JP5644253B2 - Sound field support apparatus and program - Google Patents

Description

  The present invention relates to a technique for controlling an acoustic effect in an acoustic space.

  Special halls for performances such as classical music include a jukebox type hall with multiple rows of seats arranged in parallel in front of the stage, and a wineyard type hall with a block of seats surrounding the stage. There are different types. The auditory impression that a person in the audience feels while performing on a stage in this type of acoustic space is the positional relationship between the stage and the audience in the space, the height of the ceiling, the distance between the walls, It is determined depending on various conditions such as the orientation, the material of the ceiling and the wall (for example, see Patent Document 1).

Japanese Patent Laid-Open No. 06-307107 Japanese Patent Laid-Open No. 11-191098 JP 2010-72104 A

By the way, in this kind of acoustic space, there is one having a balcony protruding to the stage side along both side walls from above the rear wall in the acoustic space. In the acoustic space having such a design, the sound reflected from the ceiling of the acoustic space is blocked by the balcony on the way to the audience seat under the balcony. For this reason, the sense of space, such as the feeling of reverberation and spread, felt by the listeners in the audience seats under the balcony, is weaker and uncomfortable than the sense of space felt by the listeners in the seats without the balcony on the upper side. There was a problem of becoming something.
The present invention has been devised under such a background, and an object thereof is to improve a feeling of space felt by a person under a balcony in an acoustic space provided with the balcony.

  The present invention performs a convolution operation that convolves the sound signal of the sound collected by the sound collection means near the sound source in the acoustic space and the filter coefficient sequence, and the operation result of the convolution operation is calculated from the sound source. Digital filter means for outputting a sound from the reflected sound shield on the propagation path of the reflected sound to the listening area in the acoustic space to the listening area, and an impulse response in a section between the sound source and the reflected sound shield And an impulse response calculation means for obtaining a second impulse response which is an impulse response in a section between the sound source and the sound collection means, and is used for a convolution operation of the digital filter means. Means for determining a filter coefficient sequence, wherein the slope of the RMS waveform of the impulse response, which is the result of convolution of the filter coefficient sequence and the second impulse response, is A sound comprising: a filter coefficient string determining unit that obtains a filter coefficient string that approximates the slope of the effective value waveform of the first impulse response, and uses the filter coefficient string as a filter coefficient string used for a convolution operation of the digital filter means. A field support device is provided.

  According to the present invention, a person behind a reflection sound shield in an acoustic space having a reflection sound shield such as a balcony is caused to feel the same spatial feeling as that felt in an acoustic space without the reflection sound shield. Can do.

  Note that Patent Document 2 discloses a technique for improving the reverberation of a passenger seat behind a balcony in an acoustic space. However, this technology generates a direct sound that should be propagated directly from the stage to the seat behind the balcony and emits the sound to the seat, which is different from the present invention.

