JP2011109487

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 JP2011109487
An object of the present invention is to suppress Doppler distortion due to a speaker included in
generated sound in a loudspeaker that is a combination of an amplifier and a speaker.
SOLUTION: A transfer characteristic from a drive signal to a sound pressure generated from a
speaker is modeled in advance including an effect of a cone displacement fluctuation of the
speaker, and a compensation device generates the cone displacement estimating means of the
speaker and the speaker By providing means for calculating a drive signal that can make the
sound pressure proportional to the sound input signal, it is possible to obtain a sound pressure
signal with Doppler distortion compensated. [Selected figure] Figure 1
Amplification apparatus having a Doppler distortion compensation function
[0001]
The present invention relates to a Doppler distortion compensator that pre-compensates, at the
amplifier stage, Doppler distortion that occurs in a speaker that emits sound.
[0002]
In a loudspeaker combined with an amplifier and a speaker, it is relatively easy to reduce
distortion generated in the amplifier.
However, in the speaker, in principle, distortion called Doppler distortion occurs, which causes
distortion in the sound emitted by the speaker. Although the Doppler distortion is often smaller
08-05-2019
1
than distortion due to other non-linear factors of the speaker, the influence on the sense of sound
is often not negligible.
[0003]
The Doppler distortion is distortion generated when the frequency of the sound emitted by the
Doppler effect due to the vibration of the diaphragm of the speaker deviates from the frequency
of the original signal. That is, it is distortion generated by frequency shift in proportion to the
moving speed of the diaphragm of the speaker.
[0004]
As an effective means for reducing this Doppler distortion, there is multi-waying of speakers. The
Doppler distortion is noticeable when low frequency components and high frequency
components are mixed. Therefore, it is not necessary to generate a low frequency signal and a
high frequency signal by one speaker by changing a speaker that generates sound for each
frequency band, and as a result, Doppler distortion can be suppressed to a low level. However,
even if multi-way is performed, usually there are not many frequency divisions such as 2 way or
3 way, and there remains a problem that it is difficult to sufficiently reduce Doppler distortion.
[0005]
For non-linear distortion of a speaker, Patent Document 1 proposes a method of compensating in
advance at the stage of an amplifier. However, in this method, in the case of performing
compensation using a PCM signal using digital processing, there is a problem that the accuracy
of compensation in a high frequency range is not good if the source signal contains a high
frequency component close to the Nyquist frequency. It was In addition, when trying to perform
Doppler distortion compensation for a multi-way speaker, there is also a problem that the
compensation accuracy near the crossover frequency is not good.
[0006]
Japanese Patent Application No. 2009-182570
08-05-2019
2
[0007]
The problem to be solved is to reduce the Doppler distortion generated by the system including
the speaker and the amplifier by previously compensating the Doppler distortion generated in
the speaker with high accuracy at the stage of the audio signal.
[0008]
In order to solve the above problems, a sound pressure estimation filter is provided that
estimates the sound pressure generated by the speaker including Doppler distortion, and
compensation is performed so that the sound pressure estimated by the sound pressure
estimation filter is proportional to the speech input signal By generating the filter output signal
and making the speaker drive signal proportional to the compensation filter output signal, the
sound pressure generated by the speaker is made proportional to the audio input signal to
compensate for Doppler distortion by the speaker.
[0009]
First, consider the case of driving a single speaker.
The same can be considered in the case of driving a plurality of equal speakers.
[0010]
The Doppler distortion can also be interpreted as the frequency of the sound observed by the
observer due to the speed of the speaker's cone displacement being shifted with respect to the
frequency of the signal that drives the speaker. It can also be interpreted that it is distortion that
occurs when the phase of the sound observed by a person changes.
These two interpretations are due to differences in thinking and are equivalent.
[0011]
Consider the system shown in FIG.
