JP5644382B2 - Audio processing device - Google Patents

Description

  The present invention relates to a sound processing apparatus that generates a masker sound for masking sounds in a plurality of areas for each area and emits the sound toward each area.

  Conventionally, when a certain sound (target sound) is heard and there is another sound with an acoustic characteristic (such as frequency characteristics) close to the target sound, the phenomenon (masking effect) that the target sound becomes difficult to hear is generally known Are known. A technique called so-called speech privacy has been proposed in which the masking effect is used to prevent the contents of conversation from leaking to a third party (see, for example, Patent Document 1). The technique described in Patent Document 1 generates a masker sound based on a sound signal collected by a microphone (hereinafter referred to as a microphone) in one acoustic space, and emits the sound from another speaker to another acoustic space. Thereby, the content of the conversation in one acoustic space can be made difficult to be heard by a listener in another acoustic space.

  Patent Document 1 can be used to prevent a third party waiting for a turn from listening to the content being conversed at the counter in, for example, a counter service. In this case, the voice of the speaker is picked up at the window, and the generated masker sound is emitted to a third party. Thereby, even if the voice of the conversation leaks from the window, the third party listens with the masker sound and cannot understand the contents of the conversation at the window.

Japanese Patent No. 42456060

  However, since Patent Document 1 emits a masker sound only in one direction, when the above-mentioned illustrated window is adjacent, if the masker sound is released to a third party, the masker is not The sound cannot be emitted. For this reason, the content of the conversation is asked to the adjacent speaker.

  SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a speech processing apparatus that can emit an accurate masker sound for each area even if there are a plurality of areas that should not be disclosed to the third party. There is.

  A speech processing apparatus according to the present invention is a speech processing apparatus that emits masker sounds masking sounds generated in a plurality of regions from a plurality of sound emitting means, and includes a plurality of sound collecting means, a generating means, and a signal Distribution means. The sound collection means collects sounds generated in a plurality of areas. The generating means generates a plurality of masker sounds each masking the sound collected by the sound collecting means. The signal distribution unit generates a plurality of synthesized signals obtained by synthesizing each masker sound generated by the generating unit at a predetermined ratio, and distributes and outputs each synthesized signal to the plurality of sound emitting units. The sound emitting means emits sound based on the synthesized signal in each of a plurality of regions. The ratio of the signal distribution means is set so that the masker sound corresponding to the sound collected in the area emitted by the sound output means for distributing and outputting becomes equal to or lower than a predetermined level.

  In this configuration, a masker sound is generated for each area, and the sound emission direction of each generated masker sound is set to other than the corresponding area, and the sound emission direction of the synthesis of each masker sound is the sound collection area. It is set to other than. As a result, the sound of the speaker in each area reaches the other area together with the masker sound related to each area, so that the content spoken in each area becomes difficult to hear in the other area. In addition, it is difficult for the third listener who is outside the picked-up area to hear sounds in all the areas.

  An audio processing apparatus according to the present invention includes a speaker array in which a plurality of speakers are arranged, and a delay supply unit that distributes the synthesized signal to each of the plurality of speakers and supplies the synthesized signals with a delay. The plurality of sound emitting means are realized by the delay supply means giving different delay amounts for each of the plurality of areas.

  In this configuration, an example is shown in which a speaker array is used as a speaker that emits masker sounds. Since the speaker array can control the directivity of the sound to be emitted, when it is installed as a single speaker, it is necessary to install each speaker in the direction in which the sound should be emitted. This reduces the burden of installation.

  In the sound processing apparatus according to the present invention, the delay supply means gives a delay amount so that the sound source position of the masker sound corresponding to the region is set in the direction of the region corresponding to the collected sound.

  In this configuration, the sound source position of the masker sound corresponding to the sound picked up in one area is the one area. As a result, the listener who hears the masker feels that the sound has been output from one area. As a result, since the sound leaking from one region and the masker sound are heard from substantially the same direction, it is possible to prevent the listener from hearing the sound.

  The speech processing apparatus according to the present invention includes a microphone array in which a plurality of microphones are arranged, and delay synthesis means for adding and synthesizing the sounds collected by the plurality of microphones. The plurality of sound collecting means are realized by the delay synthesizing means giving different delay amounts for each of the plurality of areas.

