Laid-Open Patent 10-2009-0037845
`
`(19) Korea Intellectual Property Office (KR)
`(12) Laid-Open Patent Gazette (A)
`(51) Int. Cl.
`G10L 15/02 (2006.01) G10L 15/20 (2006.01)
`(21) Application No. 10-2008-0129411 (Divisional)
`(22) Filing Date: December 18, 2008
` Examination Request Date: None
`(62) Original application: Patent 10-2007-0103166
` Original filing date October 12, 2007
`
`(11) Laid-Open No. 10-2009-0037845
`(43) Laid-Open Date: April 16, 2009
`(71) Applicant
`Samsung Electronics Co. Ltd.
`416 Maetan-dong, Yeongtong-gu, Suwon-si,
`Gyeonggi-do
`(72) Inventors
`So-young JEONG
`Apt. 401, Ghana Building, 10-59 Yangjae 1-
`dong, Seocho-gu, Seoul
`Kwang-cheol OH
`Apt. 301-1403, Halla Provence II, Jukjeon 1-
`dong, Suji-gu, Yongin-si, Gyeonggi-do
`(Continued on reverse)
`(74) Agent
`Y.P. Lee, Mock & Partners
`
`Total number of claims: 14 claims
`(54) Method and apparatus for extracting a target sound source signal from a mixed signal
`
`(57) Abstract
`According to a method and apparatus for extracting a target sound source signal from a mixed signal: a
`mixed signal is acquired from a microphone array; with respect to the mixed signal, a first signal is
`generated with emphasized directivity in a target sound source direction and a second signal is generated
`with suppressed directivity in the target sound source direction; an adaptive nonlinear filter is calculated
`for at least one of an amplitude ratio of the first signal to the second signal, frequencies of the signals, and
`a ratio of an interference signal to the mixed signal; and a target sound source signal is extracted by
`filtering the first signal with the nonlinear filter.
`
`Representative Drawing
`
`
`
`– 1 –
`
`
`
`
`
`Page 1 of 42
`
`GOOGLE EXHIBIT 1008
`
`

`

`(72) Inventors
`Jae-hoon JEONG
`Apt. 1002, Bldg. 614, Sangnok Apts. VI, Pungdeokcheon-dong, Suji-gu, Yongin-si, Gyeonggi-do
`Kyu-hong KIM
`244-1202 Ssangyong Apts., 1052-2 Yeongtong-dong, Yeongtong-gu, Suwon-si, Gyeonggi-do
`
`
`– 2 –
`
`
`
`Page 2 of 42
`
`

`

`Claims
`Claim 1
`A method of adaptively extracting a target sound source signal from a mixed signal, comprising:
`a step in which the mixed signal is acquired from a microphone array;
`a step in which a first signal is generated in which directivity is emphasized in the direction of the target
`sound source and a second signal is generated in which directivity is suppressed in the direction of the
`target sound source, with respect to the mixed signal;
`a step in which a nonlinear filter is calculated that is adaptive to at least one of: an amplitude ratio of the
`first signal to the second signal in the time-frequency domain, the frequencies of the signals, and the ratio
`of an interference signal to the mixed signal; and
`a step in which the first signal is filtered with the nonlinear filter.
`
`Claim 2
`The method according to Claim 1, wherein
`in the step in which the nonlinear filter is calculated, a coefficient of the nonlinear filter is defined so that
`the output of the nonlinear filter increases as the amplitude ratio increases.
`
`Claim 3
`The method according to Claim 1, wherein
`in the step in which the nonlinear filter is calculated, a coefficient of the nonlinear filter is defined so that
`the output of the nonlinear filter decreases as the frequency increases.
`
`Claim 4
`The method according to Claim 1, wherein
`in the step in which the nonlinear filter is calculated, a coefficient of the nonlinear filter is defined so that
`the output of the nonlinear filter decreases as the interference signal ratio increases.
`
`Claim 5
`The method according to Claim 1, wherein
`in the step in which the nonlinear filter is calculated, the nonlinear filter is calculated using a sigmoid
`function adaptive to at least one of: the amplitude ratio, the frequencies of the signals, and the interference
`signal ratio.
`
`Claim 6
`The method according to Claim 1, further comprising
`a step in which the target sound source direction is detected based on the mixed signal, using a
`predetermined sound source search algorithm.
`
`Claim 7
`The method according to Claim 6, wherein
`the predetermined sound source search algorithm specifies, as the target sound source direction, a
`direction in which a sound source having a relatively large signal-to-noise ratio is located with respect to
`the microphone array.
`
`Claim 8
`A computer-readable recording medium in which is recorded a program for executing the method of any
`of Claims 1 to 7 on a computer.
`
`
`
`– 3 –
`
`
`
`Page 3 of 42
`
`

