the correction unit may Since the comparison between the sound quality feature and the voice of the caller is sufficient, this point also contributes to the improvement of the specific accuracy of the caller.

  A sixth invention is a voice correction method for a hands-free call device. Then, in an environment where no noise related to the vehicle is generated, a step of extracting the sound volume characteristic and sound quality characteristic of the sound generator from the sound of the sound generator, and a correction condition corresponding to the sound generator are set from the extracted volume characteristic and sound quality characteristic. A step of comparing the extracted volume feature and sound quality feature with the voice of the caller in the vehicle and identifying the caller during the call, and outputting the caller's corrected voice based on the correction condition Consists of

According to the sixth aspect, the sound volume characteristic and the sound quality characteristic of the sound generator are extracted in an environment in which no noise related to the vehicle is generated, and it is possible to obtain a voice characteristic having an inflection. Also, this correction condition can be set from the characteristics of speech with inflection.
Then, during an actual call, the extracted volume characteristics and sound quality characteristics are compared with the voice of the caller, and the caller can be reliably identified. Furthermore, since the caller's corrected voice is output based on the above-described preset correction conditions, the caller's characteristics can be transmitted to the other party of the call.

  According to the present invention, since a voice characteristic having an inflection equivalent to that in a quiet environment where noise is not generated is obtained, a sound producer is surely identified, and the voice from which the characteristic is extracted is transmitted to the other party of the call A hands-free communication device that communicates with the device and a sound correction method for the device can be provided.

It is explanatory drawing in the vehicle carrying the hands-free call apparatus of a present Example. It is a figure explaining the relationship between the hands-free telephone apparatus of FIG. 1, and the exterior. It is an internal block diagram of the hands-free call apparatus of FIG. It is an operation | movement flowchart by the hands-free call apparatus of FIG. It is explanatory drawing of the sound quality correction conditions by the correction | amendment part of FIG.

Preferred embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 shows a passenger compartment of a vehicle 1, and a driver's seat 4 is shown on the right front side of the figure. A passenger seat is provided on the left side of the driver seat. The instrument panel 2 is disposed in front of each seat when viewed from the seat 4, and the windshield 6 is also provided in front of the panel 2 when viewed from the seat 4, so that the traveling direction of the vehicle 1 can be viewed from each seat. .

  In this driver's seat, the instruments 8 are arranged in front of the panel 2 and display the speed, travel distance, shift position, etc. of the vehicle 1. The handle 10 is rotatably attached to the tip of a shaft 12 extending from the lower side of the panel 2 toward the seat 4. The shift lever 14 extends from the peripheral wall of the shaft 12 toward the passenger seat and is movable in the vertical direction.

The air outlet 16 of the air conditioning device is arranged in front of the panel 2 from the driver's seat to the passenger seat, and air is sent out toward the vehicle interior as the device is activated.
The monitor 18 is installed in front of the panel 2 between the driver seat and the passenger seat. The monitor 18 can display a map, an operation menu, and the like as the navigation device is activated.

The center console 20 protrudes from the lower side of the monitor 18 toward the seat 4, and the slot 22 of the audio device is disposed in the vicinity of the monitor 18, and a CD, MD, or the like can be inserted therein.
Here, in the present embodiment, the hands-free communication device 24 is provided at an appropriate position on the seat 4 side of the console 20 in the console 20.

The call device 24 is embedded in the console 20 leaving the periphery of the operation panel, microphone, etc., and can detect the sound of the sound generator. The call device 24 can make a call with another mobile phone 60 via the mobile phone 26 owned by the sound generator shown in FIG.
Specifically, the call device 24 according to the present embodiment is configured to be connectable to a cellular phone 26 owned by a speaker via a short-distance wireless line. In addition, it may replace with this radio | wireless line and may be connected with the cable.