  In this configuration, an array microphone is used as sound collection means. Since the array microphone can control the directivity in the sound collecting direction, it is possible to save the trouble of re-installing the microphone even when the sound collecting area is uncertain.

  According to the present invention, even if there are a plurality of areas that should not let the third party know the contents of the conversation, an accurate masker sound can be emitted for each area.

It is a schematic diagram for demonstrating the outline | summary of the masker sound emission apparatus which concerns on embodiment. It is a schematic diagram for demonstrating the positional relationship of a masker sound-emitting device and the sound collection position of each area | region. It is a block diagram which shows the structure of a masker sound emission device. It is a block diagram which shows typically the function of a signal processing part and a divider | distributor. It is a schematic diagram for demonstrating the formation principle of the directivity which a speaker array forms. It is a block diagram which shows typically the structure of a part of masker sound emission device at the time of using a speaker array. It is a flowchart which shows the procedure of the process performed with a speaker array apparatus. It is a schematic diagram which shows the directivity formation method at the time of setting a virtual sound source. The schematic diagram at the time of setting a virtual sound source with respect to each masker sound is shown. It is a figure which shows another example of arrangement | positioning of each speaker. It is a schematic diagram which shows the case where an audio | voice is reflected. It is a schematic diagram which shows the structure of an array microphone.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, a preferred embodiment of a sound processing device according to the invention will be described with reference to the drawings. In the embodiments described below, the sound processing device according to the present invention will be described as a masker sound emitting device capable of emitting sound with strong directivity in a specific direction.

  FIG. 1 is a schematic diagram for explaining an outline of a masker sound emitting device according to the present embodiment. The masker sound emitting device 1 according to the present embodiment is installed, for example, with respect to the areas 100 and 101 and other areas 102 in which dialogue is made. The regions 100 and 101 are formed at a certain interval. The area 102 is an area where one or a plurality of consultants waits, and is formed with an interval from the areas 100 and 101. In the following, a person in the area 100, 101, 102 is referred to as a listener.

  The masker sound emitting device 1 is connected to microphones 13A and 13B. The microphones 13 </ b> A and 13 </ b> B are installed on a desk or the like in the areas 100 and 101, for example, and collect voices of conversation in the areas 100 and 101. The masker sound emitting device 1 generates masker sounds A and B for masking the conversational voices in the areas 100 and 101 based on the voices picked up by the microphones 13A and 13B, respectively.

  Directional speakers 16A, 16B, and 16C are connected to the masker sound emitting device 1. The speaker 16 </ b> A is disposed so that the sound emission direction is directed toward the region 101, and the speaker 16 </ b> B is disposed so that the sound emission direction is directed toward the region 100. The masker sound emitting device 1 emits the generated masker sound A from the speaker 16 </ b> A to the region 101, and emits the masker sound B from the speaker 16 </ b> B to the region 100.

  As a result, the listener in the area 101 cannot grasp the content of the conversation in the area 100 because the voice leaking from the area 100 is masked by the masker sound A. On the other hand, the listener who is in the region 100 cannot grasp the content of the conversation in the region 101 because the voice leaking from the region 101 is masked by the masker sound B.

  The speaker 16 </ b> C is arranged so that the sound emission direction is directed to the region 102. The masker sound emitting device 1 generates a masker sound A + B by combining the masker sounds A and B, and emits the masker sound A + B from the speaker 16C to the region 102.

  As a result, the listener in the area 102 cannot grasp the content of the conversation in the areas 100 and 101 because the voice leaking from the areas 100 and 101 is masked by the masker sound A + B.

  Hereinafter, the masker sound emitting device 1 will be described in detail.

  FIG. 2 is a schematic diagram for explaining the positional relationship between the speakers 16A, 16B, and 16C that emit the masker sound generated by the masker sound emitting device 1 and the regions 100, 101, and 102. Each speaker 16A, 16B, 16C is a directional speaker. The speaker 16 </ b> A is set at the focal point F <b> 1 of the region 101 in the center of the sound emission direction. The speaker 16 </ b> B is set at the focal point F <b> 2 of the region 100 in the center of the sound emission direction. The speaker 16 </ b> C is set at the focal point F <b> 3 of the region 102 in the center of the sound emission direction.

