JP5682116B2 - Apparatus and program for performing sound masking - Google Patents

Description

  The present invention relates to a technique for preventing sound leakage by using a masking effect.

  The masking effect is a phenomenon in which, for example, when two kinds of sounds are radiated in the same space, the presence of one sound prevents the other sound from being heard. Various sound masking devices have been proposed that use this masking effect to prevent the voice of a conference from leaking out from a conference room to the next room or from a partitioned conference space. This sound masking device emits a masker sound that hinders human speech from being heard outside a conference room or a conference space, for example, when a conference is held in the conference room or conference space. The masker sound in this case may be a noise sound having a spectrum in a wide band, but a sound having a spectrum envelope similar to a target sound to be prevented from listening (hereinafter referred to as a target sound) has a higher masking effect. It is known to be obtained. Patent Document 1 discloses a masker sound generation method used in a sound masking apparatus that uses such a human voice as a target sound. In the method disclosed in Patent Document 1, a masker sound is generated by performing a scramble process of dividing a speech waveform of a human voice into sections of one syllable and rearranging the divided sections. The masker sound obtained by this method has a sound quality similar to that of a human voice, but it is difficult to hear and understand the content of the story. Therefore, a high masking effect can be obtained by radiating this masker sound with the sound masking device.

JP 2008-233671 A

  However, in the sound masking device described above, for example, when the transmission path until the target sound reaches the listener is long, but the transmission path until the masker sound reaches the listener is short, it is compared with the target sound. The sound of the masker sound becomes clear. For this reason, there was a problem that the masker sound became annoying for the listener.

  This invention is made | formed in view of the situation demonstrated above, and it aims at providing the technical means which can prevent that a masker sound becomes annoying.

  In order to achieve the above object, the present invention provides a masking device for preventing a target sound generated in a first region from leaking into a second region and generating a masker sound that prevents the target sound from being heard. A masker sound signal generating unit for generating a masker sound signal used in the above, a masker sound signal generating unit and a speaker for emitting a masker sound to the second region, and the target sound is A clarity adjustment filter that performs a filter process for reducing the clarity of the masker sound signal by a degree corresponding to a reduction in clarity of the target sound that occurs in the process of leaking from the first area to the second area. A sound masking apparatus is provided. The present invention also provides a computer that generates a masker sound signal used to generate a masker sound that prevents the target sound generated in the first region from being heard in the second region, and the masker The target that is inserted between a sound signal generator and a speaker for emitting a masker sound in the second region, and that occurs in a process in which the target sound leaks from the first region to the second region Provided is a program that functions as an intelligibility adjustment filter that executes a filter process for reducing the intelligibility of the masker sound signal by a degree corresponding to a decrease in intelligibility of sound.

  According to this invention, since the clarity of the masker sound emitted to the second region approaches the clarity of the target sound, the masker sound is not annoying to the listener.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows the structure of the sound masking system containing the sound masking apparatus which is 1st Embodiment of this invention. It is a block diagram which shows the structure of the sound masking apparatus. It is a figure explaining the transfer function of the clarity adjustment filter which makes it possible to make the transfer function of the path | route of a target sound, and the transfer function of a masker sound path | route in the same embodiment the same. It is a figure which shows an example of the echo diagram of the detour path | route of the target sound used in the embodiment. It is a block diagram which shows the structural example of the clarity adjustment filter in the embodiment. It is a figure which illustrates the early reflection sound component and reverberation sound component which appear in an impulse response waveform. It is a block diagram which shows the structural example of the clarity filter used in 2nd Embodiment of this invention. It is a figure which shows the model of the space where the sound masking apparatus by 3rd Embodiment of this invention is installed.

<First Embodiment>
FIGS. 1A and 1B are diagrams showing a configuration of a sound masking system including a sound masking apparatus 10 according to the first embodiment of the present invention. This sound masking system is applied to a space composed of a conference space 110 and a passage 120 partitioned by a partition 100. FIG. 1 (a) is a view of a space composed of a conference space 110, a partition 100, and a passage 120 in which the sound masking system is installed, viewed from the horizontal direction, and FIG. 1 (b) is a view of the space viewed from above in the vertical direction. It is a figure.

  This sound masking system is a device that uses the voice of the speaker 210 in the conference space 110 as a target sound and prevents the target sound from being heard in the masking area 221 in the passage 120. In the present embodiment, an area where the passer's head frequently passes in the passage 120 is defined as a masking area 221.

