JP3213145B2 - Automotive audio equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an on-vehicle audio device, and more particularly to an on-vehicle audio device capable of performing a conversation smoothly even while listening to audio.

[0002]

2. Description of the Related Art In recent years, audio devices for vehicles have been used for CDs,
Digital audio sources such as DAT, MD, DCC and the like have been used to enhance the sound quality, and high power amplifiers have been used to increase the sound volume, making the system comparable to home audio. Also, it is customary that one set of left and right speakers is installed on the front side and one set of left and right speakers is installed on the rear side.

[0003]

However, along with the enhancement of the in-vehicle audio device, even if the volume is increased, sound can be reproduced without distortion, so that there is a tendency to listen to music at a higher volume. If a person is playing music while riding, the speaker output near him or her may be too large to hear and hear other occupants. In this case, if the audio volume is lowered, conversation becomes possible, but the music becomes difficult to hear this time. In view of the above, it is an object of the present invention to provide an in-vehicle audio device that allows a conversation to be performed smoothly even while listening to audio.

[0004]

SUMMARY OF THE INVENTION According to the present invention, there is provided an utterance detecting means for detecting an occupant's utterance, and when a utterance of a certain occupant is detected by the utterance detecting means, a speech voice is included in an audio signal. This is achieved by providing a level lowering means for lowering the level of part or all of the band and inputting it to the speaker.

[0005]

According to the present invention, when an occupant's utterance is detected,
Of the audio signals, the level of part or all of the speech voice band is reduced and input to the speaker. This allows
When a certain occupant speaks, the level of part or all of the speech voice band output from the speaker is lowered, making it easier to hear the utterance, so that even when listening to audio, it is possible to have a smooth conversation without having to lower the audio volume. it can. Also, when the conversation is interrupted, you can fully enjoy the music with the original volume and sound quality.
Since the level of the other band of the speaker output does not decrease, the music can be heard with a certain sound quality.

When the level of the conversation voice band is lowered, the level lowering means also lowers the band level below the conversation voice band. This makes it easier to hear the utterance.

The utterance is detected for each occupant, and when a certain occupant speaks, the level of a part or all of the speech voice band in the audio signal relating to the speaker near the other occupant is reduced. Thereby, the occupant who is speaking can listen to music satisfactorily.

[0008]

FIG. 1 is an overall configuration diagram of an on-vehicle audio device according to an embodiment of the present invention. In the figure, reference numeral 1 denotes a head unit, which in the present embodiment outputs digital audio signals of left and right two channels reproduced from a digital audio source such as a CD. 2a, 2b
Is a volume circuit (VO
L), the volume is increased or decreased in units of 1 dB, for example, in response to volume control by a control unit described later. Volume circuit 2
a and 2b adjust the volume in the digital domain. Reference numeral 3 denotes an operation unit for performing a volume changing operation and the like, and 4 denotes a volume data variable in accordance with the volume changing operation of the operation unit.
a, 2b and set to the volume indicated by the volume data.

Reference numerals 5a and 5b denote filters provided after the output side of the volume circuit 2a is branched into two systems, and filters 5c and 5d are provided after the output side of the volume circuit 2a is branched into two systems. This is a filter with variable characteristics. These filters 5a to 5d are configured using, for example, a DSP (Digital Signal Processor), and variably set filter coefficients by a control unit described later, thereby changing the filter characteristics. In particular,
The attenuation rate of the conversation voice band of 0.3kHz to 3kHz is -α (dB)
(Α is a positive value, and here, for example, α = 40) or 0 (dB). At this time, the attenuation rate of the other bands is kept at 0 (dB). 6a
To 6d are D / A provided individually on the output side of each filter.
Converters, 7a to 7d are power amplifiers (AMP) provided on the output side of each D / A converter, 8a is a front L channel speaker provided on the output side of the power amplifier 7a, 8
b is a rear L channel speaker provided on the output side of the power amplifier 7b, 8c is a front R channel speaker provided on the output side of the power amplifier 7c, and 8d is a rear R channel provided on the output side of the power amplifier 7d. These are channel speakers, and are installed at predetermined locations in the front and rear of the vehicle 9 as shown in the lower right of FIG. Note that the vehicle 9 is a four-seater, and the occupants for each seat are distinguished by A to D.