  As a result, the masker sound A can be heard loudly in the region 101, and the masker sound B can be heard loudly in the region 100. Also, the listener in the area 102 can hear the sound leaking from the areas 100 and 101 and the masker sound A + B coming from substantially the same direction.

  If the masker sounds are heard from different directions, the voices of the areas 100 and 101 may be heard by the listener of the area 102 by the selective listening of the voice (so-called cocktail party effect), and the content of the conversation may be grasped. There is also. However, in this embodiment, since the masker sound can be heard from the same direction as the speaker, the cocktail party effect can be suppressed.

  FIG. 3 is a block diagram showing the configuration of the masker sound emitting device 1. The masker sound emitting device 1 includes a control unit 10, a storage unit 11, an operation unit 12, a voice input unit 13, a signal processing unit 14, and a distributor 15.

  The control unit 10 includes, for example, a CPU (Central Processing Unit), and controls the operation of each of the above-described units, thereby controlling the operation of the masker sound emitting device 1. The storage unit 11 is a ROM (Read Only Memory) or a RAM (Random Access Memory), for example, and stores necessary programs and various data. The operation unit 12 receives an operation by the user of the masker sound emitting device 1.

  The voice input unit 13 has an A / D converter (not shown) and is connected to microphones 13A and 13B. The voice input unit 13 A / D-converts each input voice signal from the microphones 13 </ b> A and 13 </ b> B that has picked up the voice of conversation in the areas 100 and 101, and outputs it to the signal processing unit 14.

  The signal processing unit 14 includes, for example, a DSP (Digital Signal Processor), and performs signal processing on each input audio signal. The distributor 15 distributes and outputs the input audio signal signal-processed by the signal processing unit 14 to the speakers 16A, 16B, and 16C. FIG. 4 is a block diagram schematically illustrating the function of the signal processing unit 14 and the configuration of the distributor 15. The signal processing unit 14 includes an echo cancellation unit 141 and masker sound generation units 142 and 143.

  The echo cancel unit 141 includes an adaptive filter and removes echo from the input audio signal. For example, when the masker sound A is emitted to the area 101, the microphone 13B collects the masker sound A, so that an echo is included in the input voice signal. Therefore, the echo canceling unit 141 generates a pseudo regression sound signal that is a pseudo signal of a component in which the emitted masker sound circulates into the microphones 13A and 13B, and subtracts the pseudo regression sound signal from the input voice signal, thereby Remove. Thereby, the signal processing unit 14 can accurately extract only the voice of the conversation from the input voice signal. Note that the method of removing echoes by the echo canceling unit 141 can be changed as appropriate.

  The masker sound generators 142 and 143 generate the masker sounds A and B described above based on the input voice signal from which the echo is removed by the echo canceler 141. For example, the masker sound generation units 142 and 143 may acquire voice data having acoustic characteristics similar to the collected voice as the masker sound from the voice data stored in advance, or the collected voice Voice data having approximate acoustic characteristics may be generated as a masker sound.

  The distributor 15 synthesizes the masker sounds generated by the masker sound generation units 142 and 143 at a predetermined ratio, and emits the sound (synthesized signal) emitted to each of the areas 100, 101, and 102. In the present embodiment, the synthesized signal is a signal of masker sounds A and B and a synthesized masker sound A + B (hereinafter referred to as masker sound signals A, B, and A + B). The distributor 15 distributes the masker sound signals A, B, A + B to the speakers 16A, 16B, 16C.

  As shown in FIG. 4, the distributor 15 includes amplifiers 151 to 156. The amplifiers 151 to 153 receive the masker sound signal A generated by the masker sound generator 142, respectively. A masker sound signal B generated by the masker sound generator 143 is input to each of the amplifiers 154 to 156. The distributor 15 includes adders 15A, 15B, and 15C. The adder 15A adds the signals from the amplifiers 151 and 154 and outputs them to the speaker 16A. The adder 15B adds the signals from the amplifiers 152 and 155 and outputs the result to the speaker 16B. Adder 15C adds the signals from amplifiers 153 and 156 and outputs the result to speaker 16C.