  The left and right two-channel speakers 40L and 40R are fixed to the surface of the partition 100 on the side of the passage 120 with the speaker axis facing the masking area 221. These speakers 40 </ b> L and 40 </ b> R are speakers for emitting a masker sound that hinders listening to the target sound toward the masking area 221. The sound masking device 10 according to this embodiment is fixed to the surface of the partition 100 on the side of the conference space 110. The sound masking device 10 is a device that generates a masker sound signal for causing the speakers 40L and 40R to emit a masker sound. The reason why the sound masking device 10 is fixed to the surface of the partition 100 on the side of the conference space 110 is to make it easier for a person in the conference space 110 to operate the sound masking device 10.

  FIG. 2 is a block diagram showing the configuration of the sound masking apparatus 10 according to the present embodiment. As shown in FIG. 2, the sound masking device 10 includes a D / A converter 11, a phase adjuster 12, an EEPROM (Electronic Erasable Programmable Read Only Memory), and the like. The non-volatile memory 13, the memory I / F 14, the CPU 20, the RAM 21, the ROM 22, and the operation display unit 50 are included.

  The CPU 20 serves as a control center of the sound masking device 10 and executes various processes by executing programs stored in the ROM 22 while using the RAM 21 as a work area. The operation display unit 50 includes various operators such as buttons and volume knobs, and various displays such as an LCD (Liquid Crystal Display) and a lamp. Information indicating the contents of the operation is supplied to the CPU 20 and information is displayed on various displays according to instructions from the CPU 20.

  As a program stored in the ROM 22, there is a program that causes the CPU 20 to function as the masker sound signal generation unit 31 and the intelligibility adjustment filter 32. Here, the masker sound signal generator 31 is a means for generating a masker sound signal used to emit a masker sound that hinders the target sound from being heard. The intelligibility adjustment filter 32 performs a filtering process on the masker sound signal generated by the masker sound signal generation unit 31 and supplies the masker sound signal to the speakers 40L and 40R via the D / A converter 11 and the phase adjustment unit 12. It is. This intelligibility adjustment filter 32 makes the intelligibility of the masker sound so that the intelligibility of the masker sound emitted from the speakers 40L and 40R into the passage 120 approaches the intelligibility of the target sound leaking from the conference space 110 into the passage 120. Execute the filtering process to adjust More specifically, the intelligibility adjustment filter 32 reduces the intelligibility of the masker sound signal to the extent that the intelligibility of the target sound that occurs in the process of the target sound leaking from the conference space 110 to the passage 120 decreases. Execute.

  The non-volatile memory 13 stores masker sound data used by the masker sound signal generation unit 31 to generate a masker sound signal, and filter parameters used by the intelligibility adjustment filter 32 to execute filter processing.

  Here, masker sound data is data generated by the method disclosed in Patent Document 1, specifically, a speech waveform of a human voice is divided into sections corresponding to one syllable, and each section is divided. It is a sample data string of a waveform of a masker sound obtained by performing a scramble process for rearranging. In the present embodiment, a plurality of types of masker sound data obtained from the voices of various persons with different gender, age, language, etc. are stored in the nonvolatile memory 13. As described above, a plurality of types of masker sound data are stored in the non-volatile memory 13, and a sufficient masking effect can be expected according to the gender, age, language, etc. of the user of the conference space 110. This is to enable selection of maskable sound data. The masker sound signal generation unit 31 reads out a plurality of types of masker sound data stored in the non-volatile memory 13 and is designated by the operation of the operation display unit 50, and the clarity adjustment filter 32 as a masker sound signal. To hand over.

  The clarity adjustment filter 32 in the present embodiment is an FIR (Finite Impulse Response) filter. The filter parameter stored in the non-volatile memory 13 is a filter coefficient sequence that is convolved with the sample sequence of the masker sound signal by the articulation adjustment filter 32 that is the FIR filter. Details of this filter coefficient sequence will be described later. The memory I / F 14 is an interface that controls writing when masker sound data and filter parameters to be stored in the nonvolatile memory 13 are given from the outside.