`

`Claim 9
`An apparatus that adaptively extracts a target sound source signal from a mixed signal, comprising:
`a microphone array that acquires the mixed signal;
`a beam-former that generates, with respect to the mixed signal, a first signal, the directivity of which is
`emphasized in the direction of the target sound source, and a second signal, the directivity of which is
`suppressed in the direction of the target sound source;
`a nonlinear filter calculating unit that calculates a nonlinear filter that is adaptive to at least one of an
`amplitude ratio of the first signal to the second signal in a time-frequency domain, the frequencies of the
`signals, and a ratio of an interference signal to the mixed signal; and
`an extracting unit that extracts the target sound source signal from the first signal by filtering the first
`signal with the nonlinear filter.
`
`Claim 10
`The apparatus according to Claim 9, wherein
`the nonlinear filter calculating unit further comprises an amplitude ratio calculating unit that calculates an
`amplitude ratio of the first signal to the second signal in the time-frequency domain and uses it as an input
`variable of the nonlinear filter.
`
`Claim 11
`The apparatus according to Claim 9, wherein
`the nonlinear filter calculating unit further comprises a frequency-adaptive coefficient calculating unit that
`uses the frequencies of the signals to calculate a slope coefficient of the nonlinear filter.
`
`Claim 12
`The apparatus according to Claim 9, further comprising
`an interference noise ratio adaptive coefficient calculating unit that calculates a ratio of the interference
`signal to the mixed signal at an arbitrary time and uses it as a bias coefficient of the nonlinear filter.
`
`Claim 13
`The apparatus according to Claim 9, further comprising
`a sound source searching unit that detects the target sound source direction from the mixed signal using a
`predetermined sound source search algorithm.
`
`Claim 14
`The apparatus according to Claim 13, wherein
`the predetermined sound source search algorithm specifies, as the target sound source direction, a
`direction in which a sound source having a relatively large signal-to-noise ratio is located with respect to
`the microphone array.
`
`Specification
`Detailed Description of the Invention
` Technical Field
`<1>
`The present invention relates to an invention for a method and apparatus for extracting a sound
`source signal from a mixed signal with respect to a particular sound source; more particularly, it relates to
`a method and apparatus for processing a mixed signal in order to extract a target sound source signal that
`a user desires from a mixed signal that comprises a variety of sound sources and is input to a portable
`digital device or the like that is capable of sound signal processing or sound acquisition, such as a mobile
`phone, camcorder, digital recorder, or the like.
`
`– 4 –
`
`
`
`Page 4 of 42
`
`