It is a figure which shows the structure of the sound field assistance apparatus which is one Embodiment of this invention. It is a figure which shows the acoustic space where the same sound field assistance apparatus was installed. It is a flowchart which shows operation | movement of the sound field assistance apparatus of FIG. It is a figure which shows the mode of propagation of the direct sound and reflected sound in the acoustic space of FIG. It is a figure which shows the mode of propagation of the direct sound and reflected sound in the acoustic space of FIG. The impulse response IRss between the sound source and the microphone near the sound source in the acoustic space shown in FIG. 2, the impulse response UpsideIR 'obtained by removing the waveform portion of the direct sound from the impulse response UpsideIR between the sound source and the microphone on the balcony, and the impulse response It is a figure which shows impulse response IRss * UpsideIR 'which is a convolution calculation result of IRss and UpsideIR'. The RMS value RMS ((UpsideIR ′ × D _i ) * IRss) of the result of convolution of the impulse response UpsideIR ′ × D _i obtained by attenuating the impulse response UpsideIR ′ of FIG. 6 by the attenuation rate D _i and the impulse response IRss, RMS waveform RMS (UpsideIR '* IRss) obtained for impulse response UpsideIR' * IRss, which is the result of convolution of impulse response UpsideIR 'and impulse response IRss, and RMS waveform obtained for impulse response UpsideIR (UpsideIR) FIG. It is a figure for demonstrating other embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a configuration of a sound field support apparatus 10 according to an embodiment of the present invention. FIG. 2A is a plan view of the acoustic space 80 in which the sound field support apparatus 10 is used as viewed from the ceiling 82 side of the space 80. FIG. 2B is a cross-sectional view taken along line AA in FIG. In the acoustic space 80 shown in FIGS. 2A and 2B, if there is no balcony 86, the sound produced from the sound source 90 on the stage 81 (for example, a musical instrument played by a player on the stage 81). Are reflected on the inner surface of the acoustic space 80 (the walls 83F, 83B, 83L, 83R intersecting the ceiling 82 and the ceiling 82), and the reflected sound from each reflecting surface is not blocked by the balcony 86, but the first-floor seat 85 To the listening area 91 behind. Hereinafter, this reflected sound propagation path is referred to as a reflected sound propagation path REF. Here, in the acoustic space 80, there is a balcony 86 as a reflection sound shield in the middle of the reflection sound propagation path REF. Therefore, if no measures are taken, a part of the reflected sound reflected by the inner surface of the acoustic space 80 is shielded by the balcony 86 in the process of propagating through the reflected sound propagation path REF, and the listening area 91 will not be reached. Therefore, the feeling of space felt by those who are in the audience seats of the listening area 91 is poorer and uncomfortable than the feeling of space felt by those who are in the seats without the balcony 86 above. The purpose of this embodiment is to solve this problem.

  In the acoustic space 80 shown in FIGS. 2A and 2B, microphones 1L-1, 1L-2, 1R-1, and 1R-2 are suspended from a ceiling 82. The microphones 1L-1, 1L-2, 1R-1, and 1R-2 are positioned near the sound source 90 on the stage 81 so that the direct sound of the sound source 90 can be collected independently. ing. Specifically, the distance is closer to the sound source 90 than at least the critical distance (the distance at which the acoustic energy ratio between the direct sound and the indirect sound is equal), and the closer the sound source 90 is, the smaller the indirect sound component is. The microphones 1L-1, 1L-2, 1R-1, and 1R-2 are unidirectional microphones having high sound reception sensitivity with respect to sound waves coming from the direction of the sound source 90. Speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) are provided on the surface 78 of the balcony 86 on the listening area 91 side. The sound emission surfaces of the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) are directed to the listening area 91.

  The sound field support apparatus 10 shown in FIG. 1 acquires the sound signal S of the musical sound generated by the sound source 90 on the stage 81 from the microphones 1L-1, 1L-2, 1R-1, and 1R-2, and this sound signal. A sound signal Z ″ for reinforcing and correcting a sense of space that a person in the listening area 91 feels audibly from S is generated, and the sound signal Z ″ is generated from the speakers 2L-k (k = 1 to 8) and 2R−. This is an apparatus for outputting from k (k = 1 to 8) toward the listening area 91 under the balcony 86. In the present embodiment, before starting the generation of the sound signal Z ″ by the sound field support device 10, a setting operation for determining the content of the signal processing for generating the sound signal Z ″ is performed. The configuration of the sound field support device 10 and details of the setting work will be described later.

  Also, in the acoustic space 80 shown in FIGS. 2A and 2B, the sound signal generator 99 and the directivity are positioned on the surface 79 on the ceiling 82 side of the balcony 86 and a position that can be regarded as equivalent to the position of the sound source 90 on the stage 81. Microphones 3L-k (k = 1 to 8) and 3R-k (k = 1 to 8) are respectively disposed. The sound signal generator 99 and the microphones 3L-k (k = 1 to 8) and 3R-k (k = 1 to 8) are used when the sound field support device 10 is set and removed after the setting work is completed. It is what is done. During the setting work, the sound signal generator 99 generates various sound signals under the control of the sound field support device 10. The microphones 3L-k (k = 1 to 8) and 3R-k (k = 1 to 8) pick up sounds generated by the sound signal generator 99, and output sound signals indicating the picked-up sounds to the sound field. Supplied to the support device 10. Details will be described later.