  The masker sound signal that has undergone the processing of the intelligibility adjustment filter 32 is output from the CPU 20 to the D / A converter 11. The D / A conversion unit 11 converts the digital masker sound signals sequentially output from the CPU 20 into analog signals and outputs the analog signals to the phase adjustment unit 12. The phase adjustment unit 12 generates L channel and R channel masker sound signals from the analog masker sound signals, and outputs them to the speakers 40L and 40R. At that time, the phase adjustment unit 12 causes a phase difference of π between the L channel and R channel masker sound signals so that the masker sounds emitted from the speakers 40L and 40R become unfixed sounds.

  Next, details of the clarity adjustment filter 32 in the present embodiment will be described. As shown in FIGS. 1A and 1B, a typical sound receiving point 220 is assumed in the masking area 221. The sound receiving point 220 is, for example, a position where the head of the passerby can pass most easily in the masking area 221. As shown in FIGS. 1A and 1B, the target sound emitted by the speaker 210 in the conference space 110 is received in the masking area 221 via the detour path Ch1 passing between the partition 100 and the ceiling 230. The sound point 220 is reached. Here, the detour path Ch1 has a long path length, and the detour path Ch1 has a reflecting surface such as a ceiling 230 that reflects the target sound. For this reason, the target sound reaching the sound receiving point 220 includes a reflected sound from the ceiling 230 and the like, and has a lower degree of intelligibility than that at the time of occurrence in the conference space 110. On the other hand, the masker sound arrives at the sound receiving point 220 from the speakers 40L and 40R via the horizontal short path Ch2. For this reason, if no measure is taken, the masking area 221 has higher clarity of the masker sound than the target sound, and the clear masker sound becomes annoying. In order to avoid such an inconvenience, the intelligibility adjustment filter 32 reduces the intelligibility of the masker sound signal by a degree corresponding to the decrease in intelligibility of the target sound that occurs in the process in which the target sound leaks from the conference space 110 to the passage 120. Execute the filtering process. More details are as follows.

As shown in FIG. 3, the transfer function of the detour path Ch1 that the target sound follows until reaching the sound receiving point 220 is H1 (jω), and the transfer function of the path Ch2 that the masker sound follows is H2 (jω). Here, if the transfer function of the intelligibility adjustment filter 32 satisfies H3 (jω) satisfying the expression (1), the masker sound signal generated by the masker sound signal generation unit 31 becomes a masker sound and the sound receiving point. The transfer function of all paths until reaching 220 is H3 (jω) H2 (jω) = H1 (jω), which is the same as that of the detour path Ch1. Therefore, it can be expected that the clarity of the masker sound at the sound receiving point 220 is the same as the clarity of the target sound.
H3 (jω) = H1 (jω) / H2 (jω) (1)

In order to realize the articulation adjustment filter 32 having such a transfer function H3 (jω), a filter coefficient sequence used for the convolution calculation of the articulation adjustment filter 32 may be obtained as follows.
(1) An impulse sound is generated at a sound source position of a target sound in the conference space 110 (for example, a representative point of a speaker's mouth position in the conference space 110), and an impulse response waveform is collected at the sound receiving point 220.
(2) Fourier transform is performed on the impulse response waveform of (1) above to obtain the transfer function H1 (jω) of the detour path Ch1.
(3) Impulse sound is simultaneously generated at each position of the speakers 40L and 40R, and an impulse response waveform of the path Ch2 is collected at the sound receiving point 220.
(4) Fourier transform is performed on the impulse response waveform of (3) above to obtain the transfer function H2 (jω) of the path Ch2.
(5) Using the transfer function H1 (jω) in (2) and the transfer function H2 (jω) in (4), the calculation of the above equation (1) is performed, and the transfer function H3 ( jω).
(6) The inverse Fourier transform is performed on the transfer function H3 (jω) in (5) above to obtain a filter coefficient sequence used for the convolution calculation of the clarity adjustment filter 32.