`

` Background Art
`
`<2> An era has arrived in which it is commonplace to make phone calls, record external voices, or
`acquire videos using a portable digital device. In various digital devices, such as CE (consumer
`electronics) devices and mobile phones and the like, a microphone is used as a means for acquiring sound;
`in order to achieve a stereo sound using two or more channels rather than a monophonic sound with a
`single channel, typically a microphone array is used that comprises a plurality of microphones.
`
`By combining a plurality of microphones, the microphone array may obtain not only the sound
`<3>
`itself, but also additional properties regarding directivity such as the direction or position of the sound that
`is sought to be acquired. “Directivity” refers to using the difference between the times at which the sound
`source signal respectively arrives at each of the plurality of microphones that make up the array by
`increasing the sensitivity to the sound source signal emitted from the sound source that is located in a
`particular direction. Accordingly, it is possible to emphasize or suppress a sound source signal that is
`input from a particular direction by acquiring sound source signals using such a microphone array.
`
`Research is ongoing to address the effects of musical noise or noise caused by a rapidly changing
`<4>
`ambient environment on a filter that acquires a mixed signal, in which the target sound source and
`interference noise are mixed using a microphone array, and extracts the target sound source signal from
`the mixed signal. In addition, the International Telecommunication Union (ITU) uses the PESQ
`(Perceptual Evaluation Speech Quality) index to objectively evaluate sound quality by comparing an
`input voice to an output voice.
`
`Below, the term “sound source” refers to a source from which sound is emitted and is used to
`<5>
`denote an individual speaker included in a speaker array, and the term “sound field” refers to a virtual
`region formed by the sound emitted from the sound source, i.e., the region that sound energy reaches. In
`addition, “sound pressure” refers to a force that sound energy exerts and is expressed as a physical
`quantity of pressure.
`
`
`
` Summary of the Invention
`
` Problems the Invention is Intended to Solve
`
`The technical problem that the present invention is intended to solve is the provision of a method
`<6>
`and apparatus for separating a target sound source, which solves the problem of not being able to clearly
`separate a particular sound source signal from a mixed signal that comprises a plurality of sounds that are
`input via a microphone array.
`
`<7> Another technical problem that at least one of the embodiments of the present invention seeks to
`solve is the provision of a method and an apparatus that extract a target sound source signal from a mixed
`signal in response to a rapidly changing ambient environment. In addition, the present invention is
`intended to provide a computer-readable recording medium on which is recorded a program for carrying
`out the method on a computer. The technical problems to be solved by the present embodiments are not
`limited to the technical problems described above, and other technical problems may exist.
`
` Means of Solving the Problems
`
`To solve the above technical problems, the method of extracting a target sound source signal
`<8>
`according to the present invention comprises: a step in which a mixed signal is received as input through a
`microphone array; a step in which, with respect to the mixed signal, a first signal is generated in which
`
`Page 5 of 42
`
`

`

`directivity is emphasized in a direction of a target sound source and a second signal is generated in which
`directivity is suppressed in the direction of the target sound source; and a step in which a target sound
`source signal is extracted from the first signal by masking, based on the ratio of the first signal to the
`second signal, the interference sound source signal included in the first signal.
`
`To solve another technical problem, the present invention provides a computer-readable recording
`<9>
`medium on which is recorded a program for carrying out the above-described target sound source signal
`extraction method on a computer.
`
`<10> To solve the above technical problems, the target sound source signal extraction apparatus
`according to the present invention comprises: a microphone array that receives a mixed signal as input; a
`beam-former that generates, with respect to a mixed signal, a first signal the directivity of which is
`emphasized in a direction of a target sound source, and a second signal the directivity of which is
`suppressed in the direction of the target sound source; and a signal extracting unit that extracts a target
`sound source signal from the first signal by masking, based on a ratio of the first signal to the second
`signal, the interference sound source signal included in the first signal.
`
`<11> To solve the above technical problems, the method according to the present embodiment for
`adaptively extracting a target sound source signal from a mixed signal comprises: a step in which the
`mixed signal is extracted from a microphone array; a step in which, with respect to the mixed signal, a
`first signal is generated in which directivity is emphasized in a direction of a target sound source and a
`second signal is generated in which directivity is suppressed in the direction of the target sound source
`with respect to the mixed signal; a step in which a nonlinear filter is calculated that is adaptive to at least
`one of an amplitude ratio of the first signal to the second signal in a time-frequency domain, the
`frequencies of the signals, and a ratio of an interference signal to the mixed signal; and a step in which the
`first signal is filtered with the nonlinear filter.
`
`
`
`
`
`– 5 –
`
`Page 6 of 42
`
`