The microphones 10a to 10d are individually installed in front of the occupants and have microphones having sharp directivity for individually catching the utterances of the occupants A to D. Amplifying microphone amplifier, 12a-
12d is an A for individually A / D converting each microphone amplifier output.
A / D converters 13a to 13d are individually provided on the output side of each A / D converter, and include 0.3 kHz including a road noise component.
An HPF that cuts the following:

Reference numeral 14 denotes the outputs of the filters 5a to 5d and each H
This is an utterance detection unit that detects an occupant's utterance from the output of the PF, and is specifically configured as shown in FIG. In FIG. 2, 15 a to 15 d are 0.3 kHz for each filter output.
HPF that cuts the following, HPFs 13a to 13d
It is provided corresponding to. Reference numerals 16a to 16d denote correlation degree inspection units for inspecting the degree of correlation between the microphone input predicted from the outputs of the filters 15a to 15d and the actual microphone input for each microphone. The correlation degree inspection units 16a to 16d have the same configuration. Among them, 16a will be described. 17a is an L-channel audio signal output from the filter 15a.
A delay circuit (DL) for delaying by the time required to reach the microphone 10a, and a delay circuit 17b for delaying by the time required for the L-channel audio signal output from the filter 15b to reach the microphone 10a via the speaker 8b. , 17c are the microphones 1 through which the R channel audio signal output from the filter 15c is
The delay circuit 17d delays by the time required to reach 0a, and the microphone 17 receives the R-channel audio signal output from the filter 15d via the speaker 8d.
This is a delay circuit that delays by the time required to reach a.

Reference numerals 18a to 18d denote delay circuits 17a to 17d
A level adjustment circuit for reducing the level of each output of the microphone 10a so as to be in inverse proportion to the square of the distance between the speakers 8a to 8d and the microphone 10a. 19 adds the outputs of the level adjustment circuits 18a to 18d to estimate the output of the microphone 10a. The adders 20 and 21 are for generating a microphone input signal. Each of the adders 20 and 21 is a cascaded two-stage latch circuit that performs a latch operation on a predicted microphone input signal based on a sample clock. The previous sample data is output from the circuit 21. Reference numeral 22 denotes a comparison circuit, which is based on the previous sample data latched by the latch circuit 21 as shown in FIG.
When the current sample data latched by the above is the same or larger, "1" is output, and when the current sample data is smaller, "0" is output. Thus, when the predicted microphone input signal tends to increase, the comparison circuit 22 outputs “1” for each sample clock. Conversely, when the predicted microphone input signal tends to decrease, the comparison circuit 22 outputs the comparison circuit for each sample clock. 22
Outputs “0”.

Reference numerals 23 and 24 denote latch circuits each having a two-stage cascade configuration for performing a latch operation on an actual microphone input signal (output signal of the HPF 13a) based on a sample clock. The latch circuit 23 outputs the current sample data. The previous sample data is output from the latch circuit 24. 25 is a comparison circuit, as shown in FIG.
Outputs "1" when the current sample data latched by the latch circuit 23 is the same or larger with respect to the previous sample data latched by the latch circuit 24, and outputs "0" when the current sample data is small. I do. Thus, when the actual microphone input signal tends to increase, the comparison circuit 25 outputs “1” for each sample clock, and conversely, when the actual microphone input signal tends to decrease, the comparison circuit 25 outputs “1”. The comparison circuit 22 outputs “0”.