  By the way, the intelligibility adjustment filter 32 in the present embodiment only needs to be able to bring the intelligibility of the masker sound close to the intelligibility of the target sound. It is not necessary to satisfy 1) with high accuracy. Therefore, a method of approximating the transfer function H3 (jω) of the necessary intelligibility adjustment filter 32 with the transfer function H1 (jω) of the detour path Ch1 along which the target sound follows can be considered. That is, the filter coefficient sequence obtained by sampling the impulse response waveform of the detour path Ch1 is stored in the nonvolatile memory 13 as a filter parameter and is used by the intelligibility adjustment filter 32. However, when this method is adopted, the filter coefficient used for the convolution calculation becomes a huge number, and the calculation amount of the convolution calculation becomes huge. Therefore, in the present embodiment, only information that affects the intelligibility of the target sound among all the information included in the impulse response waveform of the detour path Ch1 is used for the processing of the intelligibility filter 32. Specifically, an echo diagram that is an envelope of the impulse response waveform of the detour path Ch1 is obtained, and a filter coefficient sequence used for the convolution calculation of the articulation adjustment filter 32 is obtained from the echo diagram.

  FIG. 4 shows an example of an echo diagram of the detour path Ch1. As shown in this figure, when the generation time of the impulse sound is set to 0, the sample g (k) (k = 0 to n−1) of the echo diagram at each time T0 + kT (k = 0 to n−1) is obtained. Ask. Here, T0 is a delay time from the generation time of the impulse sound to the rise of the volume in the echo diagram, and corresponds to the propagation time of the impulse sound in the shortest path of the speaker 210 → the upper end of the partition 100 → the sound receiving point 220. To do. The sample g (k) (k = 0 to n−1) is used as a filter coefficient sequence used for the convolution calculation of the clarity adjustment filter 32.

  FIG. 5 is a block diagram illustrating a configuration of a clarity adjustment filter 32A that is an example of the clarity adjustment filter 32 according to the present embodiment. The clarity adjustment filter 32A includes a delay unit 321, n multiplication units 322, and an addition unit 323. The delay unit 321 generates n delayed masker sound signals obtained by delaying the masker sound signal supplied from the masker sound signal generating unit 31 by a delay time T0 + kT (k = 0 to n−1). The n multipliers 322 multiply the n delayed masker sound signals by filter coefficients g (k) (k = 0 to n−1), respectively. The adder 323 adds the multiplication results of the n multipliers 322 and outputs the result to the D / A converter 11 as a masker sound signal that has undergone intelligibility adjustment.

  According to the present embodiment, since such a clarity adjustment filter 32A is employed, the clarity of the masker sound at the sound receiving point 220 can be brought close to the clarity of the target sound. Therefore, it is possible to prevent the masker sound from becoming annoying.

Second Embodiment
In the present embodiment, as the information that affects the intelligibility of the target sound, the D value is selected from all the information included in the impulse response waveform of the detour path Ch <b> 1 and used for the arithmetic processing of the intelligibility adjustment filter 32. FIG. 6 is a diagram illustrating an example of an impulse response waveform of the detour path Ch1. As shown in FIG. 6, in the impulse response waveform, the reverberant sound component appears after the initial reflection component. The volume level of the reverberant component attenuates exponentially as shown in the figure. The D value is the ratio of the energy in the interval within the initial 50 ms of the impulse response waveform to the energy of the entire impulse response waveform, and is information that is often used as an index of speech intelligibility.

  FIG. 7 is a block diagram illustrating a configuration of a clarity adjustment filter 32B that is an example of the clarity adjustment filter 32 according to the present embodiment. As shown in FIG. 7, the articulation adjustment filter 32 </ b> B includes an initial reflected sound generation unit 330, a reverberation sound generation unit 340, multiplication units 351 and 352, and an addition unit 353.

  The initial reflected sound generator 330 is an FIR filter similar to the clarity adjustment filter 32 </ b> A described above, and includes a delay unit 331, a plurality of multipliers 332, and an adder 333. The initial reflected sound generator 330 is a circuit that receives a masker sound signal from the masker sound signal generator 31 and generates an initial reflected sound signal having a delay within 50 ms from the masker sound signal.

The reverberation sound generation unit 340 is a circuit that receives a signal obtained by delaying the masker sound signal by 50 ms or more from the delay unit 331 of the initial reflected sound generation unit 330 and generates a reverberation sound signal based on the delayed masker sound signal. . The reverberation sound generator 340 includes an IIR (Infinite) including an adder 341, a delay unit 342, an LPF (Low Pass Filter) 343, and a multiplier 344.
Impulse Response (infinite impulse response) filter. The adder 341 is a circuit that adds the sample of the masker sound signal supplied from the delay unit 331 of the initial reflected sound generator 330 and the output sample of the multiplier 344 and supplies the result to the delay unit 342. The delay unit 342 is a circuit that delays the sample supplied from the adder 341 by a predetermined time T _RT and supplies the sample to the LPF 343. The LPF 343 is a circuit that removes unnecessary high frequency components from the sample supplied from the delay unit 342 and supplies the high frequency component to the multiplier 344. Multiplication section 344 multiplies an attenuation coefficient _{K RT} in samples having passed through the LPF343, supplied to the adder 341. Here, the attenuation coefficient K _RT is a coefficient within a range satisfying the condition of 0 <K _RT <1. Then, the output signal of the LPF 343 is supplied to the multiplier 352 as a signal indicating a reverberant sound component.