  Next, the configuration of the sound field support apparatus 10 will be described. The sound field support apparatus 10 is connected to microphones 1L-1, 1L-2, 1R-1, 1R-2 and speakers 2L-k (k = 1-8), 2R-k (k = 1-8). Yes. In the setting work of the sound field support device 10, the sound field support device 10 and the sound signal generator 99 and microphones 3L-k (k = 1 to 8) and 3R arranged on the stage 81 and the balcony 86 are used. -K (k = 1 to 8) is also connected. In the sound field support apparatus 10, each sound signal S input from the microphones 1L-1 and 1L-2 is converted into a digital format by an A / D converter (not shown) and added by the adder 11L. Each sound signal S input from the microphones 1R-1 and 1R-2 is converted into a digital format by an A / D converter (not shown), and added by an adder 11R. The output signal S of the adder 11L is input to a PEQ (Parametric Equalizer) 12L, and the output signal S of the adder 11R is input to a PEQ 12R.

  The PEQs 12L and 12R are connected to the acoustic space 80 → the microphones 1L-1, 1L-2, 1R-1 and 1R-2 with respect to the sound signals S of the microphones 1L-1, 1L-2, 1R-1 and 1R-2. → Sound field support device 10 → Speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) → Amplitude characteristics of transfer function (frequency response) of an acoustic feedback system that makes a round of acoustic space 80 This is a device that performs correction to prevent a sharp peak from occurring.

The amplitude characteristics of the transfer functions (frequency response) of the PEQs 12L and 12R are determined by the PEQ parameter p _s (s = 1 to 8). The PEQ parameter p _s (s = 1 to 8) includes a center frequency cf _s (s = 1 to 8) and a gain g _s (s = 1 to 8) when each band of the sound signal S is divided into eight bands. 8) (Level increase / decrease amount at the center frequency cf _s (s = 1 to 8) of each band) and Q value q _s (s = 1 to 8) (peak / dip sharpness of frequency characteristics of each band) It is a parameter to specify. The PEQ parameter p _s (s = 1 to 8) is determined by the CPU 25. Details will be described later.

  The sound signal S ′ corrected by the PEQ 12L passes through the switch 14L and the adder 15L and then branches to an 8-channel sound signal S ′. Each of the branched 8-channel sound signals S ′ is FIR (Finite Impulse). Response) is input to the filter 16L-k (k = 1 to 8). The sound signal S ′ corrected by the PEQ 12R passes through the switch 14R and the adder 15R, and then branches to an 8-channel sound signal S ′. Each of the branched 8-channel sound signals S ′ is an FIR filter 16R−. k (k = 1 to 8). When the test sound signal is output from the noise generator 30, the adders 15L and 15R supply the signals to the FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8). To play a role. The output signal of the noise generator 30 will be described later.

The FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8) output the output signals S ′ and filter coefficients h _m (m = 0, 1, 2,... M) of the PEQs 12L and 12R. A convolution operation is performed, and a convolution signal Z indicating the operation result is converted into PEQs 17L-k (k = 1 to), 17R-k (k = 1 to 8) and a delay unit 18L-k (k = 1 to). , 18R-k (k = 1 to 8) and the power amplifiers 19L and 19R are supplied to the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8). Thereby, from the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8), the sound that reinforces and corrects the spatial feeling in the listening area 91 is transmitted to the above-described reflected sound propagation path REF. And output toward the listening area 91. Here, the filter coefficients h _m (m = 0, 1, 2... M) used for the convolution calculation by the FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8) are as follows. , Determined by the CPU 25. Details will be described later.

The convolution sound signal Z output from each FIR filter 16L-k in the FIR filter 16L-k (k = 1 to 8) is input to the PEQ 17L-k. Further, the convolution sound signal Z output from each FIR filter 16R-k in the FIR filter 16R-k (k = 1 to 8) is input to the PEQ 17R-k. The PEQ 17L-k (k = 1 to 8) is a device that corrects the sound signal Z output from each FIR filter 16L-k in consideration of the installation environment of the speakers 2L-k (k = 1 to 8). . The PEQ 17R-k (k = 1 to 8) is a device that corrects the sound signal Z output from each FIR filter 16R-k in consideration of the installation environment of the speakers 2R-k (k = 1 to 8). . The PEQ parameters p _s of PEQs 17L-k (k = 1 to 8) and 17R-k (k = 1 to 8) are rewritten by the operation of the operation unit 31.