  On the other hand, the mobile phone 26 is connected to the mobile phone 60 on the other side via a public line such as a communication network for mobile phones. Thereby, the voice of the above-mentioned sound generator is transmitted from the microphone 40 to the other party through the telephone device 24, the mobile phone 26, and the mobile phone 60, and the voice of the other party passes through the mobile phone 60, the mobile phone 26, and the telephone device 24. The sound is transmitted from the speaker 38 to the speaker.

  More specifically, as shown in FIG. 3, the call device 24 of this embodiment has an antenna 30, and this antenna 30 is connected to a communication unit 32. The communication unit 32 transmits / receives a wireless call signal to / from the mobile phone 26, and the signal received via the antenna 30 is demodulated and output from the communication unit 32 toward the signal processing unit 34.

The signal processing unit 34 performs signal processing necessary for communication. When the processing result is output to the speaker D / A 36, the speaker 38 rings and the other party's voice is output to the passenger compartment. .
On the other hand, the voice of the speaker in the passenger compartment is picked up by the microphone 40 and input from the microphone A / D 42 to the signal processing unit 34. In this signal processing unit 34, processing for suppressing environmental noise in the passenger compartment and processing for removing echo components that have circulated from the speaker 38 to the microphone 40 are also performed. When the processing result is output to the communication unit 32, The modulated signal is transmitted from the antenna 30.

By the way, the communication device 24 of this embodiment further includes a control unit (sampling unit) 44, a storage unit 46, and a correction unit 48.
Specifically, the control unit 44 is configured to be able to previously extract the sound characteristics of the sound generator from the sound of the sound generator in the passenger compartment, and in response to a mode signal from a host CPU (not shown), For example, the sound volume characteristic and the sound quality characteristic of the sound generator are extracted from the voice print of the sound of the sound input.

The voice of the sound generator itself and the extracted volume characteristics and sound quality characteristics are output to the storage unit 46 and stored within the capacity range. That is, a finite number of sound generator's voices, extracted volume features and sound quality features are stored.
The correction unit 48 according to the present embodiment is arranged on the downstream side of the signal processing unit 34 when viewed in the transmission direction of the voice of the sound generator, and is configured so that correction conditions according to the extracted volume characteristic and sound quality characteristic can be set in advance. Has been. The storage unit 46 can also store this correction condition.

Further, the correction unit 48 can specify the subsequent caller, that is, the owner of the mobile phone 26 that is the driver of the vehicle 1.
Specifically, in this embodiment, the correction unit 48 identifies the speaker by comparing the volume characteristics and sound quality characteristics of the previously extracted speaker with the voice of the speaker who actually makes a call with the other party. is doing.

Further, the correction unit 48 outputs the corrected voice of the caller to the communication unit 32 based on the correction condition set for each sound generator.
Note that the control unit 44 described above also extracts the voice characteristics of the subsequent caller in addition to the speaker, and the correction unit 48 extracts the volume characteristic and sound quality characteristic of the speaker that has been extracted in advance and the speaker. The caller may be specified by comparing with the voice feature of

Referring to FIG. 4, there is shown an operation flow chart until the voice of the sound generator by the call device 24 is output to the other party. Hereinafter, the operation according to the present invention of the hands-free communication device 24 configured as described above will be described.
In step S401 in FIG. 5, a speaker who stores the cellular phone 26 in a bag, for example, gets on the vehicle 1, and the communication device 24 is activated before the engine of the vehicle 1 is started, for example, when the ignition key of the vehicle 1 is turned on. To do.

  Next, in step S402, the voice of the speaker who is on the microphone 40 is captured. Here, this voice is generated in a specific environment such as “Hoshi” or “Hoshi” in a quiet environment where no noise related to the vehicle such as road noise or wind noise of the vehicle 1 occurs, such as when the vehicle 1 is stopped. Get a short voice phrase toward the microphone 40, such as "I'll get on" or "Yes, I'm in sales department 1."

In step S403, the host CPU determines whether or not a sound generator is registered. If the sound generator is not yet registered in the vehicle 1, that is, if it is determined YES, the process proceeds to step S404. .
In step S404, the host CPU outputs a recording mode signal to the control unit 44.