In the reverberation sound generation unit 340, the sample of the masker sound signal (the sample of the masker sound signal that has been delayed by 50 ms) input to the addition unit 324 via the initial reflection sound generation unit 331 is referred to as the addition unit 341 → delay. The closed loop circuit of unit 342 → LPF 343 → multiplier 344 → adder 341 is repeated. Then, the sample of the masking sound signal, each time a round a closed loop, is multiplied by a damping coefficient K _RT. For this reason, the level of the sample of the reverberation sound component corresponding to one sample of the masker sound signal among the samples output from the LPF 343 decreases exponentially. When the reverberant component appearing in the output signal of the LPF343 is the reverberation time the time required to decay by a predetermined attenuation amount, the reverberation time is dependent on the damping coefficient K _RT, by increasing the damping coefficient K _RT If the attenuation coefficient _KRT is reduced, the reverberation time is increased. Furthermore, the reverberation time is dependent on the delay time of the delay unit 342 T _RT, the reverberation time is prolonged by increasing the delay time T _RT, the reverberation time by reducing the delay time T _RT is shortened. Therefore, in the present embodiment, the attenuation coefficient K _RT and the delay time T _RT that can realize the reverberation time comparable to the reverberation sound component appearing in the impulse response waveform of the detour path Ch1 are obtained, and the non-volatile is used as the filter parameter It is stored in the memory 13 and used for the filter processing (see FIG. 7) of the clarity adjustment filter 32B.

The multipliers 351 and 352 and the adder 353 constitute a mixer that mixes the output signal of the initial reflected sound generator 330 and the output signal of the reverberant sound generator 340 and outputs the mixed signal to the D / A converter 11. That is, the multiplication unit 351, a coefficient K _D by multiplying the initial reflected sound signals initial reflected sound generating section 330 outputs the multiplication unit 352, the coefficient 1-K _D reverberation sound signals reverberation sound generating section 340 outputs The addition unit 353 adds the multiplication result of the multiplication unit 351 and the multiplication result of the multiplication unit 352, and outputs a signal that is the addition result to the D / A conversion unit 11. Here, the coefficient _the K _D is a coefficient corresponding to the D value obtained from the impulse response waveform of the detour path Ch1, larger the D value increases, the coefficient _the K _D increases. This factor _the K _D, along with other parameters used in the calculation of the initial reflected sound generating section 330 and the reverberation sound generating unit 340, a parameter filter, stored in the nonvolatile memory 13, apparent from the nonvolatile memory 13 It is delivered to the degree adjustment filter 32C.

  According to the present embodiment, the initial reflection sound and the reverberation sound can be given to the masker sound, and the D value of the masker sound at the sound receiving point 220 can be made comparable to the D value of the target sound. Therefore, the clarity of the masker sound at the sound receiving point 220 can be brought close to the clarity of the target sound, and the masker sound can be prevented from becoming harsh. In the present embodiment, the reverberation time of the reverberation component of the masker sound at the sound receiving point 220 is set to be approximately the same as the reverberation time of the reverberation component of the target sound. Therefore, the clarity of the masker sound at the sound receiving point 220 can be made closer to the clarity of the target sound, and the masker sound can be prevented from becoming annoying.

In the present embodiment, the initial reflected sound generating unit 330 is not necessarily required to accurately reproduce the initial reflected sound component appearing in the impulse response waveform of the detour path Ch1, and the reverberant sound generating unit 340 is not limited to the detour path. It is not necessary to accurately reproduce the reverberation component appearing in the impulse response waveform of Ch1. In the present embodiment, it is to be if reflected in a mixing ratio K _D using D value obtained from the impulse response waveform of the detour path Ch1 to mixing of the initial reflected sound signal and reverberation signals.