08-05-2019
3
Now, let the displacement of the cone of the speaker 3 be p (t). Where t is time. Also, the input
signal to the compensator 1 is u (t), the output signal of the compensator 1 is v (t), the sound
pressure observed by the observer is so (t), and v (t) for minute signals is Let G (s) be the
transmission to t). Also, let s (t) be the sound pressure observed by the observer when the
Doppler distortion is not included. That is, (Equation 1) S (s) = G (s) V (s). However, V (s) is a
Laplace transform of v (t), and S (s) is a Laplace transform of s (t). The displacement of the cone
of the speaker 3 is p (t), and the transfer function H (s) from v (t) to p (t) is to be experimentally
measured in advance. Then, p (t) can be estimated from v (t). At this time, the sound pressure
observed by the observer with respect to v (t) which is not minute is as follows: so (t) = s (t-p (t) /
c). Where c is the velocity of sound.
[0012]
Here, consider a case where the variable dead time element is used to set the following equation
(3) v (t) = u (t + τ-p (t) / c). Here, τ is a fixed dead time, and it is assumed that (Expression 4)
τ−p (t) / c> 0 is satisfied. Since the signal can be delayed in the time direction but can not be
advanced, a fixed dead time τ is provided. Then, the fluctuation of the dead time causing the
Doppler distortion is canceled out, and the sound pressure observed by the observer becomes
(Equation 5) So (s) = G (s) U (s) exp (-τs). However, So (s) is a Laplace transform of so (t).
[0013]
The variable dead time element (Equation 3) can be realized by an FIR filter or the like having
variable parameters, and such variable dead time element is used in a sample rate converter or
the like.
[0014]
Furthermore, as the variable dead time element, not only pure dead time as shown in (Equation
3) is changed, but also the gain of the frequency response is made to have the inverse
characteristic of G (s), and May be compensated.
[0015]
In this case, the variable dead time element can be designed as follows.
08-05-2019
4
First, G (s) is converted to state variable representation.
[0016]
[0017]
Then, discrete time is converted including the fluctuating dead time τ-p (t).
[0018]
[0019]
Here, T is a sampling period, and p [k] is a signal obtained by discretizing time of p (t) according
to the sampling period.
Also, v [k] is the discrete-time output signal of the compensator.
Then, y [k] is a signal that has advanced for a fixed time with respect to s0 (t).
Then, v [k] is calculated as follows.
[0020]
[0021]
Then, y [k] is a signal delayed by one sample with respect to u [k], and so (t) is a signal delayed
by a fixed time with respect to u [k].
08-05-2019
5
That is, the sound pressure observed by the observer can be one that does not include Doppler
distortion faithful to the audio input signal u [k].
[0022]
Next, consider a case where two types of speakers having different characteristics are driven by
one amplifier 2 (FIG. 1).
Here, two types of speakers are assumed to be the woofer 3a and the tweeter 3b.
[0023]
Now, let pw (t) be the displacement of the woofer 3a, and pt (t) be the displacement of the cone
of the tweeter 3b. Also, u (t) is the input signal to compensator 1, v (t) is the output signal of
compensator 1, so (t) is the sound pressure observed by the observer, and v (t) for minute signals
through the woofer Let G w (s) transmit to so (t), let G t (s) transmit from v (t) to so (t) through
the tweeter. Also, let sw (t) be the sound pressure through the woofer observed by the observer
in the case where the Doppler distortion is not included, and st (t) be the sound pressure through
the tweeter. すなわち、
[0024]
[0025]
とする。
However, V (s) is Laplace transform of v (t), and Sw (s) is Laplace transform of sw (t). Further, it is
assumed that transfer functions Hw (s) from v (t) to pw (t) and transfer functions Ht (s) from v (t)
to pt (t) are experimentally measured in advance. Then, pw (t) and pt (t) can be estimated from v
(t). At this time, the sound pressure observed by the observer for v (t) which is not minute is
08-05-2019
6
[0026]
[0027]
となる。
The sound pressure so (t) observed by the observer is calculated as v [k] so as to be a signal
delayed by a predetermined time with respect to the voice input signal u [k].
[0028]
First, Gw (s) and Gt (s) are converted to state variable representations.
[0029]
[0030]
[0031]
Then, discrete time is converted including the fluctuating dead time τ-p (t).
[0032]
[0033]
[0034]
Where T is a sampling period, pw [k] is a signal obtained by discrete-time converting pw (t)
according to the sampling period, and pt [k] is a signal obtained by discrete-time converting pt (t)
according to the sampling period .