  In this distributor 15, the gains of the amplifiers 151, 152, and 153 are set to 1, 0 and 1/2, respectively. The gains of the amplifiers 154, 155, and 156 are set to 0, 1, 1/2, respectively. Thereby, only the masker sound signal A is output from the adder 15A, and only the masker sound signal B is output from the adder 15B. Further, the adder 15C outputs a masker sound signal A + B obtained by synthesizing the masker sound signal A and the masker sound signal B. Thus, the distributor 15 can output the masker sound signals A and B to the speakers 16A and 16B, respectively, and can also output the masker sound signal A + B obtained by synthesizing the masker sound signals A and B to the speaker 16C.

  The gain set in each of the amplifiers 151 to 156 is not limited to 0, 1, 1/2. For example, the gain may be set in each of the amplifiers 151 to 156 so that the masker sound A emitted to the region 100 is at a level that cannot be heard by the listener in the region 100. Further, the distributor 15 is not limited to the configuration shown in FIG. 4 and can be appropriately changed depending on the number of input signals and the number of output signals. The number of signals to be input is determined by the number of microphones, and the number of signals to be output is determined by the number of speakers. For example, if the number of microphones is “m” and the number of speakers is “n”, an m × n amplifier may be arranged and an adder capable of a predetermined composition ratio may be provided.

  Note that the audio input unit 13 and the distributor 15 may be included in the signal processing unit 14. Further, the function and operation of the signal processing unit 14 may be realized by software by causing the control unit 10 to execute a program stored in the storage unit 11.

  The speaker that emits masker sounds A, B, A + B may be a speaker array including a plurality of speakers arranged in a matrix, honeycomb, or line. The speaker array can emit sound having strong directivity in a specific direction. In this case, the control unit 10 sets the focal point of the sound emitted from the speaker array. For example, as described in FIG. 2, the control unit 10 sets the focal point F <b> 1 for the voice A so that the voice A covers the area 101, the focal point F <b> 2 for the voice B, and the voice B covers the area 100. Thus, the focal point F3 for the sound A + B is set in the area 102. The control unit 10 may set the focal points F1, F2, and F3 at preset positions, store the positions of the installed microphones 13A and 13B in advance, and set the focal points F1 and F2 to the microphones 13A and 13A. Alternatively, the focal point F3 may be set at a position that is equidistant from both positions of the microphones 13A and 13B.

  FIG. 5 is a schematic diagram for explaining the principle of directivity formed by the speaker array. The speaker array includes a speaker unit 17 in which a plurality of speakers 171 to 17n are arranged in a matrix, honeycomb, or line on one panel.

  When the same audio signal is input to each of the speakers 171 to 17n, the sound emitted from each of the speakers 171 to 17n can be simultaneously reached by adjusting the sound emission timing of the sound from each of the speakers 171 to 17n. To do. As a result, as shown by the dotted lines in the figure, the speakers 171 to 17n are arranged at a predetermined interval on the circumference centered on the focal point F, and at the same time, the same sound is emitted. As a result, the sound is emitted from the speaker unit 17 while having strong directivity in the direction of the focus F by the superposition principle.

  Using this principle, the masker sound emitting device 1 can emit the sounds A, B, A + B in the respective directions shown in FIG. As described above, when the speaker array is used, the sound emission direction of the masker sound can be changed. Therefore, the positional relationship between the masker sound emission device 1 and each of the regions 100, 101, and 102 is limited. Without this, the user can freely install the masker sound emitting device 1.

  FIG. 6 is a block diagram schematically showing a partial configuration of the masker sound emitting device 1 when a speaker array is used. When the speaker array is used, the masker sound emitting device 1 includes directivity forming units 181 to 183 that form directivities of the masker sounds A, B, and A + B between the distributor 15 and the speaker unit 17, and directivity. It is necessary to include adders 211 to 21n that synthesize output signals from the sex forming units 181 to 183. The number of directivity forming units 181 to 183 corresponds to the number of regions 100, 101, and 102 (the number of masker sounds). The number of adders 211 to 21n corresponds to the number of speakers 171 to 17n.