In the present embodiment, a calibration knob may be provided in the operation display unit 50, and the attenuation coefficient K _RT and the delay time T _{RT in} the reverberation sound generator 340 may be increased or decreased by the operation of the calibration knob. . According to this aspect, the sound source position of the target sound in the conference space 110 assumed when the attenuation coefficient K _RT and the delay time T _RT stored in the nonvolatile memory 13 are determined, and the sound receiving point 220 of the masking area 221 When a deviation occurs between the sound source position of the actual target sound and the sound receiving point 220, and the attenuation coefficient _KRT and the delay time _TRT are not optimum values, the optimum values can be corrected.

<Third Embodiment>
In the present embodiment, various layouts of the conference space 110, the partition 100, and the passage 120 are assumed, and for each assumed layout, a filter parameter used for arithmetic processing of the clarity adjustment filter 32 is obtained and stored in the nonvolatile memory 13. Let Specifically, as shown in FIGS. 8A and 8B, the height H1 from the floor 240 to the ceiling 230 of the conference space 110 and the passage 120, the height H2 of the screen surrounding the conference space 110, the conference space A plurality of types of layouts with different floor areas S of 110 are assumed, and a filter parameter is obtained and stored in the nonvolatile memory 13 for each assumed layout. 8A is a diagram of the conference space 110 viewed from the horizontal direction, and FIG. 8B is a diagram viewed from the vertical direction. In the present embodiment, the operation display unit 50 is operated to select which filter parameter corresponding to which layout is used for the calculation processing of the clarity adjustment filter 32. Note that the clarity adjustment filter 32 may be any one of the above embodiments.

  According to this embodiment, one sound masking device 10 can be used in various layouts. Also in the present embodiment, a calibration knob may be provided on the operation display unit 50, and a parameter for arithmetic processing in the clarity adjustment filter may be increased or decreased by the operation of the calibration knob.

<Other embodiments>
Although the first to third embodiments of the present invention have been described above, other embodiments are conceivable for the present invention. For example:

(1) Assuming various sound source positions of the target sound in the conference space 110, filter parameters used for the arithmetic processing of the clarity adjustment filter 32 are obtained and stored in the nonvolatile memory 13 for each assumed sound source position. Further, the sound masking device 10 is provided with a function for measuring the sound source position of the target sound. Then, the sound masking device 10 is configured so that the filter parameters corresponding to the sound source position measured by the sound source position measuring function are read from the nonvolatile memory 13 and given to the intelligibility adjustment filter 32. Various modes for realizing the sound source position measurement function are conceivable. For example, a mode in which a pressure sensor is attached to each chair in the conference space 110 is conceivable. In this aspect, the CPU 20 of the sound masking apparatus 10 monitors the output signal of each pressure sensor attached to each chair in the conference space 110. When the speaker is seated on any chair, the CPU 20 detects the chair on which the speaker is seated based on the change in the output signal of the pressure sensor attached to the chair, and is the sound source position of the target sound. Find the position of the speaker's mouth. As another mode for realizing the sound source position measurement function, a mode using a so-called array microphone is also conceivable. In this aspect, the sound source position of the target sound in the conference space 110 is obtained from the time difference in the sound collection timing of each microphone. As another mode for realizing the sound source position measurement function, a mode in which a human sensor is installed at a specific position is also conceivable.

(2) The sound masking device 10 is provided with an impulse response measurement function of the detour path Ch1. Specifically, a sound source and a speaker that generate an impulse sound and a microphone that collects an impulse response waveform are added to the sound masking device 10. In addition, the sound masking device 10 has a function of calculating a filter parameter used for calculation processing of the articulation adjustment filter 32 from the impulse response waveform of the detour path Ch1 obtained by the impulse response measurement function and writing it to the nonvolatile memory 13. Make it. In this case, the clarity adjustment filter 32 may be any one of the above embodiments. According to the present embodiment, when the sound masking device 10 is applied to the conference space 110, the partition 100, and the passage 120 having various layouts, an optimum filter parameter corresponding to the application target is acquired, and the clarity adjustment filter is obtained. It can be used for 32 arithmetic processes. Therefore, there exists an advantage that the use of the sound masking apparatus 10 can be expanded.

