close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2006173840

код для вставкиСкачать
Patent Translate
Powered by EPO and Google
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2006173840
[PROBLEMS] To provide an “audio output device” that suppresses the deterioration of
loudness compensation due to the variation of the low-frequency characteristic of a microphone.
SOLUTION: The power of a portion excluding the low-pass portion of the guidance voice signal
component included in the signal output from the microphone 7 is estimated by an output voice
power calculation unit 15 and a second high pass filter 14 to be an output voice power. . The
adder 10 cancels the output sound power from the power of the portion excluding the low band
of the signal output from the microphone 7 calculated by the first high-pass filter 8 and the input
sound power calculation unit 9 to reduce the ambient noise Calculate the power of the part
excluding the low band. The low frequency power component adder 16 adds the power of low
frequency components of the ambient noise estimated from the vehicle speed V to the power of
the ambient noise, and outputs the result as noise power. The loudness compensation control
unit 11 adjusts the gain of the volume adjustment amplifier 3 according to a predetermined
loudness compensation function based on the output sound power and the noise power. [Selected
figure] Figure 1
Voice output device
[0001]
The present invention relates to a technology for improving the intelligibility of voice heard by a
user in a voice output device that outputs voice.
[0002]
In a voice output device that outputs voice, as a technique for improving the intelligibility of
voice heard by the user, a voice signal mixed with voice and ambient noise output from the
10-04-2019
1
speaker of the voice output device is picked up using a microphone, and The gain of the sound
output from the audio output device is adjusted for each frequency band according to human
loudness characteristics according to the power of the ambient noise component in the output
audio signal and the power of the audio component output from the audio output device
Techniques for loudness compensation are known (e.g., Patent Document 1).
[0003]
In this technique, an adaptive filter for learning a transfer function of an audio signal output from
the audio output device to the speaker to the microphone from the audio signal output from the
audio output device to the speaker and the audio signal output from the microphone is used. It is
provided.
Then, using this adaptive filter, an audio component output from the speaker included in the
audio signal output from the microphone is estimated from the audio signal output from the
audio output device to the speaker.
And the component which remove | eliminated the estimated audio | voice component from the
audio | voice signal output from a microphone is estimated as an ambient noise component
contained in the audio | voice signal output from a microphone. JP 2001-236090 A
[0004]
Now, in the voice output device as described above, a general-purpose small-sized microphone
for collecting human voice is often used as a microphone for picking up a voice signal. The
frequency characteristics of such a general-purpose small-sized microphone for collecting voices
that are lower than the lower frequency band of the voice frequency band are the same products
of the same manufacturer (products of the same model number) There are often variations.
[0005]
And, for this reason, when using a microphone that can not uniquely determine such low
frequency characteristics, estimation of the speaker output sound component and the ambient
10-04-2019
2
noise component in the sound signal output from the microphone is performed. Sometimes it can
not be done correctly, and loudness compensation can not be done properly. On the other hand,
in order to cope with such a problem, the low frequency component of the audio signal output
from the microphone is cut to estimate the speaker output audio component and the ambient
noise component in the audio signal output from the microphone Although it is conceivable to do
so, for example, in an environment such as a car, since ambient noise contains many low
frequency components, there are cases where loudness compensation can not be properly
performed.
[0006]
Then, this invention makes it a subject to suppress deterioration of the loudness compensation
by the dispersion | variation in the frequency characteristic of the low-pass of a microphone in
the audio | voice output apparatus which outputs an audio | voice.
[0007]
In order to achieve the above object, the present invention provides an audio output device for
outputting an audio subjected to loudness compensation, an audio signal generation means for
generating an output audio signal which is an audio signal representing the audio to be output,
and an audio signal generation Loudness compensation means for performing loudness
compensation of an output voice signal generated by the means, a speaker for outputting voice
based on the output voice signal for which the loudness compensation means has performed
loudness compensation, a microphone for picking up ambient sound, and the microphone Output
voice component power calculation means for calculating the power of the component of the
output voice in the microphone output signal based on the microphone output signal which is a
signal to be output and the output voice signal, and the microphone output signal of the
microphone Power and the power of the component of the output voice calculated by the output
voice component power calculating means, Ambient noise component power calculation means
for calculating the power of the ambient noise component in the output signal as having no
power of a predetermined frequency region on the low frequency side of the ambient sound
component, and the ambient noise component power calculation means And noise power
calculating means for calculating, as noise power, power obtained by adding predetermined
power to the power of the ambient noise component, and the output calculated by the output
sound component power calculating means in the loudness compensation means The gain
adjustment of the output sound signal to be output to the speaker is performed according to a
predetermined loudness compensation function based on the power of the sound component and
the noise power calculated by the noise power calculating means.
