JP2009135593A - Acoustic input device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To emphasize and output sound from a sound source present in an object direction. <P>SOLUTION: A phase difference correction part 20 extracts discrete values of first and second observation signals generated by first and second directional microphones 10 and 11, and shifts only the phase difference in the case where a sound source is present in the object direction with respect to the discrete value of the second observation signal. A sound source direction determination part 21 calculates first and second observation power values being power values of the first and second observation signals. When at least one of the first and second observation power values is not larger than a threshold value, it is determined that the sound source is not present in the object direction and, when both the first and second observation power values are larger than the threshold value and a relative ratio of the first and second observation power values is smaller than another threshold value, it is determined that the sound source is present in the object direction. An output calculation part 22 outputs a synthesized signal of the first and second observation signals by amplifying it with a large amplification factor when the sound source is present in the object direction, and outputs the synthesized signal by amplifying it with a small amplification factor when the sound source is not present in the object direction. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an acoustic input device that emphasizes and outputs sound from a sound source that exists in a predetermined target direction.

  As an acoustic input device that collects sound from a sound source, for example, it is provided in a door phone sub unit of an interphone system, and the sound (sound) of a specific speaker (sound source) existing in front of the door phone sub unit is collected. There is something to do. In the acoustic input device provided in such a door phone sub unit, only the voice of the speaker is collected, and the sound other than the voice of the speaker (for example, reflected sound from the wall surface or ambient noise) is not collected. Therefore, it is necessary to predict in advance the direction in which the speaker exists (target direction) and to direct the directivity to the predicted target direction.

  As a conventional acoustic input device that collects sound from a sound source that exists in a target direction, there is one that collects sound using a single directional microphone. The directional microphone can collect sound from a sound source that exists in a sound collection region with a somewhat small directivity angle.

  As another example of a conventional acoustic input device that collects sound from a sound source, Patent Document 1 discloses a omnidirectional microphone that is a pressure microphone, a bidirectional microphone that is a speed microphone, and a non-directional microphone. A sound source direction determination apparatus is disclosed that includes a phase comparator that determines whether the phase of the waveform signal of the bidirectional microphone is advanced or delayed with respect to the waveform signal of the directional microphone.

  In the sound source direction determination apparatus of Patent Document 1, the waveform signal of the bidirectional microphone (speed microphone) is omnidirectional when the sound source is present on the front side of two microphones (omnidirectional microphone and bidirectional microphone). When the sound source is present on the back side of the two microphones, the phase is delayed by 90 degrees from the waveform signal of the omnidirectional microphone. When the phase comparator determines that the phase of the waveform signal of the bi-directional microphone is ahead of the waveform signal of the omnidirectional microphone, the phase comparator determines that the sound source is present on the front side, while the waveform signal of the omnidirectional microphone If it is determined that the sound source is slow, it is determined that the sound source exists on the back side.

From the above, according to the sound source direction determination device of Patent Document 1, it can be determined whether the sound source exists on the front side or the back side of the two microphones.
JP 2003-333680 A (paragraphs 0028 to 0048 and FIGS. 1 to 3)

  However, the conventional sound input device using a single microphone can collect and output sound from a sound source that exists within a certain range, but it is not sufficient from a practical point of view. There is a problem that noise and the like existing outside the target direction are emphasized and output in the same manner as the sound from the sound source existing in the target direction (for example, the voice of a specific speaker).

  As a conventional means for solving the above problem, it is conceivable to lengthen the acoustic tube of the microphone, but it is not practical. In addition, although signal processing such as noise removal may be performed, a new problem of excessive calculation amount occurs.

  Moreover, although the sound source direction determination device of Patent Document 1 can determine whether a sound source exists on the front side or the back side of two microphones, the sound source is detected from a sound source that exists in a specific target direction. The sound could not be output with emphasis.

  The present invention has been made in view of the above points, and an object of the present invention is to provide an acoustic input device that can emphasize and output sound from a sound source existing in a target direction.

  According to the first aspect of the present invention, each of the sound collecting areas has a predetermined directivity angle, and each sound collecting area is installed close to each other so as to include a predetermined target direction. A first directional microphone and a second directional microphone that receive the sound wave from the first wave and generate an observation signal based on the received sound wave, and receive the first observation signal from the first directional microphone. The second observation signal is received from the second directional microphone, and the phase difference between the received first observation signal and the second observation signal is used as the first observation when the sound source exists in the target direction. A phase difference correction unit that corrects only a phase difference between the signal and the second observation signal; and the first observation signal and the second observation signal that have been corrected by the phase difference correction unit. Sound source direction determination unit that determines whether or not the target direction exists When the sound source direction determination unit determines that the sound source is present in the target direction, at least one of the first observation signal and the second observation signal is amplified with a first amplification factor and output. When the sound source direction determination unit determines that the sound source does not exist in the target direction, the phase difference between the first observation signal and the second observation signal is corrected by the phase difference, And an output calculation unit for amplifying and outputting a combined signal of the first observation signal and the second observation signal at an amplification factor smaller than the first amplification factor.

  According to a second aspect of the present invention, in the first aspect of the invention, the sound source direction determination unit includes a first observation power value that is a power value of the first observation signal after being corrected by the phase difference correction unit, Both of the second observation power values, which are the power values of the second observation signal after being corrected by the phase difference correction unit, are larger than a predetermined threshold, and the first observation power value and the second observation power. When the condition that the relative ratio of values is less than a predetermined set value is satisfied, it is determined that the sound source exists in the target direction. On the other hand, when the condition is not satisfied, it is determined that the sound source does not exist in the target direction. It is characterized by that.

According to a third aspect of the present invention, in the first aspect of the invention, the sound source direction determination unit includes the discrete value x1 of the first observation signal and the discrete value of the second observation signal after being corrected by the phase difference correction unit. x2 is used to obtain a correlation function R _ij (i = 1, 2) represented by Equation 3, and a predetermined set value in which the cross correlation function R ₁₂ is set based on the autocorrelation function R ₁₁ or the autocorrelation function R ₂₂ When the above condition is satisfied, it is determined that the sound source is present in the destination direction. On the other hand, when the condition is not satisfied, it is determined that the sound source is not present in the destination direction.

  According to a fourth aspect of the present invention, in the first aspect of the present invention, the sound wave received from the sound source installed at an intermediate position between the first directional microphone and the second directional microphone is received. An omnidirectional microphone that generates a reference observation signal based on the reference difference, and the phase difference correction unit receives the reference observation signal and sets the phase difference between the received reference observation signal and the first observation signal in the target direction. Only the phase difference between the reference observation signal and the first observation signal when a sound source is present is corrected, and the sound source direction determination unit is a reference observation that is a power value of the reference observation signal after being corrected by the phase difference correction unit. A relative ratio of the first observation power value that is the power value of the first observation signal after being corrected by the phase difference correction unit with respect to the power value is included in a preset first range, and the reference observation power value Said phase with respect to When the relative ratio of the second observation power values, which are the power values of the second observation signal after being corrected by the correction unit, satisfies the condition included in the preset second range, the sound source is in the target direction If the condition is not satisfied, it is determined that the sound source does not exist in the target direction.

According to a fifth aspect of the present invention, in the first aspect of the present invention, the sound wave received from the sound source is installed at an intermediate position between the first directional microphone and the second directional microphone, and the received sound wave is converted into the received sound wave. An omnidirectional microphone that generates a reference observation signal based on the phase difference correction unit that extracts a discrete value of the received first observation signal at predetermined time intervals and receives the received second observation signal The discrete value of the second observation signal is extracted at the predetermined time interval, the reference observation signal is received, the discrete value of the reference observation signal is extracted at the predetermined time interval, and the extracted first observation The phase difference between the discrete value of the signal and the second observation signal is corrected by the phase difference between the first observation signal and the second observation signal when a sound source exists in the target direction, and the extracted reference observation signal Discrete value and first observation signal Is corrected by the phase difference between the reference observation signal and the first observation signal when a sound source is present in the target direction, and the sound source direction determination unit is corrected by the phase difference correction unit. The discrete value x1 of one observation signal, the discrete value x2 of the second observation signal after correction by the phase difference correction unit, and the discrete value x0 of the reference observation signal after correction by the phase difference correction unit are used. Correlation function R _ij (i, j = 0,1,2,2) expressed by Equation 4 is obtained, and the relative ratio of cross-correlation function R ₀₁ to autocorrelation function R ₀₀ is included in the preset first range. If the condition is satisfied the relative ratio of the cross-correlation function R ₀₂ for said autocorrelation function R ₀₀ is included in the second range set beforehand, while determines that the sound source is present in the target direction, it does not satisfy the condition If the sound source is in the target direction And judging that no standing.