(3) In each of the above embodiments, the sound masking device that generates the masker sound that prevents the target sound leaking from the conference space to bypass the partition is described, but the scope of application of the present invention is limited to this. It is not something. For example, in a situation where a plurality of conference spaces are adjacent to each other with a short screen between them, a sound masking device that prevents the target sound from leaking from each conference space to the next conference space may be configured. . In this case, since the sound source position of the target sound and the sound receiving point of the target sound are determined based on the position of the chair in each conference space, each path of the target sound and masker sound is determined, and the intelligibility filter can be easily designed. There are advantages. In each of the above embodiments, the target sound that arrives around the partition is masked. However, the sound masking device according to the present invention is provided between the sound source of the target sound and the ear of the listener receiving the target sound. It may be used to mask the target sound in a situation where there is no screen. More specifically, the target sound is transmitted from the speaker's mouth, which is the sound source, to the listener's ear, while the masker sound is transmitted from the speaker located away from the speaker's mouth, to the listener's ear. In this way, even in the absence of screens, the transmission path of the target sound and the transmission path of the masker sound are different, so if no measures are taken, the clarity of the masker sound will be higher than the clarity of the target sound. Arise. Therefore, the sound masking device according to the present invention generates a masker sound having intelligibility close to the target sound and masks the target sound.

(4) In the second embodiment, the initial reflection sound signal and the reverberation sound signal are generated from the masker sound signal, and the D value which is the mixing ratio of the sum of the masker sound signal and the initial reflection sound signal and the reverberation sound signal is adjusted. The clarity of the masker sound was adjusted. However, instead of doing this, only a reverberation sound signal may be generated from a masker sound signal, and the masker sound signal and the reverberation sound signal may be mixed and output as a masker sound. In this case, the clarity of the masker sound may be adjusted by adjusting the ratio of the energy of the masker sound signal to the energy of the reverberant sound signal (energy ratio of the direct sound component and the reverberant sound component) in the signal after mixing. That is, to increase the clarity of the masker sound, reduce the ratio of the energy of the reverberant sound signal to the energy of the masker sound signal in the mixed signal, and to reduce the clarity of the masker sound, In this case, the ratio of the energy of the reverberant sound signal to the energy of the masker sound signal is increased. Since this embodiment does not generate the initial reflected sound signal, there is an advantage that the calculation amount is small compared with the second embodiment.

(5) For example, a situation in which the sound masking device 10 is arranged at each base such as a conference space existing in a building is conceivable. In such a situation, in order to perform sound masking at each site such as a conference space, the masking sound signal generation unit 31 and the clarity adjustment filter 32 optimized by the sound masking device 10 at each site are used. It is necessary to be in a state with Therefore, various programs for causing the computer to function as the masker sound signal generation unit 31 and the intelligibility filter 32 are stored in the management server, and the sound masking apparatus 10 performs a desired program according to the operation of the operation display unit 50. It is good also as a structure which downloads and executes.

(6) In the management system of (5) above, information necessary for the sound masking device 10 at each site to calculate the filter parameters of the clarity adjustment filter 32 (for example, the echo diagram of the detour path Ch1 in the first embodiment) A function is provided for collecting and transmitting the information to the management server. Further, the management server is provided with a function of calculating filter parameters of the clarity adjustment filter 32 based on information such as an echo diagram received from the sound masking device 10 and transmitting the calculated parameters to the sound masking device 10. In the sound masking device 10, the filter parameter received from the management server is stored in the nonvolatile memory 13 and used for the execution of the intelligibility adjustment filter 32.

(7) In the above aspects (5) and (6), a function for generating a masker sound signal (the CPU 20 in the above embodiment, a program for causing the CPU 20 to function as the masker sound signal generator 31 and the clarity adjustment filter 32, etc.) ) Is provided in the sound masking device 10 at each site in the building, and a masker sound signal is generated at each site. However, a management server that distributes masker sound signals to the sound masking device 10 at each site is provided in, for example, a central management room in a building, and the sound masking device 10 disposed at each site receives the masking sound signal received from the management server. May be a device having a simple configuration that only emits the sound as a masker sound. In this case, the management server may have a function of centrally controlling the operation of the sound masking device 10 at each site. For example, a control mode in which the entire sound masking device 10 is operated only during the working hours of the building can be considered. Further, as another mode, a mode in which the sound masking device 10 at one or a plurality of bases that are in operation is operated in accordance with the working mode at each base (such as two shifts / three shifts) is also conceivable. Further, a mode in which the management server is linked with, for example, a conference reservation system is also conceivable. Specifically, the degree of confidentiality of the conference can be registered in the conference reservation system together with the holding time of each conference room, and the management server sets each conference according to the holding time of the conference in each conference room. It is conceivable that the room sound masking apparatus 10 is operated and control is performed so that the degree of masking is in accordance with the degree of confidentiality registered for the conference. Thus, the embodiment of the present invention may have a form in which a plurality of sound masking devices 10 are centrally managed by one management server.