`

`<12> To solve another technical problem, the present invention provides a computer-readable recording
`medium on which is recorded a program for carrying out the above-described adaptive target sound
`source signal extraction method on a computer.
`
`In order to solve another technical problem above, the apparatus according to the present
`<13>
`embodiment for adaptively extracting a target sound source signal from a mixed signal comprises: a
`microphone array that acquires the mixed signal; a beam-former that generates, with respect to the mixed
`signal, a first signal in which directivity is emphasized in a direction of a target sound source and a second
`signal in which directivity is suppressed in the direction of the target sound source; a nonlinear filter
`calculating unit that calculates a nonlinear filter that is adaptive to at least one of: an amplitude ratio of
`the first signal to the second signal in a time-frequency domain, the frequencies of the signals, and a ratio
`of an interference signal to the mixed signal; and an extracting unit that extracts the target sound source
`signal from the first signal by filtering the first signal with the nonlinear filter.
`
`
`
` Effect
`
`<14> As described above, with respect to a rapidly changing ambient environment or musical noise, it
`is possible to extract a target sound source signal with high PESQ (Perceptual Evaluation of Speech
`Quality) from a mixed signal using a nonlinear filter that is adaptive to an amplitude ratio, frequencies,
`and a ratio of interference noise to the mixed signal.
`
`
`
` Specific Details for Implementing the Invention
`
`<15> Hereinafter, various embodiments of the present invention are described in detail, with reference
`to the drawings.
`
`In general, an environment in which sound is recorded or a voice signal is received as input
`<16>
`through a portable digital device is more commonly an environment that also includes various noises and
`ambient interference noise rather than a quiet environment without ambient interference noise. In
`particular, because in a conventional mobile phone, in which only voice calls were possible, the caller
`was very close to the mobile phone, the introduction of interference noise through a microphone furnished
`in the mobile phone had not been a significant problem; but with the recent spread of communication
`means capable of video calling, the influence of interference noise on the caller's voice signal has
`increased in relative terms, and as a result, the problem has arisen that it is difficult to communicate
`clearly. That being the case, there is an increasing demand for a method of extracting a target sound
`source signal from a mixed signal in various sound acquisition devices such as mobile phones and CE
`(consumer electronics) devices furnished with microphones.
`
`<17> FIG. 1 illustrates a problem situation that the present invention seeks to address, wherein
`distances from a microphone array (110) to surrounding sound sources are represented by concentric
`circles. FIG. 1 shows a plurality of sound sources arranged around the microphone array (110), and each
`respective sound source differs in both distance from and direction to the microphone array (110). If it
`sought to acquire sound through the microphone array (110), various sounds emitted from these sound
`sources are mixed and input to the microphone array (110), and out of the plurality of sound sources it is
`sought to clearly acquire the sound emitted from a particular sound source.
`
`Page 7 of 42
`
`

`

`<18> Such a particular sound source may be specified according to an environment in which various
`embodiments of the present invention described below are implemented; in general, it may be specified as
`a sound source signal that is dominant among a plurality of sound source signals included in a mixed
`signal. In other words, a signal having a large gain or sound pressure in the sound source signal may be
`specified as the target sound source. As another method for specifying the target sound source, a method
`may be used that takes into account the direction or distance from the microphone array. In other words,
`the more that a sound source is located in front of the microphone array or is a sound source located
`closer to the microphone array, the greater the likelihood that it will become the target sound source. FIG.
`1 illustrates a situation in which a sound source (120) located close to the front of the microphone array
`(110) is specified as a target sound source and is extracted from the mixed signal.
`
`<19> As described above, because the specification of the target sound source may vary depending on
`the environment in which various embodiments of the present invention are implemented, a person of
`ordinary skill in the art to which the present invention pertains will recognize that various methods can be
`used in addition to the two methods above.
`
`<20> FIG. 2a and FIG. 2b are block diagrams that depict a target sound source extraction apparatus
`according to an embodiment of the present invention; the drawings respectively show the situation in
`which the direction to the target sound source is known, and the situation in which it is not known.
`
`
`
`
`
`– 6 –
`
`
`
`Page 8 of 42
`
`