Reference numeral 26 denotes an output of the comparison circuit 22 and the comparison circuit 25.
And a comparison circuit that outputs “1” when they match and outputs “0” when they do not match.
Reference numeral 7 denotes a counter that counts the number of times that the output of the comparison circuit 27 becomes “0” based on the sample clock. When a predetermined period (for example, 10,000 samples) elapses after the count value is reduced to zero at a certain timing. , And outputs the count value to a determination unit, which will be described later, and restarts counting with the count value set to zero again. When the counter 27 outputs the count value, if the count value is small, the microphone 10a
Indicates that the degree of correlation between the predicted microphone input signal and the actual microphone input signal is large, and when the count value is large, it indicates that the degree of correlation is small. If the occupant A is not speaking, the correlation should be large, and if the occupant A is speaking, the correlation should be small. Therefore, the count value is set to a predetermined reference value (for example, the counting period of the counter 27 is set to 1). By comparing with a reference value of 5000 for 10,000 samples), it is possible to determine whether or not the occupant A speaks (see FIG. 3 (2)).

The correlation degree inspection unit 16a is configured as described above for the microphone 10a.
Regarding the microphones 10b to 10d also for b to 16d,
The counter is configured in the same manner as the correlation degree inspection unit 16a, and each of the counters outputs a count value indicating the degree of correlation between the predicted microphone input signal and the actual microphone input signal every time a predetermined period elapses. Reference numeral 28 denotes a determination unit which receives the count value of the counter from each of the correlation degree inspection units 16a to 16d and determines whether or not there is an utterance for each passenger. Specifically, each correlation degree inspection unit 1
The count values of the counters input from 6a to 16d are individually compared with a predetermined reference value, and when any of them is below the reference value, it is determined that no one is speaking, and one of them is above the reference value. At this time, for example, the correlation degree inspection unit 16a
Is determined to be greater than the corresponding occupant A is speaking. If a plurality of count values exceed the reference value at the same time, it is determined that one occupant corresponding to the correlation degree inspection unit 16a having the largest count value is speaking. The determination result is output to a level lowering unit described later.

Returning to FIG. 1, reference numeral 29 denotes a speech band of 0.3 kHz to 3 kHz in the audio signal when there is a speaker based on the detection result of the speech detection section 14 (the judgment result of the judgment section 28). Is a level lowering unit for lowering the level by a certain level. Of these, reference numeral 30 denotes a control unit, and based on the detection result input from the utterance detection unit 14, when it is determined that no speaker is present, any of the filters 15a to 15d sets the attenuation rate of the entire band to 0 ( dB), and when it is determined that the occupant A is speaking, the filters 15b to 15b
15d conversation voice band 0.3kHz ~ 3kHz attenuation rate -40
(DB), and when it is determined that the occupant B is speaking, the attenuation rate of the conversation voice band of 0.3 to 3 kHz of the filters 15a, 15c, and 15d is set to -40 (dB), and the occupant C is set. Is determined to be speaking, the speech voice band of the filters 15a, 15b, and 15d is 0.3 kHz to 3 kHz.
Is set to −40 (dB), and when it is determined that the occupant D is speaking, the attenuation rate of the filters 15 a to 15 d for the conversation voice band of 0.3 kHz to 3 kHz is set to −40 (dB). Let it.

Next, the operation of the above embodiment will be briefly described. If nobody is speaking while listening to music reproduced from the digital audio source at a certain volume, the count value output from the counter of each of the correlation degree checking units 16a to 16d of the utterance detecting unit 14 may change. Is lower than the reference value, and the determination unit 28 outputs a determination result indicating that no one is speaking to the control unit 30 of the level reduction unit 29. At this time, the control unit 30 sets the conversation voice band of the filters 15a to 15d to 0.3 k.
The attenuation rate of Hz to 3 kHz is set to 0 (dB) (the attenuation rates of other bands are also 0 (dB)), and the volume and sound quality of the audio signal are not changed. Therefore, each occupant can fully enjoy the music.