  The sound signal Z ′ corrected by each PEQ 17L-k in the PEQ 17L-k (k = 1 to 8) is input to the delay unit 18L-k. The sound signal Z ′ corrected by each PEQ 17R-k in PEQ 17R-k (k = 1 to 8) is input to the delay unit 18R-k. Delay units 18L-k (k = 1 to 8) and 18R-k (k = 1 to 8) are output signals Z of PEQs 17L-k (k = 1 to 8) and 17R-k (k = 1 to 8). In addition, the sound propagation time from the microphones 1L-1, 1L-2, 1R-1, and 1R-2 to the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) It is a device that gives a corresponding delay. The delay amounts of the delay units 18L-k (k = 1 to 8) and 18R-k (k = 1 to 8) are set by the CPU 25. Details will be described later.

  The sound signal Z ″ delayed by each delay unit 18L-k in the delay unit 18L-k (k = 1 to 8) is amplified by the power amplifier unit 19L and then a D / A converter (not shown). Is converted to an analog signal, and the converted analog signal is supplied to the speaker 2L-k, and a delay is given by each delay unit 18R-k in the delay unit 18R-k (k = 1 to 8). The sound signal Z ″ is amplified by the power amplifier 19R, then converted into an analog signal by a D / A converter (not shown), and the converted analog signal is supplied to the speaker 2L-k. Then, the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) output the analog signals supplied thereto as sound.

The CPU 25 is a control center of the sound field support apparatus 10. The CPU 25 executes the sound field support program 28 stored in the ROM 27 while using the RAM 26 as a work area. The sound field support program 28 has the following two functions.
a1. Impulse response calculation function This is the impulse response UpsideIR of the section between the sound source 90 and the balcony 86 in the reflected sound propagation path REF, and the sound source 90 and the microphones 1L-1, 1L-2, 1R-1, and 1R-2. It is a function which calculates | requires the impulse response IRss of the area between.
b1. Filter coefficient sequence determination function This is because the slope of the effective value waveform of the impulse response, which is the result of convolution of the filter coefficient h _m (m = 0, 1, 2... M) and the impulse response IRss, is the effective of the impulse response UpsideIR. A filter coefficient h _m (m = 0, 1, 2,... M) that approximates the slope of the value waveform is determined, and this filter coefficient h _m (m = 0, 1, 2,... M) is determined as the FIR filter 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8).

  Next, the operation of this embodiment will be described. The feature of the operation of the present embodiment is the processing performed by the CPU 25 during the setting work of the sound field support apparatus. FIG. 3 is a flowchart showing a series of processing performed by the CPU 25 during setting work. Here, step S100 and step S120 in the series of processes shown in FIG. 3 are processes executed by the function of the impulse response calculation function. Steps S130 (S131 to S136) are processes executed by the function of the filter coefficient determination function. When the operation unit 31 gives an instruction to set the sound field support device 10, the CPU 25 starts the process illustrated in FIG. 3.

  In FIG. 3, the CPU 25 impulses in a section between the sound source 90 and the microphones 3L-k (k = 1 to 8) and 3R-k (k = 1 to 8) on the balcony 86 in the reflected sound propagation path REF. The response UpsideIR is measured (S100). In step S100, the sound signal generating device 99 placed at the position of the sound source 90 on the stage 81 is regarded as the sound source 90, and the impulse response in the section between the sound signal generating device 99 and the balcony 86 is set as an impulse response UpsideIR. More specifically, the CPU 25 generates unit impulses from the sound signal generator 99 after turning off the switches 14L and 14R. As shown in FIG. 4, when a unit impulse is generated from the signal generator 99, the microphones 3L-k (k = 1 to 8) and 3R-k do not pass through the ceiling 82 or the walls 83F, 83B, 83L, and 83R. Microphones 3L-k (k = 1 to 8) and 3R-k (k = k = 8) after passing through the direct sound D-SND and the ceiling 82 and the walls 83F, 83B, 83L, and 83R having reached (k = 1 to 8). The reflected sound R-SND reaching 1 to 8) is picked up by the microphones 3L-k (k = 1 to 8) and 3R-k (k = 1 to 8). The CPU 25 generates sound signals (direct sound D-SND and reflected sound R) output from each of the microphones 3L-k (k = 1 to 8) and 3R-k (k = 1 to 8) after generating the unit impulse. -Sound signal including SND) is set as impulse response UpsideIR.