[0008]
10-04-2019
3
According to such an audio output device, the predetermined frequency region on the low
frequency side is a frequency region where variation occurs in the frequency characteristic of the
microphone, and the predetermined power to be added by the noise power calculation means is
the ambient sound whose generation is estimated On the other hand, when the ambient sound is
generated, if it is generated in the microphone output signal by setting the power of the
frequency domain of the ambient sound component assumed in the loudness compensation
means, it does not depend on the frequency characteristics of the individual microphones. The
noise power can be calculated so that appropriate loudness compensation is performed.
[0009]
However, in such an audio output device, in the ambient sound component power calculation
unit, the power of the microphone output signal output by the microphone and the power of the
component of the output sound calculated by the output sound component power calculation
unit Therefore, the power of the ambient noise component in the microphone output signal is
replaced with the power of the predetermined frequency range on the low frequency side of the
ambient noise component, and the predetermined frequency of the low frequency side of the
ambient noise component The calculation may be performed using the power obtained by
attenuating the power of the region.
[0010]
By doing this, the predetermined frequency region on the low frequency side is made a frequency
region where variation occurs in the frequency characteristics of the microphone, and
predetermined power added by the noise power calculation means is applied to the ambient
sound whose generation is estimated. By setting the noise power to be equal to the power of the
ambient sound component assumed in the loudness compensation means when the ambient
sound is generated when it is generated in the microphone output signal, the frequency
characteristics of the individual microphones for loudness compensation can be obtained. The
noise power can be calculated taking into consideration the actual ambient sound generation
condition while reducing the dependency.
Therefore, loudness compensation can be performed according to the actual situation of ambient
sound generation.
[0011]
10-04-2019
4
Further, in the above sound output device, the noise power calculation means acquires the
vehicle speed of the vehicle, and the power determined for the acquired vehicle speed is the
power of the ambient noise component calculated by the ambient noise component power
calculation means. The added power may be calculated as noise power.
Alternatively, in the noise power calculating means, a power determined by adding the power of
the ambient noise component calculated by the ambient noise component power calculating
means to the power of the ambient noise component calculated by the ambient noise component
power calculating means May be calculated as noise power.
[0012]
By doing as described above, the ambient sound is generated for the ambient sound whose
occurrence is estimated under the situation according to the situation where each of the
predetermined powers added by the noise power calculation means is When it occurs in the
microphone output signal, the noise power can be set to the power of the ambient sound
component assumed in the loudness compensation means.
Therefore, noise power can be calculated so that more accurate loudness compensation can be
performed.
[0013]
As described above, according to the present invention, it is possible to suppress the degradation
of loudness compensation due to the variation of the frequency characteristic of the low
frequency band of the microphone in the audio output device that outputs voice.
[0014]
Hereinafter, an embodiment of a voice output device according to the present invention will be
described by taking an application to an on-vehicle device mounted on a car as an example.
FIG. 1 shows the configuration of the in-vehicle apparatus.
10-04-2019
5
As shown in the figure, the on-vehicle apparatus provides the user with a navigation device 1 for
providing a route guidance to the current position and a destination, an audio device 2 such as a
radio receiver or a CD player, a volume control amplifier 3 and an output voice adder 4. , Output
amplifier 5, speaker 6, microphone 7, first high pass filter 8, input sound power calculation unit
9, adder 10, loudness compensation control unit 11, filter 12, coefficient update unit 13, second
high pass filter 14, output voice A power calculation unit 15 and a low frequency power
component addition unit 16 are provided. Here, the navigation device 1 outputs to the volume
adjustment amplifier 3 a guidance voice signal representing a guidance voice for guiding the user
to a route or operation.
[0015]
In such a configuration, the guidance voice signal output from the navigation device 1 is input to
the output voice adder 4 after the overall gain is adjusted by the volume adjustment amplifier 3.