  The invention according to claim 6 is the invention according to any one of claims 1 to 5, wherein the sound source direction determination unit determines that the sound source is in the target direction when it is determined that the sound source does not exist in the target direction. It is determined whether or not the sound source exists in the vicinity, and when the sound source direction determination unit determines that the sound source exists in the vicinity of the target direction, the output calculation unit determines whether the first sound source moves away from the target direction. The synthesized signal is amplified and output with an amplification factor of a continuous change characteristic that monotonously decreases from the amplification factor.

  The invention according to claim 7 is the invention according to claim 1, wherein the sound wave received from the sound source installed at an intermediate position between the first directional microphone and the second directional microphone is received. An omnidirectional microphone that generates a reference observation signal based on the reference difference, and the phase difference correction unit receives the reference observation signal and sets the phase difference between the received reference observation signal and the first observation signal in the target direction. The phase difference between the reference observation signal and the first observation signal when a sound source is present is corrected, and the sound source direction determination unit is the power value of the first observation signal after being corrected by the phase difference correction unit. Both the first observation power value and the second observation power value that is the power value of the second observation signal after being corrected by the phase difference correction unit are greater than a predetermined threshold, Observation power value and second observation power When the relative ratio of the values is equal to or greater than a predetermined first set value, the cross-correlation function between the first observation signal and the second observation signal is the autocorrelation function of the first observation signal or the first When satisfying a condition that is equal to or greater than a predetermined second set value obtained from the autocorrelation function of the two observation signals, the sound source is determined to exist in each of the target direction and other regions other than the target direction, When the condition is not satisfied, it is determined that the sound source does not exist in the target direction, and when it is determined that the sound source exists in each of the target direction and other regions other than the target direction, the phase difference correction unit The relative ratio of the first observation power value to the reference observation power value, which is the power value of the reference observation signal after being corrected in step S1, is included in a preset range, and the first observation power value relative to the reference observation power value is When the relative ratio of the observed power values is not included in the range, it is determined that the sound source exists in the sound collection region of the first directional microphone, and the second observed power value with respect to the reference observed power value is When the relative ratio is included in the range and the relative ratio of the first observed power value to the reference observed power value is not included in the range, the sound source exists in the sound collection region of the second directional microphone. Then, it is determined that it is determined.

  According to an eighth aspect of the present invention, in the seventh aspect of the invention, the output calculation unit includes the sound source in a sound collection region of the first directional microphone or the second directional microphone by the sound source direction determination unit. If it is determined, the observation signal of the directional microphone in which the sound source does not exist in the sound collection region is amplified with the first amplification factor and output.

  The invention according to claim 9 is the invention according to claim 7, further comprising driving means for tilting a directional axis of the first directional microphone or the second directional microphone, and the sound source direction determination unit includes: When it is determined that the target direction and the sound collecting area of the first directional microphone or the second directional microphone among the other areas are present, the first observed power value and the The driving means is controlled until a relative ratio of second observation power values is less than the first set value, and the output calculation unit is configured to control the first directional microphone or the output calculation unit by the sound source direction determination unit. When it is determined that the sound source is present in the sound collection region of the second directional microphone, a combined signal of the first observation signal and the second observation signal after driving the driving unit is When the sound source direction determination unit determines that the sound source does not exist in the target direction, the synthesized signal is output as a second amplification smaller than the first gain. The output is amplified at a rate.

  According to a tenth aspect of the present invention, in the first aspect of the present invention, the sound wave received from the sound source is installed at an intermediate position between the first directional microphone and the second directional microphone, and the received sound wave is converted into the received sound wave. An omnidirectional microphone that generates a reference observation signal based on the reference difference, and the phase difference correction unit receives the reference observation signal and sets the phase difference between the received reference observation signal and the first observation signal in the target direction. The phase difference between the reference observation signal and the first observation signal when a sound source is present is corrected, and the sound source direction determination unit is the power value of the first observation signal after being corrected by the phase difference correction unit. When both of the first observation power value and the second observation power value that is the power value of the second observation signal after being corrected by the phase difference correction unit are larger than a predetermined threshold, The first observation signal and said When the cross-correlation function of the two observation signals satisfies a condition that is equal to or greater than a predetermined set value obtained from the autocorrelation function of the first observation signal or the autocorrelation function of the second observation signal, the sound source While determining that the sound source exists in each of the target direction and other regions other than the target direction, the sound source determines that the sound source does not exist in the target direction when the condition is not satisfied. When it is determined that each region other than the direction is present, the relative ratio of the first observation power value to the reference observation power value that is the power value of the reference observation signal after being corrected by the phase difference correction unit, And the relative ratio of the second observed power value to the reference observed power value are included in a preset range, the first directional microphone and the second directional microphone Determining that the sound source to each of the sound collection region of the directional microphone is present and said.

  The invention of claim 11 is the invention of claim 10, further comprising driving means for inclining a directional axis of the first directional microphone or the second directional microphone, and the sound source direction determination unit includes: When it is determined that the target direction and the other directional microphones are present in any of the sound collection regions of the first directional microphone and the second directional microphone, the two sound sources are moved from the target direction to the target direction. The relative ratio between the first observed power value and the second observed power value when the first directional microphone and the second directional microphone are separated from each other by a half or more of the maximum value of the directional angle. The driving means is controlled until a condition that is less than a predetermined set value is satisfied, and the output calculation unit is configured to control the first directional microphone or the second directivity by the sound source direction determination unit. When it is determined that the sound source is present in the sound collection region of the icphone, the combined signal of the first observation signal and the second observation signal after driving the driving means is amplified by the first amplification factor. On the other hand, when the sound source direction determination unit determines that the sound source does not exist in the target direction, the synthesized signal is amplified and output with a second amplification factor smaller than the first amplification factor. It is characterized by.

  According to the first aspect of the present invention, the observation signals from the two microphones are used to determine whether or not the sound source exists in the target direction, and the amplification factor when the sound source exists in the target direction (first amplification) By making the rate greater than the amplification factor when the sound source does not exist in the target direction, it is possible to emphasize and output the sound from the sound source existing in the target direction.

  According to the invention of claim 2, it is possible to accurately determine whether or not the sound source exists in the target direction by evaluating the relative ratio between the first observation power value and the second observation power value.

  According to the invention of claim 3, it is accurately determined whether or not the sound source exists in the target direction by evaluating the cross-correlation function between the discrete value of the first observation signal and the discrete value of the second observation signal. can do.

  According to the fourth aspect of the present invention, the omnidirectional microphone is installed at an intermediate position between the first directional microphone and the second directional microphone, so that the first and second reference power values relative to the reference observing power value of the omnidirectional microphone can be obtained. By using the observed power value, it can be accurately determined whether or not the sound source exists in the target direction.

  According to the invention of claim 5, by using the autocorrelation function of the reference observation signal or the cross-correlation function of the reference observation signal and the first and second observation signals, the influence of uncorrelated noise is reduced and high S / N It is possible to increase the accuracy of determination as to whether or not the sound source exists in the target direction.

  According to the invention of claim 6, it is possible to prevent the amplification factor for amplifying the synthesized signal from becoming discontinuous depending on the direction of the sound source, and as a result, the output signal can always be a continuous value. .

  According to the seventh aspect of the present invention, there is one sound source in an area where the sound collection areas of the first directional microphone and the second directional microphone overlap, and the first directional microphone and the second directional microphone. It can be determined that another sound source exists only in one of the sound collection areas.

  According to the invention of claim 8, the signal from the sound source in the target direction can be amplified and output.

  According to invention of Claim 9, the signal from the sound source of the target direction can be amplified and output.

  According to the invention of claim 10, there is one sound source in a region where the sound collection regions of the first directional microphone and the second directional microphone overlap, and the first directional microphone and the second directional microphone. It can be determined that another sound source exists only in one of the sound collection areas.

  According to the eleventh aspect of the invention, it is possible to emphasize and output the signal from the sound source in the target direction.