`

`In the target sound source extracting apparatus of FIG. 2a, it is assumed that the direction in
`<21>
`which the target sound source is located is specified by means of the various methods described in FIG. 1;
`the apparatus comprises a microphone array (210), beam-formers (220), and a signal extracting unit (230).
`
`<22> The microphone array (210) acquires, in the form of a mixed signal, sound source signals emitted
`from a plurality of sound sources located in its vicinity. Because the microphone array (210) is made up
`of a plurality of microphones, the time for the plurality of sound source signals to reach each microphone
`will vary as a function of the location and distance of the respective sound source. Let the N sound source
`signals that are input via the N microphones that make up the array be represented as X1(t), X2(t) to XN(t),
`respectively.
`
`<23> The beam-formers (220) generate, with respect to the sound source signals input through the
`microphone array (210), a signal the directivity of which is emphasized in the direction of the target
`sound source, and a signal the directivity of which is suppressed in the direction of the target sound
`source. These roles are respectively carried out by means of the emphasized signal beam-former (221)
`and the suppressed signal beam-former (222).
`
`In general, a microphone array made up of two or more microphones serves as a filter that can
`<24>
`spatially reduce noise when the direction of an interference noise signal differs from the desired target
`signal, by increasing amplitude by assigning respectively appropriate weights to each signal received at
`the microphone array in order to receive with high sensitivity a target signal mixed with background noise;
`a spatial filter of this kind is called a beam-former. In order to amplify or extract a target signal from
`noise in a different direction, it is necessary to obtain a phase difference between an array pattern and the
`signals respectively input to each microphone; a plurality of beam-forming algorithms are known for
`obtaining such signal information.
`
`<25> Typical beam-forming algorithms for amplifying or extracting a target sound source signal
`include: delay-and-sum algorithms that identify the location of a sound source from the relative delay
`time of the sound source signal that arrives at the microphones; and filter-and-sum algorithms that use a
`spatially linear filter to filter the output in order to reduce the influence due to noise and two or more
`signals in a sound field made up of sound sources. These beam-forming algorithms are widely known
`among persons of ordinary skill in the art to which the present invention pertains.
`
`In FIG. 2a, the emphasized signal beam-former (221) increases the sound pressure with respect to
`<26>
`the target sound source by increasing its directional sensitivity with respect to the specified target sound
`source. A method of adjusting directional sensitivity will be described below with reference to FIG. 3a
`and FIG. 3b.
`
`<27> FIG. 3a and FIG. 3b are block diagrams that depict a target sound source emphasizing beam-
`former according to an embodiment of the present invention; they respectively illustrate methods using a
`fixed filter and an adaptive delay term.
`
`In FIG. 3a, it is assumed that the target sound source is in front of the microphone array (310); the
`<28>
`directivity in the target sound source direction is increased by enhancing the sound pressure of the target
`sound source by adding, via the adder (320), the sound source signals input via the microphone array
`(310). In FIG. 3a, A, B, and C respectively denote a location of a respective sound source. In the present
`embodiment, because it is assumed that the target sound source is located at point A in front of the
`microphone array (310), the sound sources located at points B and C will become interference noise.
`
`Page 9 of 42
`
`