DESCRIPTION OF SYMBOLS 10 ... Sound masking apparatus, 11 ... D / A conversion part, 12 ... Phase adjustment part, 13 ... Non-volatile memory, 14 ... Memory I / F, 20 ... CPU, 21 ... RAM, 22 ... ROM, 31 ... Masker sound signal Generating unit, 32, 32A, 32B ... Clarity adjustment filter, 40 ... speaker, 50 ... operation output unit, 100 ... partition, 110 ... conference space, 210 ... speaker, 120 ... passage, 220 ... sound receiving point, 221 ... Masking area, 230 ... Ceiling, 240 ... Floor, 321, 331, 342 ... Delay part, 322, 332, 344, 351, 352 ... Multiplying part, 323, 333, 341, 353 ... Adder, 343 ... LPF, 330 ... Initial reflection generator, 340... Reverberation generator.

Claims (5)

In a sound masking device for preventing a target sound generated in a first region from being leaked to the second region,
A masker sound signal generating unit for generating a masker sound signal used for generating a masker sound that hinders listening to the target sound;
It is inserted between the masker sound signal generator and a speaker for emitting masker sound to the second region, and is generated in the process of leaking the target sound from the first region to the second region. An intelligibility adjustment filter that performs a filter process for approximating the intelligibility of the target sound to the intelligibility of a masker sound emitted to the second region, wherein the filter process includes: The transfer function of the transfer path from the area 1 to the sound receiving point in the second area and the transfer path from the speaker where the masker sound emits the masker sound to the sound receiving point in the second area A sound masking device, comprising: a clarity adjustment filter executed by a filter determined in accordance with a transfer function . The intelligibility adjustment filter adds an initial reflected sound component to the masker sound signal and outputs it as an initial reflected sound signal; and reverberant sound generation that generates a reverberant sound signal based on the masker sound signal And a mixer that mixes and outputs the initial reflection sound signal generated by the initial reflection sound generation means and the reverberation sound signal generated by the reverberation sound generation means, the mixer from the first region The sound masking device according to claim 1, wherein the initial reflected sound signal and the reverberant sound signal are mixed with a mixing ratio determined according to the clarity of the target sound leaking into the second region. . The intelligibility adjustment filter includes a reverberation sound generation unit that generates a reverberation sound signal based on the masker sound signal, and a mixer that mixes and outputs the reverberation sound signal generated by the masker sound signal and the reverberation sound generation unit. And the mixer mixes the masker sound signal and the reverberant sound signal with a mixing ratio determined according to the clarity of the target sound leaking from the first region to the second region. The sound masking apparatus according to claim 1. Various heights from the floor to the ceiling of the first region and the second region, the height of the partition separating the first region and the second region, and the floor area of the first region Assuming a plurality of different layouts, the filter parameters of the clarity adjustment filter are stored in the memory for each assumed layout, and the filter parameters corresponding to the layout designated by the operation of the operation means are stored in the memory. The sound masking device according to any one of claims 1 to 3, wherein the sound masking device is used for filter processing of an intelligibility adjustment filter. Computer
A masker sound signal generating unit for generating a masker sound signal for generating a masker sound signal used for generating a masker sound that prevents the target sound generated in the first region from being heard in the second region;
It is inserted between the masker sound signal generator and a speaker for emitting masker sound to the second region, and is generated in the process of leaking the target sound from the first region to the second region. An intelligibility adjustment filter that performs a filter process for approximating the intelligibility of the target sound to the intelligibility of a masker sound emitted to the second region, wherein the filter process includes: The transfer function of the transfer path from the area 1 to the sound receiving point in the second area and the transfer path from the speaker where the masker sound emits the masker sound to the sound receiving point in the second area An intelligibility filter executed by a filter determined in response to the transfer function;
Program to make it work.

Images