(Embodiment 1)
First, the structure of the acoustic input device of Embodiment 1 is demonstrated using FIGS. As shown in FIG. 1, the acoustic input device includes a sound collecting unit 1 that collects sound from a sound source (not shown), and an observation signal x1 (t that will be described later based on the sound collected by the sound collecting unit 1 ), X2 (t), and a signal processing unit 2 that calculates an output signal y (k, m).

  The sound collection unit 1 includes a first directional microphone 10 and a second directional microphone 11 each having a sound collection region with predetermined directivity angles α1 and α2. As shown in FIG. 2B, the first directional microphone 10 and the second directional microphone 11 are configured such that a part of each sound collection area is in a specific target direction (the first directional microphone 10 in FIG. 2B). The directional microphone 10 and the second directional microphone 11 are disposed close to each other so as to overlap each other.

  As shown in FIG. 1, each of the first and second microphones 10 and 11 installed as described above receives a sound wave from a sound source, and an observation signal x1 (which is an electric signal based on the received sound wave). t), x2 (t). The first observation signal x1 (t) is output from the first directional microphone 10 to the signal processing unit 2, and the second observation signal x2 (t) is output from the second directional microphone 11 to the signal processing unit 2. .

  Incidentally, as shown in FIG. 3, the characteristics of the first and second directional microphones 10 and 11 are the first and second observed power values Px1 (k) described later at the position of the angle φ from the central axis A1 of the directivity. ), Px2 (k) is high and almost constant when 0 ° ≦ | φ | <φ1, decreases rapidly as the angle φ increases when φ1 ≦ | φ | <φ2, and decreases when φ2 <| φ | It is constant. In FIG. 2B, the angle θ of the region where the sound collection regions overlap is 0 ° <θ <min (α1, α2). In addition, the angle which becomes a half value of the sensitivity of φ = 0 ° is defined as a directivity angle α1 (α2).

  Although the first directional microphone 10 and the second directional microphone 11 of the present embodiment have the same sensitivity, even if the sensitivities are different, the observation signal based on the received sound wave is used. What is corrected with the sensitivity coefficient may be the new observation signals x1 (t) and x2 (t). Specifically, when the sensitivity coefficient of the first directional microphone 10 is m1 and the sensitivity coefficient of the second directional microphone 11 is m2 (m1 ≠ m2), (the first directional microphone 10 receives the wave) (Observation signal xm1) based on the sound wave received by the second directional microphone 11 is defined as the first observation signal x1 (t) transmitted to the signal processing unit 2 May be the second observation signal x2 (t) transmitted to the signal processing unit 2. At this time, the sensitivity coefficients m1 and m2 are set to satisfy m1: m2 = √ (Px2): √ (Px1) using Px1 and Px2 described later. The same applies to the following description.

  The signal processing unit 2 shown in FIG. 1 receives the first and second observation signals x1 (t) and x2 (t) and receives the discrete value x1 (k, m) of the first observation signal and the second observation signal. Phase difference correction unit 20 for correcting the phase difference of discrete values x2 (k, m) of the first and second values based on discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals. The sound source direction determination unit 21 that determines whether or not the sound source exists in the target direction using the observed power values Px1 (k) and Px2 (k) of 2, and the output signal according to the determination result of the sound source direction determination unit 21 and an output calculation unit 22 that outputs y (k, m).

  The phase difference correction unit 20 receives the first observation signal x1 (t) from the first directional microphone 10, and receives the received first observation signal x1 (t) as shown in FIG. 5B. The frame is divided into frames having a frame length L, and a discrete value x1 (k, m) of the first observation signal is extracted at each predetermined time interval Δt for each frame. Similarly, the second observation signal x2 (t) is received from the second directional microphone 11, and the second observation signal x2 (t) is divided into frames having the frame length L, and the second observation signal x2 (t) is divided into frames for each frame. A discrete value x2 (k, m) of the observed signal is extracted at a predetermined time interval Δt. The discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals are the mth discrete values in the kth frame.

  The phase difference correction unit 20 that has extracted the discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals and the extracted discrete value x1 (k, m) of the first observation signal. The phase difference between the discrete values x2 (k, m) of the second observation signal is the phase difference between the first observation signal x1 (t) and the second observation signal x2 (t) when a sound source is present in the target direction. Only correct.

  For example, when the target direction is a direction B inclined by an angle φ from the central axis A of the first directional microphone 10 and the second directional microphone 11 as shown in FIG. As shown in FIG. 2, when the sound source S exists in the target direction, the first directional microphone 10 and the second directional microphone 11 are installed close to each other, so that the sound waves from the sound source S are incident substantially in parallel. To do. Therefore, a phase difference 2dsinφ (2d: distance between the first directional microphone 10 and the second directional microphone 11) is generated between the first observation signal x1 (t) and the second observation signal x2 (t). . If the speed of sound is c, a time lag occurs by the time τ = 2dsinφ / c (see FIG. 5A). From the above, as shown in FIG. 5B, the discrete value x2 (k, m) of the second observation signal is delayed by the time τ, and the new discrete value x2 ′ (k, m, By setting m) = x2 (k, m−τ / Δt), the phase difference between the discrete value x1 (k, m) of the first observation signal and the discrete value x2 (k, m) of the second observation signal Can be corrected.

  In the present embodiment, as shown in FIG. 2B, since the target direction is the direction of the central axis A of the first directional microphone 10 and the second directional microphone 11, the sound source is directed to the target direction. Since there is no phase difference between the first observation signal x1 (t) and the second observation signal x2 (t) in the presence of S1, x2 ′ (k, m) = x2 (k, m). .

  The discrete values x1 (k, m) and x2 '(k, m) of the first and second observation signals corrected by the phase difference correction unit 20 are output to the sound source direction determination unit 21.

  The sound source direction determination unit 21 shown in FIG. 1 calculates an average value for a predetermined time (frame length) of the sum of squares of the discrete value x1 (k, m) of the first observed signal for each frame as shown in Equation 5. The first observed power value Px1 (k) is calculated, and the average value of the second sum of squares of the discrete value x2 ′ (k, m) of the second observed signal is calculated for each frame to obtain the second The observed power value is Px2 (k). In Equation 5, i = 1 and 2.

  The sound source direction determination unit 21 that has calculated the first and second observed power values Px1 (k) and Px2 (k) uses the first and second observed power values Px1 (k) and Px2 (k) to determine whether the sound source is It is determined whether or not it exists in the target direction.

  Here, as shown in FIG. 2B, in a region where the sound collection region of the first directional microphone 10 and the sound collection region of the second directional microphone 11 overlap (φ1 in FIG. 2B), The relative ratio between the first observed power value Px1 (k) and the second observed power value Px2 (k) is small, and the first observed power value Px1 in a region other than the above (for example, φ2 in FIG. 2B). The relative ratio between (k) and the second observed power value Px2 (k) increases.

Therefore, the sound source direction determination unit 21 shown in FIG. 1 has both the first and second observation power values Px1 (k) and Px2 (k) larger than the predetermined threshold value Pth, and the first observation power value Px1 (k ) And the second observed power value Px2 (k) | 10log ₁₀ (Px1 (k) / Px2 (k)) | (dB) is 0 or more and less than the predetermined set value ε (dB) (0 ≦ | 10log) _{When the condition of 10} (Px1 (k) / Px2 (k)) | <ε) is satisfied, it is determined that the sound source exists in the target direction. On the other hand, when the above condition is not satisfied, the sound source direction determination unit 21 determines that the sound source does not exist in the target direction.

  The output calculation unit 22 receives the first observation signal x1 (t) from the first directional microphone 10 and the second observation signal x2 (t) from the second directional microphone 11, and outputs a digital signal processor ( (Hereinafter referred to as “DSP”), the discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals from the first and second observation signals x1 (t) and x2 (t). ), The discrete value x2 (k, m) of the second observation signal is delayed by time τ, and the new discrete value x2 ′ (k, m) = x2 (k, m− τ / Δt), and the sum (x1 (k, m) + x2 ′ (k, m)) of the discrete values x1 (k, m) and x2 ′ (k, m) of the first and second observation signals is obtained. In this embodiment, since no delay occurs, x2 ′ (k, m) = x2 (k, m).