  Next, the CPU 25 performs a delay amount setting process (S110). In the delay amount setting process, the CPU 25 individually measures and measures the sound wave propagation time from the sound source 90 to each of the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8). The propagation time is set as the delay amount of the delay units 18L-k (k = 1 to 8) and 18R-k (k = 1 to 8).

  Next, the CPU 25 measures the impulse response IRss (S120). In this step S120, the sound signal generator 99 placed at the position of the sound source 90 on the stage 81 is regarded as the sound source 90, and the sound signal generator 99 and the microphones 1L-1, 1L-2, 1R-1, and 1R- Let the impulse response in the interval between each of the two be the impulse response IRss. More specifically, the CPU 25 generates unit impulses from the sound signal generator 99 after turning off the switches 14L and 14R. As shown in FIG. 5, when the unit impulse is generated from the sound signal generator 99, the microphones 1L-1, 1L-2, 1R-1, and the like without passing through the ceiling 82 and the walls 83F, 83B, 83L, and 83R, and The direct sound D-SND reaching 1R-2 and the reflected sound reaching the microphones 1L-1, 1L-2, 1R-1, and 1R-2 after passing through the ceiling 82 and the walls 83F, 83B, 83L, and 83R The R-SND is picked up by the microphones 1L-1, 1L-2, 1R-1, and 1R-2. After generating the unit impulse, the CPU 25 generates the sound signal S (direct sound D-SND and reflected sound R) from the signal sampling point 13L after the PEQ 12L or the signal sampling point 13R after the PEQ 12R (for example, the signal sampling point 13R). -A sound signal S) including SND is captured, and the captured sound signal S is defined as an impulse response IRss. FIG. 6A shows a waveform of the impulse response IRss.

The CPU 25 performs a filter coefficient sequence determination process (S130). The filter coefficient sequence determination process is performed by using filter coefficients h _m (m = 0, 1, 2,...) Used for the convolution calculation by the FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8). M) is determined. Here, the impulse response UpsideIR measured in step S <b> 100 is an impulse response in a section between the sound source 90 and the balcony 86 in the acoustic space 80. Therefore, in principle, the sound 2 generated by the sound source 90 and the convolution sound signal Z, which is the convolution calculation result of the filter coefficient h _m (m = 0, 1, 2,... M), are used as the speaker 2L-k (k = 1). ~ 8) and 2R-k (k = 1 to 8) to be output as sound to the listening area 91, the same spatial feeling as when the balcony 86 is eliminated should be created. However, in this embodiment, since the microphones 1L-1, 1L-2, 1R-1 and 1R-2 are used as sound collection means for collecting the sound generated by the sound source 90, the FIR filter 16L-k ( k = 1 to 8) and 16R-k (k = 1 to 8) are convolutional sound signals Z that are convolutional sound signals Z, not only the direct sound D-SND of the sound generated by the sound source 90 but also its reflected sound R- Even the SND is redundantly folded. For this reason, when this convolution sound signal Z is output to the listening area 91 as a sound, the sense of space created in the area 91 is still uncomfortable.

Therefore, in the filter coefficient sequence determination process, the slope of the effective value waveform of the impulse response that is the result of convolution of the filter coefficient h _m (m = 0, 1, 2,... M) and the impulse response IRss is the impulse response UpsideIR. A filter coefficient h _m (m = 0, 1, 2,... M) approximating the slope of the effective value waveform is obtained, and this filter coefficient h _m (m = 0, 1, 2,... M) is obtained from the FIR filters 16L-k and 16R. The filter coefficient h _m (m = 0, 1, 2,... M) used for the −k convolution operation is used.