08-05-2019
7
Also, v [k] is the discrete-time output signal of the compensator.
Then, y [k] in the following equation is a signal that has advanced for a fixed time with respect to
s0 (t).
[0035]
[0036]
Then, v [k] is calculated as follows.
[0037]
[0038]
Then, so (t) is a signal delayed for a fixed time with respect to u [k].
That is, the sound pressure observed by the observer can be one that does not include Doppler
distortion that is faithful to the audio input signal.
However, it is necessary that the compensator configured by Equations 13, 14, and 16 be stable.
That is, the signal v [k] needs not to diverge with time.
It is difficult to pinpoint the conditions for this compensator to be stable, but using this approach
will impose constraints on Gw (s) and Gt (s).
Although two types of speakers are used here, the woofer 3a and the tweeter 3b, the number of
types of speakers can be further increased by the same method.
08-05-2019
8
[0039]
The amplification apparatus having the Doppler distortion compensation function according to
the present invention has an advantage that the Doppler distortion can be compensated with
high accuracy even for a plurality of types of speakers (multi-way speakers).
[0040]
FIG. 14 is a block diagram showing the configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a fourth embodiment of the present invention.
FIG. 1 is a block diagram showing a configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a first embodiment of the present invention.
FIG. 5 is a block diagram showing another configuration of the amplifier and the speaker
including the Doppler distortion compensator according to the first embodiment of the present
invention.
FIG. 7 is a block diagram showing the configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a second embodiment of the present invention. FIG.
7 is a block diagram showing another configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a second embodiment of the present invention. FIG.
10 is a block diagram showing the configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a third embodiment of the present invention. FIG.
14 is a block diagram showing another configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a fourth embodiment of the present invention. FIG.
13 is a block diagram showing the configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a fifth embodiment of the present invention. FIG.
14 is a block diagram showing the configuration of an amplifier and a speaker including a
Doppler distortion compensator according to a sixth embodiment of the present invention.
[0041]
A first embodiment of the present invention is shown in FIG. The voice input signal u (t) is a
signal subjected to volume processing. The transfer function from the signal v (t) to the cone
08-05-2019
9
displacement of the speaker 3 is obtained in advance by an experimental method or the like, and
the transfer function is H (s). Then, the cone displacement of the speaker 3 can be estimated by
applying the transfer function H (s) to the signal v (t). The cone displacement of the speaker 3
estimated is p (t). The compensator 12 is constituted by the following FIR filter.
[0042]
[0043]
Here, the function f (t) is a previously designed time function, and a sinc function or the like can
be used.
N is an integer parameter that determines the order of the filter, and T is a sampling period.
Further, the signal u [k] is a signal obtained by sampling the audio input signal u (t), and the
signal v (t) is a signal obtained by passing the signal v [k] through zero-order hold. By properly
designing the function f (t), the compensator 12 operates as a dead time component within a
specific frequency range. Since the dead time in the compensator 12 is changed according to the
amount of change in the delay time generated by the cone displacement of the speaker 3, a
sound pressure signal with Doppler distortion compensated can be generated from the speaker 3
as a result. .
[0044]
In the first embodiment of the present invention, the cone displacement of the speaker 3 is
estimated based on the signal v (t). However, since the amount of compensation in the
compensator 12 is small, as shown in FIG. The cone displacement of the speaker 3 may be
estimated based on the input signal u (t).
[0045]
A second embodiment of the present invention is shown in FIG.