`

`If a sound source signal emitted from a point A located in front of the microphone array (310) is
`<29>
`input to the microphone array (310) as part of the mixed signal, then the phase and magnitude of the input
`sound source signals will be nearly identical. As a result, for the input sound source signals, the gain of
`the signal is enhanced via the adder (320), and a signal is output the phase of which is unchanged. On the
`other hand, when a sound source signal emitted from a point B or C is input to the microphone array
`(310), there is a difference in the time that the sound source signal reaches each of the respective
`microphones because there is a difference in the angle and distance between the sound source and each
`microphone that makes up the array. In other words, the sound source signal emitted from the point B or
`C will arrive sooner at the microphone located closer to the sound source, and will arrive relatively later
`to the microphone located further from the sound source. When signals having a difference in arrival
`times are added via the adder (320), either the signals are partially canceled due to the difference in
`arrival time between the respective signals, or the gain is reduced due to the difference between the
`phases of the signals. Even if the phase difference between the signals does not match precisely, it has the
`effect that the gain of the signal is relatively reduced compared to the sound source signal from the point
`A. Therefore, as with the present embodiment, the directional sensitivity of the target sound source
`located in front of the microphone array (310) may be improved simply by means of the microphone
`array (310), which has fixed gaps, and the adder (320).
`
`<30> FIG. 3b shows a target sound source emphasizing beam-former that enhances directivity with
`respect to a target sound source direction; for ease of description, a first-order differential microphone
`structure is used that is made up of only two microphones.
`
`
`
`– 7 –
`
`
`
`Page 10 of 42
`
`

`

`<31> First, when the sound source signals input from the microphone array are denoted as X1(t) and
`X2(t) respectively, the delaying unit (330) delays the input signal X1(t) by a certain amount of time via
`adjustment of the adaptive delay term, and then the delayed input signal X1(t) is subtracted from the input
`signal X2(t) via the subtractor (340), and a sound source signal is generated having directivity in a
`particular direction. Finally, when the sound source signal generated as a result of the subtraction is
`filtered through a low-pass filter (LPF) (350), an emphasized sound source signal is output that is
`independent of the frequency change of the sound source signal. (Acoustic signal processing for
`telecommunication, Steven L. Gay and Jacob Benesty, Kluwer Academic Publishers, 2000.) Such a
`beam-former is called a delay-and-subtract beam-former; such a beam-former may be readily understood
`by a person of ordinary skill in the art to which the present invention pertains, and therefore will be
`briefly described below only to the extent necessary for the present embodiment.
`
`In general, directional control factors, such as the gap between microphones constituting the array
`<32>
`and a delay time of the respective sound source signals applied to each microphone, are widely known as
`factors that determine the directional response of a microphone array. The relationship between these
`directional control factors is defined as in Equation 1 below.
`
`Equation 1
`
`
`
`<33>
`
`<34> Here, τ is the adaptive delay term that determines the directional response, d is the gap between
`microphones, α1 is an adjustment variable introduced to define the relationship between the sound
`pressure field and the directional control factors, and c is the speed of sound in air, that is, 340 m/sec.
`
`In FIG. 3b, the delaying unit (330) determines the delay term according to Equation 1 based on
`<35>
`the direction of the sound source signal that emphasizes directivity, and delays the input signal X1(t) by
`the value of the delay term thus determined. The subtractor (340) then subtracts the delayed input signal
`X1(t) from the input signal X2(t). As a function of such delay, a time difference arises among the
`microphones that make up the array; as a result, an emphasized signal with increased directivity in a
`particular direction (namely the direction of the target sound source) can be obtained from the sound
`source signal that has been input to the microphone array.
`
`<36> The sound pressure field of the input signal X1(t) delayed by the delaying unit (330) is defined as
`a function of the angular frequency of the signal and the angle at which the sound source signal from the
`sound source is incident to the microphone array. The sound pressure field is difficult to control because
`the sound pressure field changes according to various variables including the gap between microphones
`and the angle of incidence of the sound source signal; in particular, among these variables, the frequency
`and amplitude of the sound source signal vary with the properties of the sound source signal. It is
`therefore necessary to be able to control the sound pressure field only with the adaptive delay term of
`Equation 1, regardless of changes in the frequency or amplitude of the sound source signal.
`
`<37> The low-pass filter (350) suppresses changes in the sound pressure field due to changes in an
`abnormal frequency by holding constant a frequency component included in the sound pressure field. As
`a result, when the sound source signal that is output via the subtractor (340) is filtered through the low-
`pass filter (350) again, the directivity in the target sound source direction can be adjusted with only the
`
`Page 11 of 42
`
`