  When the sound source direction determination unit 21 determines that the sound source exists in the target direction, the output calculation unit 22 has a discrete value x1 (k, m) of the first observation signal and a discrete value x2 ′ of the second observation signal. A composite signal (x1 (k, m) + x2 ′ (k, m)) of (k, m) is amplified with a first amplification factor G1 and output as an output signal y (k, m) of a digital output signal. On the other hand, when the sound source direction determination unit 21 determines that the sound source does not exist in the target direction, the output calculation unit 22 performs the first amplification on the combined signal (x1 (k, m) + x2 ′ (k, m)). Amplification is performed at a second amplification factor G2 smaller than the factor G1, and the result is output as an output signal y (k, m).

  Next, the operation of the acoustic input device of this embodiment will be described with reference to FIG. First, when the first and second microphones 10 and 11 of the sound collection unit 1 receive a sound wave from a sound source, the first and second observation signals x1 (t) and x2 (t) are signal-processed from the sound collection unit 1. Transmitted to part 2. In the signal processing unit 2, the phase difference correction unit 20 extracts the discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals (S1, S2). Thereafter, the phase difference correction unit 20 shifts the phase difference of the second observation signal by the phase difference when the sound source exists in the target direction with respect to the discrete value x2 (k, m) of the second observation signal. A discrete value x2 ′ (k, m) is set (S3).

The sound source direction determination unit 21 calculates the first and second observation power values Px1 (k) and Px2 (k) for each frame (S4, S5). Thereafter, the sound source direction determination unit 21 examines whether or not the condition that both the first and second observation power values Px1 (k) and Px2 (k) are larger than the threshold value Pth is satisfied (S6). If the above condition is not satisfied, it is determined that the sound source is present in the direction other than the target direction (outside the sound collection area) (S9). On the other hand, when the above condition is satisfied, the sound source direction determination unit 21 examines whether or not the condition that the relative ratio | 10log ₁₀ (Px1 (k) / Px2 (k)) | is less than the threshold ε is satisfied (S7). If the above condition is satisfied, it is determined that the sound source exists in the target direction (in the sound collection area) (S8). If the above condition is not satisfied, it is determined that the sound source exists in a direction other than the target direction (S9).

  When it is determined that the sound source exists in the target direction, the output calculation unit 22 outputs G1 (x1 (k, m) + x2 ′ (k, m)) as the output signal y (k, m) (S10), When it is determined that the sound source does not exist in the target direction, G2 (x1 (k, m) + x2 ′ (k, m)) is amplified and output as the output signal y (k, m) (S11).

  As described above, according to the present embodiment, the observation signals (first observation signal x1 (k, m), second observation) from the two microphones (first directional microphone 10 and second directional microphone 11). Signal x2 (k, m)) is used to determine whether the sound source exists in the target direction, and the amplification factor (first amplification factor G1) when the sound source exists in the target direction By making it larger than the amplification factor (second amplification factor G2) when it does not exist in the direction, the sound from the sound source existing in the target direction can be emphasized and output.

Further, by evaluating the relative ratio | 10log ₁₀ (Px1 (k) / Px2 (k)) | between the first observed power value Px1 (k) and the second observed power value Px2 (k), the sound source It can be accurately determined whether or not it exists in the direction.

  Furthermore, by setting the two microphones 10 and 11 so that the sound collecting areas partially overlap each other, sharper directivity can be obtained as compared with a single directional microphone.

  In addition, the directivity angles α1 and α2 of the first and second directional microphones 10 and 11 of the first embodiment may have various directivity angles. Specifically, it is not limited to the one shown in FIG. 2, but as shown in FIG. 7 (with a directivity angle of 90 °) or as shown in FIG. 8 (with a directivity angle of 180 °). But you can. The same applies to the following embodiments.

  As a modification of the first embodiment, the output calculation unit 22 illustrated in FIG. 1 does not output the output signal y (k, m) of the digital signal, but the first observation signal x1 (t) 2 (x1 (t) + x2 ′ (t)) is amplified with a new second observation signal x2 ′ (t) = x2 (t−τ) obtained by delaying the second observation signal x2 (t) by time τ. May be output. That is, when it is determined that the sound source exists in the target direction, the output calculation unit 22 outputs the output signal y (t) = G1 (x1 (t) + x2 ′ (t)), and the sound source exists in the target direction. When it is determined not to output, the output signal y (t) = G2 (x1 (t) + x2 ′ (t)) (G2 <G1) is output. The same applies to the following embodiments.

(Embodiment 2)
In the acoustic input device according to the second embodiment, the sound source direction determination unit 21 illustrated in FIG. 1 has the cross-correlation function R ₁₂ of the discrete values x1 (k, m) and x2 ′ (k, m) of the first and second observation signals. The difference from the sound input device of the first embodiment is that (Equation 6) is calculated for each frame. The correlation function R _{ij in Equation} 6 is an autocorrelation function when i = j, and a cross-correlation function when i ≠ j. In addition, about the component similar to Embodiment 1, the same code | symbol is attached | subjected and description is abbreviate | omitted.

As shown in FIG. 9, the cross-correlation function R ₁₂ is large in an area where the sound collection areas of the first directional microphone 10 and the second directional microphone 11 overlap (−θ / 2 <φ <θ / 2). It takes a value and becomes a small value in other areas than the above. Therefore, the sound source direction determination unit 21 determines that the sound source exists in the target direction when the condition that the cross-correlation function R ₁₂ > threshold Rth is satisfied, and determines that the sound source does not exist in the target direction when the above condition is not satisfied. judge.

The determination of the threshold value Rth, the discrete values x1 (k, m) of the first observation signal discrete value x2 (k, m) of the autocorrelation function _{R 11} or the second observation signal of the autocorrelation function _{R 22} of Used. Since the autocorrelation functions R ₁₁ and R ₂₂ are large values regardless of the direction in which the sound source exists, Rth = ζR _ii (i = 1, 2) and 0 <ζ <1 are determined.

As described above, according to this embodiment, the discrete values x1 (k, m) of the first observation signal discrete value x2 '(k, m) of the second observation signal by evaluating the cross-correlation function R ₁₂ in It is possible to accurately determine whether or not the sound source exists in the target direction.

(Embodiment 3)
As shown in FIG. 10A, the acoustic input device according to the third embodiment is installed in the sound collection unit 1a at an intermediate position between the first directional microphone 10 and the second directional microphone 11, and the sound wave from the sound source. Is different from the acoustic input device of the first embodiment (see FIG. 1) in that it includes a non-directional microphone 12 that generates a reference observation signal that is an electric signal based on the received sound wave. In addition, about the component similar to Embodiment 1, the same code | symbol is attached | subjected and description is abbreviate | omitted.

  The omnidirectional microphone 12 of the present embodiment uses the one having the same sensitivity as that of the first and second directional microphones 10 and 11, but it is based on the received sound wave even if the sensitivity is different. The observation signal corrected with the sensitivity coefficient may be used as a new observation signal x0 (t). Specifically, when the sensitivity coefficient of the omnidirectional microphone 12 is m0, (the observation signal based on the sound wave received by the omnidirectional microphone 12 × m0) is the reference observation signal transmitted to the signal processing unit 2a. x0 (t) may be used. At this time, the sensitivity coefficient m0 is set to satisfy m0: m1 = √ (Px1): √ (Px0) using Px0 described later. The same applies to the following description.

  In the signal processing unit 2a of the present embodiment, as shown in FIG. 11, the phase difference correction unit 20a receives the reference observation signal x0 (t) from the omnidirectional microphone 12, and receives the received reference observation signal x0 (t). Discrete values x0 (k, m) are extracted at a predetermined time interval Δt, and the discrete reference values x0 (k, m) of the extracted reference observation signals are used as reference observation signals x0 (t) when a sound source exists in the target direction. And the first observation signal x1 (t) are shifted by the phase difference.

  In addition, the sound source direction determination unit 21a calculates an average value in a predetermined time (frame length L) of the square sum of the discrete values x0 (k, m) of the reference observation signal shifted by the phase difference correction unit 20a to perform the reference observation. The power value is Px0 (k). If the distance from the omnidirectional microphone 12 to the sound source is constant, the reference observation power value Px0 (k) is constant regardless of the direction of the sound source.