  Masker sound signals A, B, A + B are supplied to the directivity forming units 181 to 183 by the distributor 15 described above. Since the directivity forming units 181 to 183 have the same configuration, the directivity forming unit 181 will be described below. The directivity forming unit 181 includes a delay unit 20 and amplifiers 201 to 20n. The number of amplifiers 201 to 20n corresponds to the number of speakers 171 to 17n.

  The control unit 10 calculates the timing (delay amount) of sound emission from the speakers 171 to 17n so that the masker sound is emitted to each set focal point. The delay unit 20 distributes the input masker sound signal with a delay amount calculated by the control unit 10 to each of the connected amplifiers 201 to 20n. The amplifiers 201 to 20n adjust the gain of the output signal of the delay unit 20. The output signal of the delay unit 20 is D / A converted by a D / A converter (not shown) and input to the amplifiers 201 to 20n.

  The adders 211 to 21n add the output signals of the directivity forming units 181 to 183 and output the added signals to the speakers 171 to 17n of the speaker unit 17. The speaker unit 17 emits sound based on the signals output from the adders 211 to 21n. Thereby, as in the case shown in FIG. 1, the sounds A, B, A + B are emitted in each direction, and the sound collecting means of the present invention is realized. The output signals from the adders 211 to 21n are output to the speakers 171 to 17n after the amplitude (volume) is adjusted to an optimum value by a power amplifier (not shown).

  Next, the operation of the masker sound emitting device 1 will be described. FIG. 7 is a flowchart showing a procedure of processing executed by the masker sound emitting device 1.

  In the masker sound emitting device 1, the control unit 10 determines whether or not at least one of the microphones 13A and 13B has collected the voice of the conversation in the areas 100 and 101 (S1). When the control unit 10 determines that no sound is collected (S1: NO), the control unit 10 ends this process. When the control unit 10 determines that the voice has been collected (S1: YES), the signal processing unit 14 generates a masker sound (S2). When the microphone 13A picks up sound, the signal processing unit 14 generates a masker sound A. When the microphone 13B collects sound, the signal processing unit 14 generates a masker sound B. Then, the signal processing unit 14 synthesizes masker sounds A and B to generate a masker sound A + B (S3).

  Next, the control unit 10 outputs the generated masker sound signals A, B, and A + B to the speakers 16A, 16B, and 16C, and outputs the masker sounds A, B, and A + B from the speakers 16A, 16B, and 16C to the regions 100 and 101, respectively. , 102 (S4). Then, this process ends.

  Although the preferred embodiment has been described above, the specific configuration and the like of the masker sound emitting device 1 can be appropriately changed in design, and the actions and effects described in the above-described embodiment are the most preferred resulting from the present invention. However, the functions and effects of the present invention are not limited to those described in the above embodiment.

  For example, in the above-described embodiment, when the speaker array is used, the focal points F1 and F2 may be set so that the masker sounds A and B are emitted in that direction. You may comprise so that the sound A and B may be emitted so that it may sound as if it came from the position of the area | region 100,101.

  FIG. 8 is a schematic diagram showing a directivity forming method when a virtual sound source is set. When the sound is emitted from the virtual sound source 17 ′ set in the space, the sound emission timing of the sounds from the speakers 171 to 17 n is adjusted so that the sound wave front is formed. Thereby, the listener U feels as if sound has been emitted from the virtual sound source 17 ′.

  Using this principle, the control unit 10 emits sound from the speakers 171 to 17n so that the virtual sound source of the masker sound A is in the vicinity of the region 100 and the virtual sound source of the masker sound B is in the vicinity of the region 101. Timing (delay amount) is calculated.

  FIG. 9 is a schematic diagram when a virtual sound source is set for each masker sound. As shown in FIG. 9, the listener in the region 100 (or 101) feels that the masker sound B (or A) can be heard from the sound source 101 '(or 100') near the region 101 (or 100). As a result, the cocktail party effect can be prevented and the masking effect can be enhanced.

  Note that the virtual sound source preferably coincides substantially with the areas 100 and 101, but may be as close to the areas 100 and 101 as possible from the relationship between the function and the installation position of the masker sound emitting device 1.

