close

Вход

Забыли?

вход по аккаунту

?

JP2005260744

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 JP2005260744
PROBLEM TO BE SOLVED: To provide an in-phase method and an in-phase device for
microphone reception signals in a microphone array, which can make the reception signals of
respective microphones constituting the microphone array in-phase with high accuracy. .
SOLUTION: The in-phase device for the microphone sound reception signal in a microphone
array having three or more microphones mounted on a portable device and arranged on the xy
plane of a three-dimensional coordinate system based on the portable device is provided The first
means for calculating the coordinate estimation value of the sound source in the threedimensional coordinate system and the sound reception signal of each microphone are in-phased
based on the coordinate estimation value of the sound source calculated by the first means every
time. A second means for converting to a signal which is estimated using the natural gradient
learning method to estimate the coordinates of the sound source such that the difference
between the inphase signals obtained by the second means is minimized It is [Selected figure]
Figure 2
Method and apparatus for in-phase acquisition of microphone reception signal in microphone
array
[0001]
The present invention relates to an in-phase method and an in-phase device for microphone
reception signals in a microphone array.
[0002]
The inventors of the present application have been studying the purpose of high-quality sound
reception of remote speech in a situation where the spatial position of the microphone and the
04-05-2019
1
direction of arrival of the sound source change from moment to moment.
[0003]
The inventors of the present application have proposed a method of estimating the direction of
the microphone array with respect to the direction of arrival of the sound source and correcting
the time difference between the sound reception signals (see reference [1]).
This method adaptively estimates rotation angles of two coordinate axes representing the spatial
position of the microphone based on the minimization of one evaluation function.
[0004]
Reference [1]: Horiuchi, Mizumachi, Nakamura, "Sequential time difference correction algorithm
for multi-microphone sound reception signals," Journal of sound lectures, Vol.
I, pp. 691-692, Mar. 2003.
[0005]
Most of the conventional direction estimation methods (see reference [2]) up to that point
perform estimation in frame units, but in the above proposed method proposed by the inventors
of the present application, estimation is performed for each sample. There is. This is one of the
features of the above proposed method.
[0006]
Reference [2]: Oga, Yamazaki, Kanada, "Sound system and digital processing," The Institute of
Electronics, Information and Communication Engineers, 1995.
[0007]
However, the above proposed method proposed by the inventors of the present application uses
04-05-2019
2
the LMS algorithm, and the stability and tracking speed largely depend on the step size
parameter, and in general, the expert knowledge in its selection is necessary.
[0008]
By the way, in reference [3], Amari et al. Considered the reason why the steepest descent method
(probability descent method) using a gradient on space with singular points does not give the
optimum gradient direction, and used the Riemann metric. We propose a natural gradient
learning method.
[0009]
Reference [3]: Amari, "Natural Gradient Works Efficiently in Learning," Neural Computation, vol.
10, pp.
251-276, 1998.
Horiuchi, Mizumachi, Nakamura, "Sequential time difference correction algorithm for multimicrophone sound reception signals," "Ongaku Shu, Vol.
I, pp. 691-692, Mar. 2003. Ohga, Yamazaki, Kanada, “Acoustic system and digital processing,”
The Institute of Electronics, Information and Communication Engineers, 1995. Amari, “Natural
Gradient Works Efficiently in Learning,” Neural Computation, vol. 10, pp. 251-276, 1998.
[0010]
An object of the present invention is to provide an in-phase method and an in-phase device for
the microphone sound reception signal in the microphone array which can make the sound
reception signal of each microphone constituting the microphone array in phase with high
accuracy.
[0011]
The invention according to claim 1 is an in-phase arrangement of microphone reception signals