When receiving the mode, the control unit 44 extracts the sound volume characteristic and the sound quality characteristic of the speaker from the short voice phrase captured by the microphone 40. As a result, it is possible to recognize the level of the voice of the pronunciation and the gender.
At the same time, the correction unit 48 sets a volume correction condition and a sound quality correction condition according to the sound generator from the extracted volume feature and sound quality feature.

Specifically, the volume correction condition is for adjusting the volume to be suitable for a sounder. For example, when the sounder is a low voice, the amplification amount is increased while the volume is high. In order to make it easier to listen to, the amount of amplification is narrowed down.
Also, the sound quality correction condition is for performing noise suppression suitable for the sound producer. More specifically, noise is generally present in a lower frequency range, as indicated by hatching in FIG.

Here, it is said that the main band frequency of men is about 120 Hz, whereas the main band frequency of women is about 220 Hz. It can be seen that there are more areas overlapping with the noise than in FIG.
Therefore, the correction unit 48 of the present embodiment sets the sound quality correction condition according to the main band frequency of the voice of the sound generator, and uses a filter that leaves the main band frequency and cuts outside this band.

In other words, for example, if the speaker is a male, the amount of noise cut is reduced, and by simply cutting the noise shown in FIG. To do.
On the other hand, for example, when the speaker is a woman, the amount of noise cut is increased compared to the man. This is because, as shown in FIG. 5B, when the main band frequency is high, the characteristics of the woman can be expressed even if the noise is simply cut.

Subsequently, in step S405, the storage unit 46 stores the voice of the sound generator, its voice characteristics, and correction conditions, and the process proceeds to step S406.
In step S406, it is determined whether or not the mobile phone 26 and the outside of the vehicle 1 are in a call state via a public line, and the call device 24 is in a standby state until the call state is reached.

Thereafter, when the mobile phone 26 and the outside of the vehicle 1 reach a call state, that is, when it is determined YES, the process returns to step S402, and the microphone 40 captures the voice of the caller in the passenger compartment.
In step S403, it is determined whether or not the caller is registered. If the caller has already been registered in the vehicle 1, the host CPU outputs a normal mode signal to the control unit 44. Then, the process proceeds to step S407.

In step S407, the correcting unit 48 receives the voice of the caller captured by the microphone 40. Note that the control unit 44 may extract all registered voice data of the sound generator from the storage unit 46 and output the voice data to the correction unit 48 in order to more reliably identify the caller.
Then, the correcting unit 48 compares the sound volume characteristics and sound quality characteristics of the speaker extracted in advance with the subsequent voices of the caller, identifies the caller, and proceeds to step S408.

In step S <b> 408, the correction unit 48 outputs the corrected voice of the caller to the communication unit 32 based on the set correction condition.
More specifically, the volume of the caller is corrected according to the volume correction conditions described above. If the caller is a loud speaker, the amount of amplification is increased and the speaker is a loud speaker. In such a case, the amount of amplification is reduced so that it can be easily heard.

Furthermore, the sound quality of the caller is also corrected according to the sound quality correction conditions described above. If the caller is a male speaker, the amount of noise cut is reduced, and the sounder is a female speaker. Increase the amount of noise cut.
Next, in step S409, the communication unit 32 modulates the signal from the correction unit 48 and outputs the modulated signal to the antenna 30, and exits a series of routines.

  As described above, according to the present embodiment, the hands-free call device 24 includes the control unit 44 and the correction unit 48. The control unit 44 can previously extract the volume characteristic and the sound quality characteristic from the voice of the sound generator. For example, the voice of the voice with inflection can be obtained in the same manner as the voice in a quiet environment where no noise is generated with respect to the vehicle 1. Features can be obtained. This cannot be obtained by simply amplifying the sound as in the prior art.