Here, for example, when the occupant A speaks for a conversation with the occupant D, the count value output from the counter 27 of the correlation degree inspection unit 16a of the utterance detection unit 14 exceeds the reference value, and the judgment is made. The unit 28 determines that the occupant A has spoken, and outputs the determination result to the control unit 30. Then, the control unit 30 sets the conversation voice band for the filters 15b to 15d.
The attenuation rate of 0.3 kHz to 3 kHz is set to -40 (dB).
Therefore, the level of the conversation voice band of the music output from the speakers 8b to 8d is lowered, so that the occupants B to D can hear the voice of the occupant A.
Since the level does not change in the band above 3kHz, the occupants B to D
Can enjoy music with a certain volume and sound quality.
At this time, since the output level of the speaker 8a near the occupant A does not change in all bands, it is possible to satisfactorily listen to music irrespective of its own utterance.

Next, when the occupant D confirms the call from the occupant A and utters an answer, the count value output from the counter of the correlation degree inspection unit 16d of the utterance detection unit 14 exceeds the reference value, and the determination unit 28 Determines that the occupant D has spoken, and outputs the determination result to the control unit 30. Then, the control unit 30 controls the filters 15a to 15c for the conversation voice band of 0.3 kHz.
The decay rate of 33 kHz is set to −40 (dB), so that the occupants A to C can hear the voice of the occupant D. At this time, the level of 0 to 0.3 kHz and the band above 3 kHz does not change. Therefore, the occupants A to C can enjoy music with a certain volume and sound quality. At this time, the speaker 8 near the occupant A
Since the output level of “a” does not change in all bands, it is possible to listen to music satisfactorily irrespective of its own utterance. Thereafter, the occupants A and D can proceed with the conversation in the same manner.

According to this embodiment, when someone speaks, the other occupants decrease the level of the conversation voice band of 0.3 kHz to 3 kHz of music output from a nearby speaker by a fixed amount.
The user can easily hear the uttered voice, and can have a smooth conversation while listening to the audio without performing an operation of lowering the audio volume. In addition, when the conversation is interrupted, the music can be fully enjoyed by the original volume and sound quality.
Since the band levels of 0.3 kHz and 3 kHz or more do not drop, music can be enjoyed with a certain volume and sound quality. Then, since the output level of the speaker near the speaker does not change in all bands, it is possible to satisfactorily listen to music irrespective of the utterance of the speaker.

In the above embodiment, the HPF 1
After rectifying and smoothing the outputs of 3a to 13d, the outputs of the HPFs 15a to 15d were individually rectified and smoothed in the identification unit 14 by inputting them to the correlation degree inspection units 16a to 16d of the utterance detection unit 14. And each correlation degree inspection unit 16a
To 16d. Also, the judgment unit 2
8 has output the determination result of the content in which the speaker does not exist until then, when the determination result of the content in which the speaker exists at a certain point in time is output, the output is held for a certain period of time, After a certain period of time, the determination may be made again. Conversely, although the determination result of the content where the speaker is present is output, when it is determined that the speaker no longer exists at a certain point in time, when the state where the speaker does not exist continues for a certain period of time, For the first time, a determination result of a content in which no speaker exists may be output. Further, the level adjustment circuits 18a to 18d in the correlation degree inspection units 16a to 16d of the utterance detection unit 14 may adjust the level in consideration of not only the distance between the microphone and each speaker but also the directivity of the microphone. Good.

The utterance detecting section may be modified as shown in FIG. 4, reference numerals 40a to 40d denote HPFs provided on the output side of the filters 15a to 15d in FIG. 41a-
Reference numeral 41d is a reference level generator provided separately for the microphones 10a to 10d. The reference level generator 41d calculates the level of the predicted microphone input signal and generates a reference level that is higher than the level by a certain level. Since the reference level generation units 41a to 41d have the same configuration, the reference level generation unit 41a will be described. Among them, reference numeral 42a indicates only the time required for the L-channel audio signal to reach the microphone 10a via the speaker 8a. A delay circuit for delaying the delay 42b, a delay circuit for delaying the time required for the L channel audio signal to reach the microphone 10a via the speaker 8b, and a delay circuit 42c for the R channel audio signal reaching the microphone 10a via the speaker 8c A delay circuit 42d for delaying by the time required until the R channel audio signal reaches the microphone 10a via the speaker 8d.