  The sound source direction determination unit 21a that has calculated the reference observation power value Px0 (k) has a relative ratio (Px1 (k) / Px0 (k) of the first observation power value Px1 (k) to the reference observation power value Px0 (k). ) Is included in the preset first range, and the relative ratio (Px2 (k) / Px0 (k)) of the second observed power value Px2 (k) to the reference observed power value Px0 (k) is preset. If the condition included in the second range is satisfied, it is determined that the sound source exists in the target direction. On the other hand, when the above condition is not satisfied, the sound source direction determination unit 21a determines that the sound source does not exist in the target direction.

  Here, the first range of the present embodiment is a range greater than the initially set threshold value δ and 1 or less (δ <Px1 (k) / Px0 (k) ≦ 1). The second range of the present embodiment is also a range greater than the initially set threshold value δ and 1 or less (δ <Px2 (k) / Px0 (k) ≦ 1).

  As described above, according to the present embodiment, the reference observation power value Px0 of the omnidirectional microphone 12 is provided by installing the omnidirectional microphone 12 at an intermediate position between the first directional microphone 10 and the second directional microphone 11. By using the first and second observed power values Px1 (k) and Px2 (k) for (k), it can be accurately determined whether or not the sound source exists in the target direction.

(Embodiment 4)
In the acoustic input device according to the fourth embodiment, the sound source direction determination unit 21a illustrated in FIG. 11 includes the first and second observation signals x1 (k, m), x2 (k, m) and the reference observation signal x0 (k, m). Is used to determine the correlation function R _ij (i, j = 0,1,2,2) expressed by Equation 7 and determine whether or not the sound source exists in the target direction using this correlation function R _ij Thus, it is different from the acoustic input device of the third embodiment. In addition, about the component similar to Embodiment 3, the same code | symbol is attached | subjected and description is abbreviate | omitted.

The correlation function R _{ij in Equation} 7 is an autocorrelation function when i = j, and a cross-correlation function when i ≠ j.

The sound source direction determination unit 21a of the present embodiment is different from the embodiment 3, the reference observation signal x0 (k, m) a reference observation signal x0 (k, m) for the autocorrelation function _{R 00} between the first observation signals x1 The relative ratio (R ₀₁ / R ₀₀ ) of the cross-correlation function R ₀₁ of (k, m) is included in the first preset range, and the reference observation signal x 0 (k, m) with respect to the autocorrelation function R ₀₀ When the relative ratio (R ₀₂ / R ₀₀ ) of the cross-correlation function R ₀₂ between the second observation signal x 2 (k, m) satisfies the condition included in the preset second range, the sound source is in the target direction. It is determined that it exists. On the other hand, when the above condition is not satisfied, the sound source direction determination unit 21a determines that the sound source does not exist in the target direction.

Here, the first range of the present embodiment is a range greater than the initially set threshold value ξ and 1 or less (ξ <R ₀₁ / R ₀₀ ≦ 1). The second range of the present embodiment is also a range greater than the initially set threshold value ξ and 1 or less (ξ <R ₀₂ / R ₀₀ ≦ 1).

As described above, according to the present embodiment, the autocorrelation function R _{00 of the} reference observation signal x 0 (k, m), the reference observation signal x 0 (k, m), and the first and second observation signals x 1 (k, m), x 2. By using the cross-correlation functions R ₀₁ and R ₀₂ of (k, m), the influence of uncorrelated noise can be reduced and a high S / N ratio can be achieved, and whether or not the sound source exists in the target direction. The determination accuracy can be increased.

(Embodiment 5)
In the acoustic input devices of the first and third embodiments, the output signal y (k, m) (the power value Py (k) of the output signal is not as shown in FIG. 10B) with respect to the direction in which the sound source exists. It will be continuous.

  Therefore, in the sound input device according to the fifth embodiment, when the sound source direction determination unit 21 illustrated in FIG. 1 determines that the sound source does not exist in the target direction, the sound input device determines whether the sound source exists in the vicinity of the target direction. On the other hand, when the output calculation unit 22 determines that the sound source exists in the vicinity of the target direction by the sound source direction determination unit 21, a continuous change that monotonously decreases between the first amplification factor G1 and the second amplification factor G2. The sound input device is different from the sound input device according to the first embodiment in that the synthesized signal is amplified and output with the amplification factor of the characteristic. In addition, about the component similar to Embodiment 1, the same code | symbol is attached | subjected and description is abbreviate | omitted.

As shown in FIG. 12, the amplification factor G is such that both the first and second observation power values Px1 (k) and Px2 (k) are larger than a predetermined threshold value Pth, and the relative ratio | 10log ₁₀ (Px1 (k) / Px2 (k)) | is 0 or more and less than the threshold value ε (0 ≦ | 10log ₁₀ (Px1 (k) / Px2 (k)) | <ε), as in the first embodiment, the first gain G1 Become.

On the other hand, when the relative ratio | 10log ₁₀ (Px1 (k) / Px2 (k)) | is equal to or greater than the threshold ε, first, the relative ratio | 10log ₁₀ (Px1 (k) / Px2 (k)) | When the threshold is less than ε1 (ε1> ε) (ε ≦ | ₁₀ log ₁₀ (Px1 (k) / Px2 (k)) | <ε), it is determined that the sound source exists in the vicinity of the target direction, and the gain G decreases monotonously. It varies according to the function f. Subsequently, when the relative ratio | 10log ₁₀ (Px1 (k) / Px2 (k)) | is equal to or greater than the threshold ε1 (ε1 ≦ | 10log ₁₀ (Px1 (k) / Px2 (k)) |), the sound source is in the target direction. Is determined to be present at a distant position, and the amplification factor G becomes the second amplification factor G2.

  As described above, according to the present embodiment, it is possible to prevent the amplification factor G for amplifying the combined signal (x1 (k, m) + x2 ′ (k, m)) from becoming discontinuous depending on the direction in which the sound source exists. As a result, the output signal y (k, m) (power value Py (k) of the output signal) can always be a continuous value (see FIG. 13).

As for the acoustic input device of the second embodiment, the amplification factor G is the first amplification factor G1 as in the second embodiment when the cross-correlation function R ₁₂ > threshold value Rth as shown in FIG. .

On the other hand, if it is not the cross-correlation function _{R 12>} threshold Rth, first, when the cross-correlation function _{R 12} is the threshold value Rth1 (Rth1 <Rth) greater than threshold value Rth following (Rth1 _{<R 12} ≦ Rth), the sound source is object direction It is determined that it exists in the vicinity, and the amplification factor G varies according to the monotonically increasing function e. Subsequently, when the cross-correlation function _{R 12} is 0 or larger than the threshold Rth1 less (0 ≦ _{R 12} ≦ Rth1), the sound source is determined to be present at a position distant from the destination direction, the gain G is a second amplification factor G2 It becomes. Note that the sensitivities of the first directional microphone 10, the second directional microphone 11, and the omnidirectional microphone 12 are all equal.

  In the acoustic input device according to the third embodiment, the amplification factor G has a relative ratio (Px1 (k) / Px0 (k)) and a relative ratio (Px2 (k) / Px0 (k) as shown in FIG. ) Is greater than the threshold value δ and equal to or less than 1 (δ <Px1 (k) / Px0 (k) ≦ 1 and δ <Px2 (k) / Px0 (k) ≦ 1), as in the third embodiment. A gain G1 of 1.

  On the other hand, when the relative ratio (Px1 (k) / Px0 (k)) and the relative ratio (Px2 (k) / Px0 (k)) are not within the above ranges, first, the relative ratio (Px1 (k) / Px0 (k) )) Or the relative ratio (Px2 (k) / Px0 (k)) is greater than the threshold δ1 (δ1 <δ) and less than or equal to the threshold δ (δ1 <Px1 (k) / Px0 (k) ≦ δ or δ1 <Px2 When (k) / Px0 (k) ≦ δ), it is determined that the sound source exists in the vicinity of the target direction, and the amplification factor G varies according to the monotonically increasing function g. Subsequently, either the relative ratio (Px1 (k) / Px0 (k)) or the relative ratio (Px2 (k) / Px0 (k)) is 0 or more and the threshold δ1 or less (0 ≦ Px1 (k) / Px0 (k) ) ≦ δ1 or 0 ≦ Px2 (k) / Px0 (k) ≦ δ1), it is determined that the sound source exists at a position away from the target direction, and the amplification factor G becomes the second amplification factor G2. Note that the sensitivities of the first directional microphone 10, the second directional microphone 11, and the omnidirectional microphone 12 are all equal.