The voice input signal u [k] is a discrete time signal whose sampling period is T, and is a volumeprocessed signal. The transfer function from the signal v [k] to the cone displacement of the
08-05-2019
10
speaker 3 is previously obtained by an experimental method or the like, and the transfer function
is H [z]. Then, the cone displacement of the speaker 3 can be estimated by applying the transfer
function H [z] to the signal v [k]. The cone displacement of the speaker 3 estimated is p [k]. In
addition, a transfer function G (s) for a minute signal from the signal v [k] to the sound pressure
is also obtained in advance experimentally. The compensator 12 calculates a signal v [k] which is
an output signal in accordance with Equation 8. Further, the state variable vector x [k] in
equation 8 is updated every sample according to equation 7. At this time, since the dead time in
the compensator 12 is changed according to the amount of change in the delay time generated
by the cone displacement of the speaker 3, a sound pressure signal with Doppler distortion
compensated as a result is generated from the speaker 3. be able to. In the compensator 12,
since the frequency characteristics of the amplifier 2 and the speaker 3 are also compensated,
the sound pressure signal generated from the speaker 3 is not only the Doppler distortion in the
speaker 3 but also the frequency characteristics of the amplifier 2 and the speaker 3 It becomes
a compensated sound pressure signal.
[0046]
In the second embodiment of the present invention, the cone displacement of the speaker 3 is
estimated based on the signal v [k]. However, since the amount of compensation in the
compensator 12 is small, as shown in FIG. The cone displacement of the speaker 3 may be
estimated based on the input signal u [k]. At this time, it is assumed that the transfer function G
(s) for a minute signal from the signal v [k] to the sound pressure is normalized so as to be
approximately 1 in the main range.
[0047]
A third embodiment of the present invention is shown in FIG. This is to drive two types of
speakers 3a and 3b using two amplifiers 2a and 2b, and prepare compensators 1a and 1b for
each speaker 3a and 3b.
[0048]
The combination of the compensator 1a, the amplifier 2a, the speaker 3a and the combination of
the compensator 1b, the amplifier 2b and the speaker 3b are the same as those of the second
embodiment of the present invention. The cone displacement estimator 11c calculates the
08-05-2019
11
estimated value pw [k] of the cone position of the speaker 3a, and the compensation element 12a
is designed from the characteristic of the amplifier 2a and the speaker 3a combined. Further, the
cone displacement estimator 11d calculates the estimated value pt [k] of the cone position of the
speaker 3b, and the compensation element 12b is designed from the characteristic of the
combination of the amplifier 2b and the speaker 3b.
[0049]
In the third embodiment of the present invention, since Doppler distortion compensation is
performed independently for two types of speakers 3a and 3b, it has an advantage that there is
no restriction on the combination of two types of speakers 3a and 3b. ing.
[0050]
In the third embodiment of the present invention, the loudspeaker 3a and the loudspeaker 3b are
estimated using the signal vw [k] and the signal vt [k], but the compensators 1a and 1b in FIG.
The cone position estimated values pw [k] and pt [k] may be calculated based on the speech input
signal u [k] approximately by using the same configuration as that of the compensator 1.
[0051]
A fourth embodiment of the present invention is shown in FIG.
Although a frequency discrimination filter referred to as a network is not explicitly described in
FIG. 1, when a network is installed between the amplifier 2 and the speakers 3a and 3b, they are
included in the speakers 3a and 3b. It shall be interpreted.
The voice input signal u [k] is a discrete time signal whose sampling period is T, and is a volumeprocessed signal. The transfer function from the signal v [k] to the cone displacement of the
speakers 3a and 3b is previously obtained by an experimental method or the like, and the
transfer functions are Hw [z] and Ht [z]. Then, the cone displacement of the speaker 3a and the
speaker 3b can be estimated by applying the transfer function H [z] to the signal v [k]. The cone
displacement of the speaker 3 estimated is pw [k] and pt [k]. Also, transfer functions Gw (s) for
minute signals from the signal v [k] to the sound pressure through the speaker 3a and transfer
functions Gt (s) for the minute signal from the signal v [k] to the sound pressure through the
speaker 3b ) Are also determined in advance experimentally. The compensator 12 calculates a
signal v [k] which is an output signal according to Equation 16. Further, the state variable vectors
08-05-2019
12
xw [k] and xt [k] in Eq. 16 are updated every sample according to Eq. 13 and Eq. As a result, the
sound pressure signal generated from the speakers 3a and 3b is a sound pressure signal in which
not only the Doppler distortion in the speakers 3a and 3b but also the frequency characteristics
of the amplifier 2 and the speakers 3a and 3b are compensated.