The output audio adder 4 adds the guidance audio signal input from the volume adjustment
amplifier 3 and the audio signal input from the audio device 2 and outputs the result as an
output audio signal to the output amplifier 5. Then, the output amplifier 5 amplifies the output
audio signal to be input, and outputs the amplified output audio signal from the speaker 6. Then,
the output sound (synthesized sound of the guidance voice S and the audio sound A) output from
the speaker 6 and the ambient noise N such as car noise are picked up by the microphone 7 and
input to the first high pass filter 8 as an electric signal. . Here, for convenience of description, the
signal output from the microphone 7 is hereinafter referred to as a "microphone output signal".
The microphone output signal is input to the input sound power calculation unit 9 after the low
frequency component is attenuated by the first high pass filter 8. The input sound power
calculation unit 9 calculates the power per unit time of the microphone output signal whose low
frequency component has been attenuated, and inputs the power to the adder 10.
[0016]
On the other hand, the filter 12 is a filter that simulates the impulse response (transfer function)
of the transfer system of the guidance voice signal from the output of the navigation device 1 to
the output of the microphone 7 and calculates and updates the impulse response of the filter 12
Is the coefficient updating unit 13, and the filter 12 and the coefficient updating unit 13
constitute an adaptive filter. Then, the second high pass filter 14 attenuates the low frequency
component of the audio signal output from the filter 12 and outputs it to the output audio power
10-04-2019
6
calculation unit 15. Here, the frequency characteristics of the first high pass filter 8 and the
second high pass filter 14 are set equal.
[0017]
Next, the output voice power calculation unit 15 calculates the power per unit time of the voice
signal whose low-pass component output from the second high-pass filter 14 is attenuated, and
the coefficient update unit 13 calculates LMS (mean square error) The filter of the filter 12 so
that the difference between the output of the input sound power calculator 9 output from the
adder 10 and the output of the output voice power calculator 15 is minimized by an algorithm or
an NLMS (normalized mean square error) algorithm, etc. The process of calculating the
coefficients and setting them in the filter 12 is repeated.
[0018]
Here, when the impulse response of the filter 12 accurately simulates the impulse response of
the transfer system of the guidance voice signal from the output of the navigation device 1 to the
output of the microphone 7, the output of the filter 12 is an output of the microphone 7 The
output of the output voice power calculation unit 15 excludes the low-pass component
attenuated by the second high-pass filter 14 in the guide voice signal component in the signal
output from the microphone 7 It becomes the power per unit time of the part.
Further, the output of the adder 10 is a low pass attenuated by the first high pass filter 8 of an
ambient noise component obtained by combining an audio signal component and an ambient
noise component such as a car noise in a signal output from the microphone 7 It becomes the
power of the part excluding the component.
[0019]
Now, the output voice power calculation unit 15 sets the power per unit time of the portion
excluding the low-pass component attenuated by the second high pass filter 14 of the guidance
voice signal component output from the filter 12 as the output voice power as the loudness Also
output to the compensation control unit 11. On the other hand, the power per unit time of the
portion excluding the low band component attenuated by the first high pass filter 8 of the
ambient noise component output from the adder 10 is output to the low band power component
adder 16. The low frequency power component adder 16 determines the ambient noise low
10-04-2019
7
frequency component power by low frequency power addition processing described later, adds
the determined ambient noise low frequency component power to the power output from the
adder 10, and outputs it as noise power. It is output to the loudness compensation control unit
11.
[0020]
Then, based on the input output voice power and noise power, the loudness compensation
control unit 11 adjusts the gain of the volume adjustment amplifier 3 according to a
predetermined loudness compensation function so that the user can clearly hear the guidance
voice. However, separately from the output voice power calculation unit 15, the power per unit
time of the guidance voice signal component before the low-pass component attenuation is
performed by the second high-pass filter 14 output from the filter 12 is the second output voice
power In the loudness compensation control unit 11, based on the second output sound power
and the noise power, the gain of the volume adjustment amplifier 3 is calculated according to a
predetermined loudness compensation function. May be adjusted to be clearly audible to the
user.
[0021]
Here, the frequency characteristics of the microphone 7 and the frequency characteristics of the
first high pass filter 8 and the second high pass filter 14 will be described. Now, it is assumed
that the product used as the microphone 7 exhibits the frequency characteristics shown in FIG.