Reference numerals 43a to 43d denote delay circuits 42a to 42d.
43a is a level adjustment circuit for individually adjusting the level of each output, and 43a adjusts its own output level for the L channel audio signal to be the same as the level of the output component of the speaker 8a included in the output of the HPF 13a,
43b is a speaker 8b whose output level for the L channel audio signal is included in the output of the HPF 13b.
43c.
Is adjusted so that its own output level with respect to the R channel audio signal is the same as the level of the output component of the speaker 8c included in the output of the HPF 13c.
The adjustment is performed so that the level of the output component of the speaker 8d included in the output of the PF 13d becomes the same.

An adder 44 adds the outputs of the level adjustment circuits 43a to 43d to create a predicted microphone input signal for the microphone 10a, and 45 a rectifier and a smoother for the adder output to perform the level of the predicted microphone input signal. , A reference level generating circuit for generating a reference level signal raised by a certain level (for example, 8 dB) with respect to the output of the rectifying / smoothing circuit 45. The reference level signal will be described later. Output to the judgment unit. The other reference level generating units 41b to 41d also use the microphones 10b to 10d.
For d, the reference level generator 4 for the microphone 10a
It is configured similarly to 1a.

47a to 47d are HPFs 13a to 13d
Rectifier / smoothing circuit that individually rectifies and smoothes the outputs of
A determination unit 8 compares the reference level signal input from each of the reference level generation units 41a to 41d with the actual microphone input signal level input from the rectification / smoothing circuits 47a to 47d, and determines whether or not the occupant speaks. It is. Specifically, the actual microphone input signal level input from the rectifier / smoothing circuit 47a for the microphone 10a is used as the reference level generator 41.
a, and if the actual microphone input signal level is lower than the reference level,
It is determined that the occupant A is not speaking, and conversely, if the actual microphone input signal level is higher than the reference level, it is determined that the occupant A is speaking. Similarly, the microphone 10b
The actual microphone input signal level input from the rectification / smoothing circuit 47b is compared with the reference level input from the reference level generation unit 41b, and if the actual microphone input signal level is lower than the reference level, When it is determined that the occupant B is not speaking, and when the actual microphone input signal level is higher than the reference level, it is determined that the occupant B is speaking, and the rectifying / smoothing circuit 47c for the microphone 10c is determined. Is compared with the reference level input from the reference level generator 41c. If the actual microphone input signal level is lower than the reference level, the occupant C does not speak. Conversely, when the actual microphone input signal level is higher than the reference level, it is determined that the occupant C is speaking,
For the microphone 10d, the actual microphone input signal level input from the rectifier / smoothing circuit 47d is used as the reference level generator 4.
If the actual microphone input signal level is lower than the reference level, it is determined that the occupant D is not speaking, and conversely, the actual microphone input signal level is If it is higher than the reference level, it is determined that the occupant D is speaking.

When the actual microphone input level is lower than the corresponding reference level for any of the microphones, a determination result indicating that there is no speaker is output, and the actual microphone input level corresponds to only one microphone. When the reference level is higher than the reference level, a judgment result indicating that the corresponding occupant is speaking is output, and when the actual microphone input level is higher than the corresponding reference level for a plurality of microphones at the same time, the corresponding result is output. Among them, a determination result that one occupant corresponding to the microphone having the largest difference between the microphone input level and the reference level is speaking is output. The presence / absence of an occupant-specific utterance can also be detected by the utterance detection unit configured as shown in FIG.