In the acoustic input device according to the fourth embodiment, as shown in FIG. 16, the amplification factor G has a relative ratio (R ₀₁ / R ₀₀ ) and a relative ratio (R ₀₂ / R ₀₀ ) that are both higher than a threshold ξ. In the case of large 1 or less (ξ <R ₀₁ / R ₀₀ ≦ 1 and ξ <R ₀₂ / R ₀₀ ≦ 1), the first amplification factor G1 is obtained as in the fourth embodiment.

On the other hand, when the relative ratio (R ₀₁ / R ₀₀ ) and the relative ratio (R ₀₂ / R ₀₀ ) are not in the above ranges, first, the relative ratio (R ₀₁ / R ₀₀ ) or the relative ratio (R ₀₂ / R ₀₀ ) If either threshold ξ1 (ξ1 <ξ) greater than threshold value xi] or less _{(ξ1 <R 01 / R 00} ≦ ξ or _{ξ1 <R 02 / R 00 ≦} ξ) of determining a sound source is present in the vicinity of the target direction The amplification factor G varies according to the monotonically increasing function h. Subsequently, either the relative ratio (R ₀₁ / R ₀₀ ) or the relative ratio (R ₀₂ / R ₀₀ ) is 0 or more and the threshold ξ1 or less (0 ≦ R ₀₁ / R ₀₀ ≦ ξ 1 or 0 ≦ R ₀₂ / R ₀₀ ≦ In the case of ξ1), it is determined that the sound source exists at a position away from the target direction, and the amplification factor G becomes the second amplification factor G2. Note that the sensitivities of the first directional microphone 10, the second directional microphone 11, and the omnidirectional microphone 12 are all equal.

  As described above, even in the acoustic input device as in the third and fourth embodiments, the amplification factor G for amplifying the combined signal (x1 (k, m) + x2 ′ (k, m)) as in the fifth embodiment. Can be prevented from becoming discontinuous depending on the direction in which the sound source exists. As a result, unlike FIG. 10B, the output signal y (k, m) (power value Py (k) of the output signal) is obtained. It can always be a continuous value.

(Embodiment 6)
In the first to fifth embodiments, the case of one sound source has been described. In the sixth embodiment, the case of two sound sources will be described. Here, as a combination of a plurality of sound sources, (1) only a target sound source exists, (2) a target sound source and one noise source exist, and (3) a target sound source and two noise sources exist. (4) No target sound source and no noise source, (5) No target sound source and only one noise source, (6) No target sound source and two noise sources May exist. Among them, the combination of the two sound sources is the case of (2) and (6). Specifically, as shown in FIG. 17A, the target sound source S exists in an area where the sound collection areas of the first directional microphone 10 and the second directional microphone 11 overlap, and the first directional microphone. A case where the noise source N1 exists only in the ten sound collection regions will be described.

In the case shown in FIG. 17A, among the sound source direction determination conditions of the first embodiment, Px1 (k)> Pth (first condition) and Px2 (k)> Pth (second condition) are satisfied, The second observation signal x2 (t) of the second directional microphone 11 includes only a signal based on the target sound source S, whereas the first observation signal x1 (t) of the first directional microphone 10 includes Since not only the target sound source S but also a signal based on the noise source N1 is included, | 10log ₁₀ (Px1 (k) / Px2 (k) | <ε (third condition) is not satisfied.

As described above, when the first and second conditions are satisfied but the third condition is not satisfied, the sound source direction determination condition R ₁₂ > Rth (fourth condition) of the second embodiment is further used. The target sound source S exists in an area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, and the noise source N1 exists only in the sound collection area of the first directional microphone 10. However, since the signal based on the target sound source S is included in the observation signals x1 (t) and x2 (t) of both the first and second directional microphones 10 and 11, the fourth condition is satisfied. Therefore, the sound source direction determination unit 21 can determine that the target sound source S exists in an area where the sound collection regions of the first and second directional microphones 10 and 11 overlap.

  Subsequently, the sound source direction determination unit 21 needs to determine in which sound collection area the noise source N1 is present among the first and second directional microphones 10 and 11. Therefore, as shown in FIG. 17A, the omnidirectional microphone 12 is disposed between the first directional microphone 10 and the second directional microphone 11. Since the omnidirectional microphone 12 is omnidirectional, the reference observation signal x0 (t) includes a signal based on both the target sound source S and the noise source N1. Therefore, the sound source direction determination unit 21 among the observed power values Px1 (k) and Px2 (k) of the first and second directional microphones 10 and 11, δ <Pxi (k) / Px0 (k) ≦ 1 ( i = 1, 2) It is determined that the noise source N1 is present in the sound collection region of the first and second directional microphones 10 and 11 satisfying (fifth condition).

  When it is determined that the target sound source S exists in the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap and the noise source N1 exists only in the sound collection area of the first directional microphone 10, The output calculation unit 22 amplifies the discrete value x2 (k, m) of the second observation signal of the second directional microphone 11 including only the signal based on the target sound source S with the amplification factor G1, and outputs the output signal y (k , M) = G1 × x2 (k, m) is output.

  On the other hand, when it is determined that the target sound source S does not exist in the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, the output calculation unit 22 selects the first and second directional microphones 10 and 11. A composite signal (x1 (k, m) + x2 (k, m)) of the discrete values x1 (k, m) and x2 (k, m) of the 11 observation signals is amplified with an amplification factor G2 (G1> G2). Output signal y (k, m) = G2 (x1 (k, m) + x2 (k, m)) is output.

  As described above, according to the present embodiment, the target sound source S exists in the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, and the noise source N1 is the first and second directional microphones 10 and 11. Even when the signal exists only in one of the sound collection regions, only the signal based on the target sound source S can be amplified and output.

  As a modification of the sixth embodiment, the output signal y (k, m) can be output as follows. The sound source direction determination unit 21 includes the target sound source S in an area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, and the noise source N1 is present only in the sound collection area of the first directional microphone 10. If it is determined that it exists, the first and second directional microphones 10 and 11 are observed after the directional axis of the first directional microphone 10 is tilted until the third condition is satisfied as shown in FIG. A composite signal (x1 (k, m) + x2 (k, m)) of the discrete values x1 (k, m) and x2 (k, m) of the signal is amplified by the amplification factor G1, and the output signal y (k, m) = G1 (x1 (k, m) + x2 (k, m)) is output. The case where the target sound source S does not exist in the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap is the same as that in the sixth embodiment.

(Embodiment 7)
In the sixth embodiment, the case of two sound sources has been described. In the seventh embodiment, the case of three sound sources will be described. The combination of the three sound sources is (3) only when the target sound source and the two noise sources exist, but it is necessary to distinguish between the two sound sources ((2) and (6)). Specifically, as shown in FIG. 18A, the target sound source S exists in an area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, and the sound collection of the first directional microphone 10 is performed. A case where the noise source N1 exists only in the area and the noise source N2 exists only in the sound collection area of the second directional microphone 11 will be described.

In the case shown in FIG. 18A, among the sound source direction determination conditions of the first embodiment, Px1 (k)> Pth (first condition) and Px2 (k)> Pth (second condition) are satisfied. The first observation signal x1 (t) of one directional microphone 10 includes a signal based on the target sound source S and the noise source N1, and the second observation signal x2 (t) of the second directional microphone 11 is included in the first observation signal x1 (t). Includes a signal based on the target sound source S and the noise source N2, and thus | 10log ₁₀ (Px1 (k) / Px2 (k) | <ε (third condition) is not always satisfied.

As described above, when the first and second conditions are satisfied but the third condition is not satisfied, the sound source direction determination condition R ₁₂ > Rth (fourth condition) of the second embodiment is further used. The target sound source S exists in an area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, the noise source N1 exists only in the sound collection area of the first directional microphone 10, and the second Even when the noise source N2 exists only in the sound collection region of the directional microphone 11, the signal based on the target sound source S is the observation signal x1 (t of both the first and second directional microphones 10 and 11). ), X2 (t), the fourth condition is satisfied. Therefore, the sound source direction determination unit 21 can determine that the target sound source S exists in an area where the sound collection regions of the first and second directional microphones 10 and 11 overlap.