[0052]
In the fourth embodiment of the present invention, the two-way speaker by the speakers 3a and
3b is driven by one amplifier 2. However, a speaker with more ways may be driven by one
amplifier 2. In that case, the signal v [k] is calculated so as to make the sound input signal u [k]
follow the overall sound pressure generated by all the speakers.
[0053]
In the fourth embodiment of the present invention, the cone displacement of the speakers 3a and
3b is estimated based on the signal v [k]. However, since the compensation amount in the
compensator 12 is small, as shown in FIG. The cone displacement of the speaker 3 may be
estimated based on the audio input signal u [k]. At this time, it is assumed that the transfer
function Gw (s) + Gt (s) for a minute signal from the signal v [k] to the sound pressure is
normalized so as to be approximately 1 in the main range.
[0054]
A fifth embodiment of the present invention is shown in FIG. This is a system which drives a 3
Way speaker by 3 types of speakers 3a, 3b and 3c by two amplifiers 2a and 2b. Although filters
called networks are not explicitly shown in FIG. 8, when networks are installed, they are included
in the speakers 3a, 3b and 3c.
[0055]
The speakers 3a and 3b are driven by the method shown in the fourth embodiment of the
present invention. Therefore, the Doppler distortion generated by the speakers 3a and 3b is
compensated by the compensator 1a. Further, the speaker 3c is driven by the method shown in
08-05-2019
13
the second embodiment of the present invention. Therefore, the Doppler distortion generated by
the speaker 3c is compensated by the compensator 1b. As a result, all Doppler distortion
generated by the speakers 3a, 3b and 3c will be compensated.
[0056]
A sixth embodiment of the present invention is shown in FIG. This is a system which drives a 3
Way speaker by 3 types of speakers 3a, 3b and 3c by two amplifiers 2a and 2b. Although filters
called networks are not explicitly shown in FIG. 8, when networks are installed, they are included
in the speakers 3a, 3b and 3c.
[0057]
The speakers 3a and 3b are driven by the method shown in the fourth embodiment of the
present invention. Therefore, the Doppler distortion generated by the speakers 3a and 3b is
compensated by the compensator 1a. As for the speaker 3c, Doppler distortion is compensated
and driven by a method other than the method shown in the second embodiment of the present
invention. For example, Doppler distortion is compensated in the compensator 4 by the method
described in Japanese Patent Application No. 2009-182570, and the speaker 3c is driven.
Therefore, the Doppler distortion generated by the speaker 3 c is compensated by the
compensator 4. As a result, all Doppler distortion generated by the speakers 3a, 3b and 3c will be
compensated.
[0058]
The method described in the fourth embodiment of the present invention has the advantage of
being able to accurately compensate for the Doppler distortion generated by a plurality of
speakers, but it is possible to compensate for the speaker's cone displacement that can be
compensated from the sampling period T of signal processing. The amplitude is limited. In the
case of the sixth embodiment of the present invention, assuming that the speaker 3a is a
squawker and the speaker 3b is a tweeter, both the squawker and the tweeter have small cone
displacement amplitudes, so the present invention fourth embodiment The method shown in the
embodiment can be used without any problem. On the other hand, when the speaker 3c is a
woofer, the amplitude of the cone displacement of the woofer is large in many cases, so that the
sampling period T of the signal processing can not be made very short. On the other hand, the
08-05-2019
14
method described in Japanese Patent Application No. 2009-182570 has a problem that the
accuracy of Doppler distortion compensation near the crossover frequency is poor when there
are a plurality of types of speakers. There is no. Moreover, since the Doppler distortion
compensated by the compensator 4 is a single type of loudspeaker, the problem of the
compensation accuracy near the crossover frequency with the method described in Japanese
Patent Application No. 2009-182570 does not occur.
[0059]
By using the Doppler distortion compensator according to the present invention, it is possible to
realize a loudspeaker that generates a sound in which the Doppler distortion generated in the
speaker has been canceled.
[0060]
1, 1a, 1b: distortion compensation device 11, 11a, 11b, 11c, 11d: cone displacement estimator
12, 12a, 12b, 12c: compensator 2, 2a, 2b: amplifier 3, 3a, 3b, 3c ... speaker 4 ... distortion
compensation device
08-05-2019
15