  Further, when the masker sound emitting device 1 is a speaker array, the speakers 171 to 17n of the speaker unit 17 are arranged on a plane in FIG. 5, but the shape can be changed as appropriate. FIG. 10 is a diagram illustrating another arrangement example of the speakers 171 to 17n. As shown to Fig.10 (a), the structure which arrange | positions each speaker 171-17n in the shape of a mortar-like cross section may be sufficient. Moreover, as shown in FIG.10 (b), the structure which has arrange | positioned each speaker 171-17n in circular arc shape may be sufficient. In the case of FIG. 10, the virtual sound source can be as close to the regions 100 and 101 as possible.

  In addition, partition plates (partitions) may be installed in the areas 100 and 101, and the sounds A and B may be reflected by the partitions and then emitted to the areas 101 and 100. FIG. 11 is a schematic diagram showing a case where sound is reflected. FIG. 11 shows a case where a partition 105 is installed in the area 100.

  In this case, it can be virtually as if the sound source of masker sound A exists at the position where the sound A is reflected by the partition 105. For this reason, since the listener in the area 101 can hear the masker sound A and the conversation voice in the area 100 from substantially the same direction, it is difficult to grasp the contents of the conversation in the area 100.

  Note that the installation angle of the partition 105 may be adjusted so that the sound A is emitted to the area 101, or the focus (reflection position) of the sound A on the partition 105 may be adjusted. .

  In the above-described embodiment, the microphones 13A and 13B are provided in the regions 100 and 101, respectively, but an array microphone can be used instead of the microphones 13A and 13B. FIG. 12 is a schematic diagram showing the configuration of an array microphone that replaces the microphone 13A (or microphone 13B). The audio input unit 13 performs A / D conversion on audio signals input from the plurality of microphones M1 to Mn forming the microphone array, and adds after adding predetermined delays by the delay units (delay synthesis means) 301 to 30n. Is synthesized by a unit (delay synthesis means) 31. This signal becomes the input audio signal. The voice input unit 13 can set the directivity of sound collection in various directions by changing the delay amount of each microphone.

  Thereby, by setting the directivity of sound collection in each of the areas 100 and 101 shown in FIG. 1, sound can be collected in each of the areas 100 and 101, and the sound emission means of the present invention is realized. In this case, the control unit 10 may set the sound emission direction of the masker sound according to the set sound collection direction of the array microphone.

1-masker sound emitting device, 10-control unit (setting unit, calculation unit), 13-voice input unit (input unit), 14-signal processing unit, 15-distributor (signal distribution unit), 16A, 16B, 16C-speaker, 17-speaker unit, 171-17n-speaker, 20-delay unit (delay supply unit), 100, 101, 102-region, 142-masker sound generation unit (generation unit)

Claims (4)

A sound processing device that emits masker sounds masking sounds generated in a plurality of areas from a plurality of sound emitting means,
A plurality of sound collecting means for collecting sounds generated in the plurality of regions;
Generating means for generating a plurality of masking sounds each masking the sound collected by the sound collecting means;
A plurality of synthesized signals obtained by synthesizing each masker sound generated by the generating means at a predetermined ratio, and a signal distributing means for distributing and outputting each synthesized signal to the plurality of sound emitting means,
The sound emitting means is
Sounds based on the synthesized signal are emitted to a plurality of areas,
The signal distribution means includes
The audio processing apparatus according to claim 1, wherein the ratio is set so that a masker sound corresponding to a sound collected in an area emitted by the sound output means for distributing and outputting is equal to or lower than a predetermined level. A speaker array in which a plurality of speakers are arranged;
A delay supply unit that distributes the combined signal to each of the plurality of speakers and supplies the combined signal with a delay;
The plurality of sound emitting means include
The audio processing apparatus according to claim 1, wherein the delay supply unit implements the delay supply unit by assigning different delay amounts to the plurality of regions. The delay supply means includes
The audio processing device according to claim 2, wherein a delay amount is provided so that a sound source position of a masker sound corresponding to the region is set in a direction of the region corresponding to the collected sound. A microphone array in which a plurality of microphones are arranged;
Delay synthesizing means for adding a delay to each of the sounds collected by the plurality of microphones,
The plurality of sound collecting means are:
The speech processing apparatus according to any one of claims 1 to 3, wherein the delay synthesizing unit is implemented by giving different delay amounts to the plurality of regions.

Images