2a and the frequency characteristics shown in FIG. 2b by the individual microphones 7. That is,
this product is a product having a frequency characteristic of a frequency band higher than a
predetermined frequency fco (for example, 250 Hz) approximately equal, but showing a
frequency characteristic different for each individual microphone in a frequency band below
frequency fco, ie, frequency fco It is assumed that the product has non-negligible variation in the
frequency characteristics of the following frequency bands.
[0022]
In such a case, as shown in FIG. 2c, the frequency characteristics of the first high pass filter 8 and
the second high pass filter 14 are components of the frequency band below the upper limit fco of
the frequency band in which the frequency of the microphone 7 has variation. Attenuate greatly
10-04-2019
8
and cut. That is, as the first high-pass filter 8 and the second high-pass filter 14, one having a
frequency characteristic that passes only the component of the frequency band exceeding the
upper limit fco of the frequency band without substantially attenuating is used.
[0023]
Next, low-pass power addition processing performed by the low-pass power component addition
unit 16 will be described. The low speed power component addition unit 16 receives the vehicle
speed signal V of the vehicle from the ECU or the like of the vehicle. Further, in the low
frequency power component addition unit 16, the correspondence between the vehicle speed
and the value of the ambient noise low frequency component power to be added to the power
output from the adder 10 is set in advance. Then, the low frequency power component addition
unit 16 obtains the value of the ambient noise low frequency component power for which the
correspondence with the vehicle speed indicated by the input vehicle speed signal V is set, and
adds the obtained ambient noise low frequency component power value It adds to the power
which the unit 10 output, and outputs to the loudness compensation control part 11 as noise
power.
[0024]
Here, the correspondence between the vehicle speed and the value of the ambient noise lowfrequency component power to be added to the power output from the adder 10 is determined in
advance as follows, and set in the low-frequency power component adder 16. That is, as shown in
FIG. 2d, the power frequency distribution per unit time of the car noise at each vehicle speed V
(a, b, c: a> b> C in the figure) is determined by measurement or calculation. Here, this
measurement or calculation is performed on the premise of using the microphone 7 having a
specific frequency characteristic of the frequency characteristics that can be acquired by the
microphone. Further, the loudness compensation function set in the loudness compensation
control unit 11 is set such that the optimum loudness compensation is performed when the
microphone 7 having a frequency characteristic equal to this specific frequency characteristic is
used.
[0025]
Now, for each vehicle speed V, of the power per unit time of the car noise obtained for the
10-04-2019
9
vehicle speed V, the frequency of the frequency of the microphone 7 shown in FIG. The
component of is extracted and the power of the whole extracted component is calculated. Then,
the value of the power of all components below the frequency below fco obtained is set in the
low frequency power component adder 16 as the value of the ambient noise low frequency
component power corresponding to the vehicle speed V for which the power is obtained.
[0026]
The embodiments of the present invention have been described above. By the way, in the above
embodiment, according to the vehicle speed, the value of the ambient noise low-frequency
component power added to the power output from the adder 10 is changed by the low-frequency
power component adder 16. The value of the ambient noise low-frequency component power to
be added by the component addition unit 16 may be a fixed value. In this case, the frequency
characteristic of the power per unit time of the car noise is determined in advance with respect to
the average vehicle speed V, and among the power per unit time of the determined car noise,
there is a frequency with the frequency variation of the microphone 7 The component below the
frequency below the upper limit fco of a zone is extracted, and the power of the whole extracted
component is computed. Then, the value of the calculated power is set in the low band power
component adder 16 as the value of the ambient noise low band component power to be added
by the low band power component adder 16.
[0027]
As described above, according to the present embodiment, the ambient sound component relative
to the ambient sound estimated as the power of the frequency domain component below fco of
the ambient sound component that varies depending on the frequency characteristics of the
individual microphones 7 The noise power is determined using the power predetermined as the
power to be generated as the power of the component of the frequency range lower than fco, so
that appropriate loudness compensation is performed without depending on the frequency
characteristics of the individual microphones 7 be able to. In addition, by changing the ambient
noise low-frequency component power added to the power output from the adder 10 according
to the vehicle speed, the noise power can be calculated and the loudness compensation can be
performed according to the actual ambient noise generation state. Become.