  Subsequently, the sound source direction determination unit 21 determines whether the noise source N1 is present in any one of the sound collection regions of the first and second directional microphones 10 and 11 as in the sixth embodiment, or in the present embodiment. Thus, it is necessary to determine whether different noise sources N1 and N2 exist in the sound collection areas of the first and second directional microphones 10 and 11, respectively. Therefore, as shown in FIG. 18A, the omnidirectional microphone 12 is disposed between the first directional microphone 10 and the second directional microphone 11. Since the omnidirectional microphone 12 is omnidirectional, the reference observation signal x0 (t) includes a signal based on the target sound source S and the noise sources N1 and N2, and the first directional microphone 10 includes the target sound source S and the noise source N1. Is included in the first observation signal x1 (t), and the second directional microphone 11 includes a signal based on the target sound source S and the noise source N2 in the second observation signal x2 (t). Therefore, the sound source direction determination unit 21 includes the omnidirectional microphone 12 and the first directivity when the noise sources N1 and N2 are outside the sound collection region of the first directional microphone 10 and the second directional microphone 11. The observed power values Px1 (k), Px2 (1) of the first and second directional microphones 10 and 11 with the ratio δ2 of the observed power values of the directional microphone 10 or the omnidirectional microphone 12 and the second directional microphone 11 as a threshold. k) satisfies δ2 <Pxi (k) / Px0 (k) ≦ δ (i = 1, 2) (fifth condition), respectively, the sound collecting areas of the first and second directional microphones 10 and 11 respectively. It is determined that different noise sources N1 and N2 exist.

  The target sound source S exists in an area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, and the noise sources N1 and N2 are collected by the first directional microphone 10 and the second directional microphone 11, respectively. When it is determined that the sound source exists only in the sound region, as shown in FIG. 18A, the noise sources N1 and N2 are separated from the central axis of the area where the sound collection regions of the first and second directional microphones 10 and 11 overlap. When the distance is greater than Max (α1 / 2, α2 / 2), as shown in FIG. 18B, the directional axes of the first and second directional microphones 10 and 11 are tilted until the third condition is satisfied. Only signals based on the target sound source S can be collected by the first and second directional microphones 10 and 11. At this time, the output calculation unit 22 is a composite signal (x1 (k, m) + x2) of the discrete values x1 (k, m) and x2 (k, m) of the observation signals of the first and second directional microphones 10 and 11. (K, m)) is amplified with an amplification factor G1, and an output signal y (k, m) = G1 (x1 (k, m) + x2 (k, m)) is output.

  On the other hand, when it is determined that the target sound source S does not exist in the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, the output calculation unit 22 selects the first and second directional microphones 10 and 11. A composite signal (x1 (k, m) + x2 (k, m)) of the discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals of 11 is amplified by a gain G2 (G1> G2). ) And output an output signal y (k, m) = G2 (x1 (k, m) + x2 (k, m)).

  As described above, according to the present embodiment, the target sound source S exists in the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap, and the first directional microphone 10 and the second directional microphone 11 are present. Even when different noise sources N1 and N2 exist in the respective sound collection regions, signals based on the target sound source S can be emphasized and output.

  As shown in FIGS. 19A and 19B, at least one of the noise source N1 and the noise source N2 is Max from the central axis of the area where the sound collection areas of the first and second directional microphones 10 and 11 overlap. If they are not separated by (α1 / 2, α2 / 2) or more, even if the directional axes of the first and second directional microphones 10 and 11 are tilted, only the signal of the target sound source S cannot be collected. However, the signal based on the target sound source S is included in both the first and second observation signals x1 (t) and x2 (t), and the signal based on the noise source N1 is only in the first observation signal x1 (t). Since the signal based on the noise source N2 is included only in the second observation signal x2 (t), a composite signal (x1) of the discrete values x1 (k, m) and x2 (k, m) of the first and second observation signals. By amplifying (k, m) + x2 (k, m)) with the amplification factor G1 and outputting the signal, the signal based on the target sound source S can be emphasized as compared with the case where the sound is collected by the omnidirectional microphone 12 alone. .

  Note that the sound input devices of Embodiments 6 and 7 perform the determination operation as shown in FIG. 20 to (1) when only the target sound source exists, or (2) when there is the target sound source and one noise source. (3) When there is a target sound source and two noise sources, (4) When there is neither a target sound source nor a noise source, (5) When there is no target sound source and there is only one noise source, (6 ) The case where there is no target sound source and there are two noise sources can be determined separately.

It is a block diagram which shows the structure of the acoustic input device of Embodiment 1,2. It is a figure for demonstrating operation | movement of an acoustic input device same as the above. It is a figure for demonstrating the characteristic of a 1st directional microphone and a 2nd directional microphone in the acoustic input device of Embodiments 1-8. It is a figure for demonstrating phase adjustment in an acoustic input device same as the above. It is a figure for demonstrating phase adjustment in an acoustic input device same as the above. 3 is a flowchart illustrating an operation of the sound input device according to the first embodiment. It is a figure for demonstrating the characteristic of the modification of a 1st directional microphone and a 2nd directional microphone in the acoustic input device of Embodiments 1-8. It is a figure for demonstrating the characteristic of the other modification of a 1st directional microphone and a 2nd directional microphone in the acoustic input device of Embodiments 1-8. FIG. 10 is a diagram for explaining the operation of the acoustic input device according to the second embodiment. It is a figure for demonstrating operation | movement of the acoustic input device of Embodiment 3. It is a block diagram which shows the structure of the acoustic input device of Embodiment 3,4. It is a figure which shows the amplification factor in the acoustic input device of Embodiment 5. It is a figure which shows the output signal in an acoustic input device same as the above. It is a figure which shows the gain in the acoustic input device of the modification same as the above. It is a figure which shows the gain in the acoustic input device of the other modification same as the above. It is a figure which shows the gain in the acoustic input device of the other modification same as the above. FIG. 10 is a diagram for explaining the operation of the sound input device according to the sixth embodiment. FIG. 10 is a diagram for explaining the operation of the acoustic input device according to the seventh embodiment. It is a figure for demonstrating operation | movement of an acoustic input device same as the above. 10 is a flowchart illustrating an operation of the acoustic input device according to the sixth and seventh embodiments.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,1a Sound collecting part 10 1st directional microphone 11 2nd directional microphone 12 Non-directional microphone 2, 2a Signal processing part 20, 20a Phase difference correction part 21, 21a Sound source direction determination part 22 Output calculation part

Claims (11)

Each of them has a sound collection area with a predetermined directivity angle, and is installed close to each other so that a part of each sound collection area includes a predetermined target direction, and receives sound waves from a sound source. A first directional microphone and a second directional microphone that generate an observation signal based on the received sound wave;
A first observation signal is received from the first directional microphone, a second observation signal is received from the second directional microphone, and a phase difference between the received first observation signal and the second observation signal is received. A phase difference correction unit that corrects only the phase difference between the first observation signal and the second observation signal when a sound source is present in the target direction,
A sound source direction determination unit that determines whether or not the sound source exists in the target direction, using the first observation signal and the second observation signal that have been corrected by the phase difference correction unit;
When the sound source direction determination unit determines that the sound source is present in the target direction, while at least one of the first observation signal and the second observation signal is amplified by a first amplification factor and output, When the sound source direction determination unit determines that the sound source does not exist in the target direction, the phase difference between the first observation signal and the second observation signal is corrected by the phase difference, and the corrected second An acoustic input device comprising: an output calculation unit that amplifies and outputs a combined signal of one observation signal and a second observation signal with an amplification factor smaller than the first amplification factor.   The sound source direction determination unit includes a first observation power value that is a power value of the first observation signal after being corrected by the phase difference correction unit, and a second value after being corrected by the phase difference correction unit. The condition that any of the second observation power values that are the power values of the observation signals is greater than a predetermined threshold and the relative ratio between the first observation power value and the second observation power value is less than a predetermined set value. The sound input device according to claim 1, wherein when the condition is satisfied, the sound source is determined to exist in the target direction, and when the condition is not satisfied, it is determined that the sound source does not exist in the target direction. The sound source direction determination unit uses the discrete value x1 of the first observation signal and the discrete value x2 of the second observation signal that have been corrected by the phase difference correction unit, and the correlation function R _ij ( i = 1, 2), and when the cross-correlation function R ₁₂ satisfies the condition that the cross-correlation function R ₁₂ is equal to or larger than a predetermined set value set based on the auto-correlation function R ₁₁ or the auto-correlation function R ₂₂ , the sound source The sound input device according to claim 1, wherein if the condition is not satisfied, it is determined that the sound source does not exist in the target direction.