Further, the correction unit 48 presets a correction condition according to the sound generator from the extracted volume feature and sound quality feature, and this correction condition can also be set from the feature of the voice having an inflection.
And since this correction | amendment part 48 compares the extracted volume characteristic and sound quality characteristic with the voice of the caller of the vehicle 1 at the time of an actual call, it becomes possible to identify a caller reliably. Further, since the correction unit 48 outputs the caller's corrected voice to the communication unit 32 based on the above-described preset correction conditions, the outside of the caller, that is, the other party of the call, The voice that brings out the features will be transmitted.

  As a result, if a speaker is registered on the vehicle 1 side, the speaker who has boarded the vehicle 1 can send a specific short voice phrase to the microphone 40 in a quiet environment such as when the vehicle 1 is stopped. When speaking in advance, the call device 24 outputs a voice that draws out the features of the subsequent caller without performing any operation on the hands-free call device 24, so the convenience of the call device 24 is improved. In addition, its reliability is greatly improved.

Further, if the sound quality correction conditions are classified according to the level of the main band frequency, such as male and female, noise suppression suitable for a speaker can be performed.
Further, if the outside of the main band is cut by the filter, only the main band remains, so that the voice that further draws out the voice quality characteristics of the caller is transmitted to the other party of the call.
Furthermore, the volume correction condition is set according to the volume of the voice. For example, if the amplification amount is increased in the case of a loud voice and the volume is adjusted so that it can be easily heard in the case of a loud voice, Can be adjusted.

In addition, if the storage unit 46 stores the voice of the sounder, the confirmation of the sounder is facilitated, and since the number of storages is finite, the correction unit 48 includes the finitely extracted volume feature and sound quality feature. The comparison with the subsequent caller's voice is sufficient, which contributes to the improvement of the accuracy of the caller.
Further, if the volume feature and sound quality feature extracted by the storage unit 46 are stored in a finite amount, the correction unit 48 can compare the finitely extracted volume feature and sound quality feature with the subsequent voice of the caller. This also contributes to improving the accuracy of the caller.

The present invention is not limited to the above embodiments, and various modifications can be made without departing from the scope of the claims.
For example, in the above embodiment, the control unit 44 and the correction unit 46 are provided separately. However, the present invention is not necessarily limited to this form, and the control unit and the correction unit are provided in the same circuit. Also good.

Further, the hands-free communication device of the present invention may be installed on the instrument panel 2 or on the front pillar in addition to the center console 20.
In any of these cases, similar to the above, it is possible to obtain the sound characteristics of the sound generator and specify the sound generator.

1 Vehicle 24 Hands-free call device 44 Control unit (sampling unit)
46 storage unit 48 correction unit

Claims (6)

A sampling unit that previously extracts the sound volume characteristics and sound quality characteristics of the sound generator from the sound of the sound generator;
A correction condition corresponding to the sound generator is set in advance from the extracted volume feature and sound quality feature, while the extracted volume feature and sound quality feature is compared with the voice of the caller to identify the caller and the correction A hands-free call device comprising: a correction unit that outputs a correction voice of the caller based on a condition. The hands-free call device according to claim 1,
The hands-free call device according to claim 1, wherein the correction condition for the sound quality feature is set according to a main band frequency of the voice of the sound generator. The hands-free call device according to claim 2,
The correction unit has a filter that cuts out of the main band. The hands-free call device according to any one of claims 1 to 3,
A hands-free call device, further comprising a storage unit for storing a finite number of voices of the sound generator. The hands-free call device according to any one of claims 1 to 3,
A hands-free call device, further comprising a storage unit for storing a finite number of the extracted volume features and sound quality features. Extracting the sounder's volume characteristics and sound quality characteristics from the sound of the sounder in an environment free from noise related to the vehicle;
Setting a correction condition according to the sound generator from the extracted volume feature and sound quality feature;
Comparing the extracted volume feature and sound quality feature with the voice of the caller in the vehicle during a call to identify the caller;
Outputting a corrected voice of the caller based on the correction condition.

Images