in a microphone array having three or more microphones mounted on a portable device and
disposed on the xy plane of a three-dimensional coordinate system based on the portable device.
04-05-2019
3
The first device for calculating the estimated value of the coordinate of the sound source in the
three-dimensional coordinate system for each predetermined time in the digitizing device, and
the sound reception of each microphone based on the estimated coordinate of the sound source
calculated by the first means A second means for converting the signal into an in-phase signal,
the first means utilizing a natural gradient learning method such that the difference in the inphase signals obtained by the second means is minimized It is characterized in that it estimates
the coordinates of the sound source.
[0012]
The invention according to claim 2 is an in-phase arrangement of microphone reception signals
in a microphone array having three or more microphones mounted on a portable device and
disposed on the xy plane of a three-dimensional coordinate system based on the portable device.
In the method, the first step of calculating the coordinate estimation value of the sound source in
the three-dimensional coordinate system at predetermined time intervals, and the sound
receiving coordinate estimation value of the sound source calculated in the first step A second
step of converting the signal into an in-phased signal, the first step utilizing a natural gradient
learning method such that the difference in the in-phased signal obtained by the second step is
minimized It is characterized in that it estimates the coordinates of the sound source.
[0013]
According to the present invention, the sound reception signals of the microphones constituting
the microphone array can be made in phase with high accuracy.
[0014]
Hereinafter, embodiments of the present invention will be described with reference to the
drawings.
[0015]
[1] Definition of Microphone Array Coordinate System FIG. 1 shows a microphone array
coordinate system.
[0016]
The microphone array is mounted on a portable device.
04-05-2019
4
Three or more microphones Mi (i = 1, 2,..., M) are disposed at arbitrary positions on the xy plane
of a three-dimensional space (three-dimensional coordinate system) based on the portable device.
The coordinates (polar coordinates) (ri, π / 2, φi) of each microphone Mi are given in advance.
[0017]
Further, the coordinates (polar coordinates) of the sound source S are taken as (R, θ, φ).
Here, it is assumed that the sound source is single.
[0018]
The arrival time τi from the sound source S to the microphone Mi can be expressed as in the
following equation (1) using the distance d (S, Mi) between the sound source S and the
microphone Mi.
[0019]
[0020]
Where c is the velocity of sound.
[0021]
Further, using this arrival time τi, the sound reception signal xi (t) of each microphone Mi can
be expressed as the following equation (2).
[0022]
[0023]
Here, s (t) represents a source signal, and ni (t) represents observation noise.
04-05-2019
5
[0024]
[2] Description of in-phase circuit for making microphone sound reception signal in phase in
microphone array Fig. 2 shows an in-phase circuit for making microphone sound reception signal
in phase in the microphone array.
[0025]
The sound reception signal xi (t) of each microphone Mi is sent to the in-phase circuit 10.
The in-phase circuit 10 corrects the sound reception signal xi (t) so that the delay time difference
between the respective sound reception signals xi (t) becomes zero.
That is, the in-phase circuit 10 makes the respective received sound signals xi (t) in phase.
Then, the signal in phase is output as yi (t).
Hereinafter, the process performed by the in-phase circuit 10 will be described.
[0026]
First, it is considered to make the reception signal xi (t) in phase.
Using the Sinc function, the in-phased signal yi (t) can be expressed as the following equation (3).
[0027]
[0028]
04-05-2019
6
However, it is given by Sinc (x) = sin (πx) / πx.
N is a filter length, D is a fixed delay, and T is a sampling interval.
Further, ^ τi is an arrival time τi calculated from estimated coordinates (^ R, ^ θ, ^ φ) of the
sound source S described later.
[0029]
For this yi (t), the squared error E as shown in the following equation (4) is defined.
[0030]
[0031]
The estimated coordinates (^ R, ^ θ, ^ φ) of the sound source S are given by the following