An omnidirectional microphone installed at an intermediate position between the first directional microphone and the second directional microphone to receive a sound wave from the sound source and generate a reference observation signal based on the received sound wave;
The phase difference correction unit receives the reference observation signal, and calculates the phase difference between the received reference observation signal and the first observation signal, and the reference observation signal and the first observation when a sound source exists in the target direction. Correct only the phase difference of the signal,
The sound source direction determination unit includes a power value of the first observation signal after being corrected by the phase difference correction unit with respect to a reference observation power value that is a power value of the reference observation signal after being corrected by the phase difference correction unit. Is a power value of the second observation signal after being corrected by the phase difference correction unit with respect to the reference observation power value that is included in the first range set in advance. When the relative ratio of the second observed power values satisfies the condition included in the second range set in advance, the sound source is determined to exist in the target direction, and when the condition is not satisfied, the sound source is The sound input device according to claim 1, wherein the sound input device is determined not to exist in the target direction. An omnidirectional microphone installed at an intermediate position between the first directional microphone and the second directional microphone to receive a sound wave from the sound source and generate a reference observation signal based on the received sound wave;
The phase difference correction unit extracts a discrete value of the received first observation signal at a predetermined time interval, receives the received second observation signal, and calculates the discrete value of the second observation signal for the predetermined time. Extracting at an interval, receiving the reference observation signal, extracting a discrete value of the reference observation signal at the predetermined time interval, and calculating a phase difference between the extracted discrete value of the first observation signal and the second observation signal , Correcting only the phase difference between the first observation signal and the second observation signal when a sound source is present in the target direction, and calculating the phase difference between the discrete value of the extracted reference observation signal and the first observation signal, Correct only the phase difference between the reference observation signal and the first observation signal when there is a sound source in the target direction,
The sound source direction determination unit includes a discrete value x1 of the first observation signal after correction by the phase difference correction unit, a discrete value x2 of the second observation signal after correction by the phase difference correction unit, and the A correlation function R _ij (i, j = 0, 1, 2) represented by Equation 2 is obtained using the discrete value x0 of the reference observation signal after being corrected by the phase difference correction unit, and a mutual correlation with respect to the autocorrelation function R ₀₀ is obtained. When the relative ratio of the correlation function R ₀₁ is included in the first range set in advance and the relative ratio of the cross-correlation function R _{02 to} the autocorrelation function R ₀₀ satisfies the condition included in the second range set in advance The sound input device according to claim 1, wherein it is determined that the sound source is present in the target direction while the sound source is not present in the target direction when the condition is not satisfied.

The sound source direction determination unit determines whether the sound source exists in the vicinity of the target direction when it is determined that the sound source does not exist in the target direction.
When the sound source direction determination unit determines that the sound source exists in the vicinity of the target direction, the output calculation unit has a continuously changing characteristic that monotonously decreases from the first amplification factor as the distance from the target direction increases. The acoustic input device according to claim 1, wherein the synthesized signal is amplified and output with an amplification factor. An omnidirectional microphone installed at an intermediate position between the first directional microphone and the second directional microphone to receive a sound wave from the sound source and generate a reference observation signal based on the received sound wave;
The phase difference correction unit receives the reference observation signal, and calculates the phase difference between the received reference observation signal and the first observation signal, and the reference observation signal and the first observation when a sound source exists in the target direction. Correct only the phase difference of the signal,
The sound source direction determination unit includes a first observation power value that is a power value of the first observation signal after being corrected by the phase difference correction unit, and a second value after being corrected by the phase difference correction unit. Any of the second observation power values that are the power values of the observation signals is greater than a predetermined threshold, and the relative ratio between the first observation power value and the second observation power value is a predetermined first set value. In this case, the cross-correlation function between the first observation signal and the second observation signal is a predetermined value obtained from the autocorrelation function of the first observation signal or the autocorrelation function of the second observation signal. When the condition that is equal to or greater than the second set value is satisfied, it is determined that the sound source exists in each of the target direction and other regions other than the target direction, and when the condition is not satisfied, the sound source Judge that it does not exist in the direction,
When it is determined that the sound source is present in each of the target direction and other regions other than the target direction, the first relative to the reference observation power value that is the power value of the reference observation signal after being corrected by the phase difference correction unit. When the relative ratio of one observation power value is included in a preset range and the relative ratio of the second observation power value to the reference observation power value is not included in the range, the first directional microphone And the relative ratio of the second observation power value to the reference observation power value is included in the range, and the first observation power value relative to the reference observation power value is The sound input device according to claim 1, wherein when the relative ratio is not included in the range, it is determined that the sound source exists in a sound collection region of the second directional microphone.   When the sound source direction determination unit determines that the sound source exists in the sound collection region of the first directional microphone or the second directional microphone, the output calculation unit determines that the sound source is in the sound collection region. The acoustic input device according to claim 7, wherein an observation signal of a non-existing directional microphone is amplified by the first amplification factor and output. Drive means for tilting the directional axis of the first directional microphone or the second directional microphone;
The sound source direction determination unit determines that the sound source exists in one of the target direction and the sound collection region of the first directional microphone or the second directional microphone among the other regions. And controlling the driving means until a condition that a relative ratio between the first observed power value and the second observed power value is less than the first set value is satisfied,
When the sound source direction determination unit determines that the sound source exists in the sound collection region of the first directional microphone or the second directional microphone, the output calculation unit is A synthesized signal of the first observation signal and the second observation signal is amplified and output at the first amplification factor, while the sound source direction determination unit determines that the sound source does not exist in the target direction. 8. The acoustic input device according to claim 7, wherein the combined signal is amplified by a second amplification factor smaller than the first amplification factor and output. An omnidirectional microphone installed at an intermediate position between the first directional microphone and the second directional microphone to receive a sound wave from the sound source and generate a reference observation signal based on the received sound wave;
The phase difference correction unit receives the reference observation signal, and calculates the phase difference between the received reference observation signal and the first observation signal, and the reference observation signal and the first observation when a sound source exists in the target direction. Correct only the phase difference of the signal,
The sound source direction determination unit includes a first observation power value that is a power value of the first observation signal after correction by the phase difference correction unit, and a second after correction by the phase difference correction unit. When any of the second observed power values that are power values of the observed signals is larger than a predetermined threshold, the cross-correlation function between the first observed signal and the second observed signal is When the condition that is equal to or greater than a predetermined set value obtained from the autocorrelation function of the observation signal or the autocorrelation function of the second observation signal is satisfied, the sound source is in each of the target direction and other regions other than the target direction. On the other hand, when the condition is not satisfied, the sound source is determined not to exist in the target direction when the condition is not satisfied.
When it is determined that the sound source is present in each of the target direction and other regions other than the target direction, the first relative to the reference observation power value that is the power value of the reference observation signal after being corrected by the phase difference correction unit. When the relative ratio of the first observed power value and the relative ratio of the second observed power value to the reference observed power value are included in a preset range, the first directional microphone and the second directional microphone The sound input device according to claim 1, wherein the sound source is determined to exist in each of the sound collection regions of the directional microphone. Drive means for tilting the directional axis of the first directional microphone or the second directional microphone;
When the sound source direction determination unit determines that the sound source exists in both the target direction and the sound collection area of the first directional microphone and the second directional microphone among the other areas. The first observation power value when the two sound sources are separated from the target direction by more than one half of the maximum value of the directivity angle of the first directional microphone and the second directional microphone. And controlling the driving means until a condition that a relative ratio of the second observed power value is less than a predetermined set value is satisfied,
When the sound source direction determination unit determines that the sound source exists in the sound collection region of the first directional microphone or the second directional microphone, the output calculation unit is A synthesized signal of the first observation signal and the second observation signal is amplified and output at the first amplification factor, while the sound source direction determination unit determines that the sound source does not exist in the target direction. 11. The acoustic input device according to claim 10, wherein the synthesized signal is amplified by a second amplification factor smaller than the first amplification factor and output.

Images