In the modification shown in FIG.
8 has output the determination result of the content in which the speaker does not exist until then, when the determination result of the content in which the speaker exists at a certain point in time is output, the output is held for a certain period of time, After a certain period of time, the determination may be made again. Conversely, although the determination result of the content where the speaker is present is output, when it is determined that the speaker no longer exists at a certain point in time, when the state where the speaker does not exist continues for a certain period of time, For the first time, a determination result of a content in which no speaker exists may be output.

Further, in the above-described embodiment and the modified example, the level of the conversation voice band of the speaker output near the occupant other than the speaker is reduced by a fixed amount. However, the output of all speakers including the speaker near the speaker is reduced. May be lowered by a certain amount. Also, the level reduction was all of the conversation voice band of 0.3 kHz to 3 kHz,
Only some of them may be used. Furthermore, if the level of the band of 0 to 0.3 kHz is also lowered in addition to part or all of the speech voice band, it becomes easier to hear when one occupant speaks to another occupant. .

Further, the output levels of the volume circuits 2a and 2b of FIG. 1 are detected (a volume value managed by the control unit 4 can be used for the detection). The control of 30 may be performed, and when it is less than a certain value, the attenuation rates of the filters 5a to 5d may be fixed to 0 (dB) over the entire band. Further, the control unit 30 may vary the attenuation rate and the attenuation band set in the filters 5a to 5d according to the output levels of the volume circuits 2a and 2b. Specifically, the attenuation rate is increased or the attenuation band is widened in proportion to the output levels of the volume circuits 2a and 2b.

[0030]

As described above, according to the present invention, the conversation detecting means for detecting the conversation of the occupant, and the level for lowering the level of the conversation voice band in the audio signal when the conversation detection means detects the conversation of the occupant. And a lowering means for detecting conversation of the occupant and, when the conversation is detected, lowering the conversation voice band level of the audio signal. , The conversation can be smoothly performed even during the audio listening without performing the operation of lowering the audio volume. Also, when the conversation is interrupted, the original volume,
Depending on the sound quality, the music can be fully enjoyed, and even during conversation, the music can be heard with a certain sound quality because the remaining band level of the audio signal does not decrease.

Further, when the level lowering means is configured to lower the level of the conversation voice band when lowering the level of the conversation voice band, the utterance is more easily heard.

In addition, the utterance is detected for each occupant, and when a certain occupant speaks, the level of a part or all of the speech voice band of the audio signal relating to the speaker near the other occupant is reduced. Therefore, the occupant who is speaking can listen to music satisfactorily.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of an on-vehicle audio device according to an embodiment of the present invention.

FIG. 2 is a specific configuration diagram of an utterance detection unit in FIG. 1;

FIG. 3 is an explanatory diagram of a correlation degree inspection method of a correlation degree inspection unit in FIG. 2;

FIG. 4 is a configuration diagram of a modified example of an utterance detection unit.

[Explanation of symbols]

 Reference Signs List 1 head unit 8a to 8d speaker 10a to 10d microphone 14 utterance detection unit 29 level reduction unit

Continuation of the front page (72) Inventor Minoru Sekine 1-8-1, Nishigotanda, Shinagawa-ku, Tokyo Alpine Co., Ltd. (56) References JP-A-61-88607 (JP, A) (58) Fields investigated ( Int.Cl. ⁷ , DB name) H04R 3/04 G11B 20/04 101

Claims (2)

(57) [Claims] 1. A amplifies the audio signal obtained from the audio source, the vehicle audio apparatus to output from the speaker installed at a plurality of locations in the vehicle compartment, and utterance detection means for detecting an occupant another utterance, speech Level reduction means for reducing the level of part or all of the speech voice band of the audio signal related to the loudspeaker near the occupant near the occupant and inputting it to the loudspeaker; apparatus. 2. The in-vehicle audio device according to claim 1, wherein the level lowering means lowers a band level below the conversation voice band when lowering the level of the conversation voice band.