equation (5) as R, θ, and φ when the square error E is the smallest.
[0032]
[0033]
A natural gradient learning method is introduced into the minimization of the squared error E.
The estimated coordinates (^ R, ^ θ, ^ φ) of the sound source S are defined as in the following
equation (6).
[0034]
[0035]
04-05-2019
7
Here, T represents transposition.
[0036]
The minute change amount dω of the coordinates of the sound source S is given by the following
equation (7).
[0037]
[0038]
Using the natural gradient learning method, the (n + 1) -th estimated coordinate ω n + 1 is given
by the following equation (8) using the n-th estimated coordinate ω n.
[0039]
[0040]
Here, G <-1> is given by the following equation (9).
[0041]
[0042]
In the above equation (8), ∇ (nabula) is a differential operator.
G <−1> ∇E (ωn) in the above equation (8) corresponds to a minute change amount of the
coordinates of the sound source S.
[0043]
The in-phase circuit 10 first calculates the estimated coordinates (^ R, ^ θ, ^ φ) of the sound
source S of this time based on the above equation (8).
04-05-2019
8
Next, based on the estimated coordinates (^ R, ^ θ, ^ φ) of the obtained sound source S and the
coordinates (ri, π / 2, φi) of each microphone Mi, the sound source S is obtained from the above
equation (1) The arrival time τi to the microphone Mi is calculated from
Then, based on the calculated arrival time τi, the signal yi (t) which is in-phased is calculated
from the above equation (3).
[0044]
[3] Verification of Basic Performance The basic performance of the estimation algorithm of the
position of the sound source S used in the above embodiment is verified by computer simulation.
[0045]
First, several conditions were set when performing computer simulation.
The microphone has coordinates (−d, 3 <1⁄2> d / 3, 0) and (d, 3 <1⁄2> d / 3, 0) (−d, 2 × 3 <0)
on the xy plane. Three were arranged at 1/2> d / 3, 0).
ただし、dは0.05mである。
[0046]
As the target source signal, shaped gaussian noise with an average of 0, a variance of 0.05, and a
band of 125 Hz to 6 kHz was used.
As the noise signal, random band noise in the same band as the target source signal is used, and
the correlation between the channels and the target source signal is uncorrelated.
In addition, the signal used was one having a sampling frequency of 48 kHz and 16 bits
04-05-2019
9
quantized.
[0047]
The target sound source signal assumes that θ is 45 deg in a fixed direction, and φ arrives from
30 deg and -30 deg alternately, and the sound receiving signal xi (t) at each microphone Mi (i =
1, 2, 3) Was created by giving appropriate time shifts to the target sound source signal and the
noise signal on a computer and adding them.
[0048]
The initial values of θ and φ are 0 deg, and the initial value of R is 1.
Moreover, it compared with the conventional method (method shown by the reference [1]) based
on LMS algorithm.
The step size parameter μ in the conventional method is 0.01.
[0049]
FIG. 3 shows the result of direction estimation in the case where the SNR of the target source
signal with respect to the noise signal is 20 dB.
The lower part of FIG. 3 shows a part of the upper part of FIG. 3 in an enlarged manner.
Here, the horizontal axis represents time, and the vertical axis represents the estimated direction
of arrival.
[0050]
It can be confirmed from FIG. 3 that in the method (proposed algorithm) according to the present
04-05-2019
10
embodiment, the sound source direction is estimated at high speed as compared with the
conventional method (Conventional) based on the LMS algorithm.
In general, the choice of step size parameters is not easy, as it relates to the stability and
convergence speed of the system.
In the proposed algorithm, the natural gradient (Natural Gradient) learning method is used, so
that the coefficient update amount can be appropriately varied in each time and estimated
coordinates, and the tracking speed can be improved.
[0051]
It is a schematic diagram which shows the coordinate system of a microphone array.
It is a block diagram showing composition of a circuit for making a microphone sound reception
signal in phase in a microphone array.
It is a graph which shows a simulation result.
Explanation of sign
[0052]
Mi microphone S sound source 10 In-phase circuit
04-05-2019
11
Документ
Категория
Без категории
Просмотров
0
Размер файла
18 Кб
Теги
jp2005260744
1/--страниц
Пожаловаться на содержимое документа