[0028]
10-04-2019
10
Now, as the frequency characteristics of the first high pass filter 8 and the second high pass filter
14 used in the above embodiment, for example, as shown in FIG. 3a, the frequency variation of
the frequency of the microphone 7 is not more than the upper limit fco of the frequency band. As
the components below the frequency, a frequency characteristic may be used in which the
amount of attenuation gradually increases toward the low frequency side. In this case, the value
of the ambient noise low frequency component power set in the low frequency power component
addition unit 16 in correspondence with each vehicle speed is the unit time of the car noise
obtained for the vehicle speed V shown in FIG. As the inverse of the attenuation characteristic of
the frequency function of the first high-pass filter 8 and the second high-pass filter 14 for
components below the upper limit fco of the frequency band where there is variation of the
frequency of the microphone 7 among the per-power It is possible to set the value of the total
power of the components below the frequency below fco shown in FIG.
[0029]
By doing this, it is possible to calculate the noise power that reflects the power of the frequency
component below the fco of the actual ambient sound component while reducing the dependence
on the frequency characteristics of the individual microphones 7. More loudness compensation
can be performed according to the actual situation of ambient sound generation. Further, in the
above embodiment, the loudness compensation is performed only by the overall gain adjustment
of the guidance voice in the volume adjustment amplifier 3, but this is performed for each
frequency band of the guidance voice signal output to the speaker 6. An equalizer may be
provided to perform gain adjustment, and loudness compensation may also be performed to
perform gain adjustment for each frequency band. In this case, the input sound power calculation
unit 9 and the output sound power calculation unit calculate the power for each frequency band.
Also, in this case, the low frequency power component adder 16 adds power for each frequency
band below the upper limit fco of the frequency band in which there is variation in the frequency
of the microphone 7.
[0030]
The power of each frequency band below fco added by the low-pass power component adder 16
is fco of the power per unit time of the car noise obtained for each vehicle speed V shown in FIG.
2e and FIG. 3b. The low frequency power component adder 16 is previously set according to the
power distribution of the following frequency components.
[0031]
10-04-2019
11
However, when performing loudness compensation that also performs gain adjustment for each
frequency band as described above, the power of each frequency band below fco to be added by
the low-frequency power component addition unit 16 may be determined as follows. Good.
That is, as shown in FIG. 3c, the power distributions 311 and 312 in which the offset OS is given
to the power distribution 300 of the frequency band below fco set in advance as shown in FIG.
The offset OS is determined so as to be continuous with the power distribution 301, 302 of the
frequency band exceeding f. Then, according to the power distribution 311, 312 obtained by
adding the offset OS obtained to the power distribution 300 shown in FIG. In this way, when
performing loudness compensation including gain adjustment for each frequency band, it is
possible to suppress discontinuous gain adjustment being performed near the frequency fco and
outputting an unnatural guidance voice. it can. Even when the loudness compensation is
performed only by the overall gain adjustment as described above, the power to be added by the
low-pass power component adder 16 is determined according to the power distributions 311 and
312 obtained in this manner. You may
[0032]
By doing this, noise power calculation and loudness compensation can be performed according to
the actual situation of ambient sound generation. In the above embodiment, the loudness
compensation is performed on the guidance voice output from the navigation device 1. However,
in the technology for voice clarification according to the present embodiment, the voice clarity
output from any device is clarified. Applicable for
[0033]
It is a block diagram showing composition of an in-vehicle device concerning an embodiment of
the present invention. It is a figure which shows the process which the audio | voice output
device concerning embodiment of this invention performs. It is a figure which shows the process
which the audio | voice output device concerning embodiment of this invention performs.
Explanation of sign
10-04-2019
12
[0034]
DESCRIPTION OF SYMBOLS 1 ... Navigation apparatus, 2 ... Audio apparatus, 3 ... Volume
adjustment amplifier, 4 ... Output audio adder, 5 ... Output amplifier, 6 ... Speaker, 7 ...
Microphone, 8 ... 1st high pass filter, 9 ... Input sound power calculation part , 10: Adder, 11:
Loudness compensation control unit, 12: Filter, 13: Coefficient updating unit, 14: Second high
pass filter, 15: Output voice power calculation unit, 16: Low frequency power component
addition unit.
10-04-2019
13
Документ
Категория
Без категории
Просмотров
0
Размер файла
23 Кб
Теги
description, jp2006173840
1/--страниц
Пожаловаться на содержимое документа