`

`adaptive delay term in Equation 1, regardless of the frequency or amplitude of the sound source signal. In
`other words, it is possible to generate the emphasized sound source signal Y(t) having increased
`directivity in the target sound source direction via the target sound source emphasizing beam-former
`shown in FIG. 3b.
`
`<38> Above, with reference to FIG. 3a and FIG. 3b, two embodiments of the target sound source
`emphasizing beam-former have been described that enhance the directivity of the target sound source. In
`contrast to this, there is a beam-former that reduces the sound source signal that is incident from the
`direction in which the target sound source is located by suppressing the directivity of the target sound
`source; this is referred to as a target sound source suppressing beam-former.
`
`<39> FIG. 4a and FIG. 4b are block diagrams that depict a target sound source suppressing beam-
`former according to an embodiment of the present invention; they respectively illustrate methods using a
`fixed filter and an adaptive delay term.
`
`In FIG. 4a, it is assumed that, as in FIG. 3a, the target sound source exists in front of the
`<40>
`microphone array (410). In addition, it is assumed that sound sources are respectively located at A, B, and
`C. In addition, as in FIG. 3a, because it is assumed in the present embodiment that the target sound source
`is located at point A in front of the microphone array (410), the sound sources located at points B and C
`will become interference noise. In FIG. 4a, signal values of + and – are respectively alternately assigned
`to the sound source signals input via the microphone array (410), and all signals are then added via the
`adder (420), thus suppressing directivity in the direction of the target sound source.
`
`
`
`– 8 –
`
`
`
`Page 12 of 42
`
`

`

`The signal values of + and – illustrated in FIG. 4a may be assigned by multiplying the input signal by a
`matrix such as (–1, +1, –1, +1). A matrix in which signs are assigned alternately so as to attenuate sound
`source signals input to nearby microphones in this way is called a blocking matrix.
`
`<41> The directivity suppression process may be described in greater detail as follows. First, when the
`sound source signal emitted from point A is input to the microphone array (410) as part of a mixed signal,
`the sound source signal input through those of the four microphones that are near one another will have a
`very similar phase and size. In other words, the input signals will be similar to each other between the
`microphones located in the first and second, second and third, and third and fourth positions. Accordingly,
`if respectively opposite signs are assigned to sound source signals that are input via adjacent microphones
`and added via the adder (420), this will have the effect that the nearby signals cancel each other out.
`Accordingly, directivity in the target sound source direction is suppressed by reducing the gain or sound
`pressure of the sound source signal that is input from the sound source A located in front of the
`microphone array (410).
`
`<42> On the other hand, when a sound source signal emitted from point B or C is input to the
`microphone array (410), there is a delay by a respective specified time, according to a distance from the
`sound source, at each microphone that makes up the array. In other words, there is a difference between
`the arrival times of the sound source signals emitted from point B or C arriving at the microphone. Even if
`the signals having such a time difference are added via the adder (420) after opposite signals have been
`assigned to nearby microphones, the canceling effect of the signals at point B or C is not as large, due to
`the time difference between the signals. Therefore, as with the present embodiment, the directional
`sensitivity for a target sound source located in front of the microphone array (410) may be suppressed by
`first multiplying by the opposite sign the signals near a microphone array (410) that has a fixed gap, and
`then adding them via the adder (420).
`
`<43> FIG. 4b is a target sound source suppressing beam-former for suppressing directivity in a target
`sound source direction; because it uses the first-order differential microphone structure described in FI