  Specifically, the following processing is performed. First, the CPU 25 calculates an effective value waveform RMS (UpsideIR) of the impulse response UpsideIR obtained in step S100 (S131). In step S131, an effective value waveform RMS (UpsideIR) is calculated as follows. First, the CPU 25 squares the amplitude at each time t after the unit impulse generation time t = 0 in the impulse response UpsideIR and converts it to energy. Next, a moving average energy obtained by moving and averaging these energies every τ (for example, τ = 25 milliseconds) is obtained. A waveform composed of moving average energy at each time t is defined as an RMS waveform (UpsideIR).

Next, the CPU 25 selects one of the 16 types of impulse responses UpsideIR obtained in step S100 (for example, the impulse response UpsideIR in the section between the sound source 90 and the microphone 3L-1) as a representative value, and this impulse response. The impulse response UpsideIR ′ excluding the waveform portion corresponding to the direct sound in UpsideIR is calculated (S132). More specifically, the offset time T _OFFSET is added to each time t = 0, t = 1, t = 2... T = M in the impulse response UpsideIR, and each time t = 0 + T _OFFSET , t = 1 + T in the impulse response UpsideIR. The amplitude of _OFFSET , t = 2 + T _OFFSET , t = M + T _OFFSET ... Is the amplitude of t = 0, t = 1, t = 2. FIG. 6B is a diagram illustrating a waveform of the impulse response UpsideIR ′.

  The CPU 25 selects one of the FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8) as a setting target (S133).

  Here, as described above, in this embodiment, the microphones 1L-1, 1L-2, 1R-1, and 1R-2 are used as sound collection means. As shown in FIG. 6 (C), an impulse response which is a result of convolution operation of an impulse response IRss and an impulse response UpsideIR ′ between the sound source 90 and the microphones 1L-1, 1L-2, 1R-1 and 1R-2. The slope of IRss * UpsideIR ′ is gentler than the slope of impulse response UpsideIR ′. Then, the gentle inclination of the impulse response IRss * UpsideIR ′ causes the sense of space in the listening area 91 to be uncomfortable.

Therefore, in step S134, the CPU 25 attenuates the impulse response UpsideIR ′ by the attenuation rate D _i (i = 1, 2, 3, 4: where D ₁ <D ₂ <D ₃ <D ₄ ). Impulse response UpsideIR ′ × D _i (i = 1, 2, 3, 4: where D ₁ <D ₂ <D ₃ <D ₄ ) is obtained, and the impulse response UpsideIR ′ × D _{i is} selected from these four types. The impulse response (UpsideIR ′ × D _i ) * IRss RMS value RMS ((UpsideIR ′ × D _i ) * IRss), which is the calculation result of convolution of the impulse response IRss, with the RMS waveform (UpsideIR) Impulse response UpsideIR ′ × D _i that is closest to the slope of the filter is selected, and filter coefficients h _m (m = 0, 1, 2,... M) simulating the selected impulse response UpsideIR ′ × D _i are obtained (S134). .

Here, the attenuation factor D _i in this embodiment is a constant value for attenuating the value at the end of the effective value waveform RMS. In the present embodiment, to attenuate the effective value waveform RMS by the attenuation value D _i means that the amplitude at the end of the effective value waveform RMS is a value obtained by multiplying the amplitude by the attenuation factor D _i, and the starting point of the effective value waveform RMS and each amplitude between the end, which means that the value obtained by interpolation in accordance with the value obtained by multiplying the attenuation factor D _i at the end (prorated).

The step S134 will be described in more detail. The CPU 25 follows the same procedure as the calculation of the RMS waveform RMS (UpsideIR) in the step S131, and the RMS waveform RMS ((UpsideIR ′ × D ₁ ) * IRss), RMS ( (UpsideIR ′ × D ₂ ) * IRss), RMS ((UpsideIR ′ × D ₃ ) * IRss), and RMS ((UpsideIR ′ × D ₄ ) * IRss) are calculated. Next, for each of these, the CPU 25 converts the energy at each time t of the effective value waveform RMS ((UpsideIR ′ × D _i ) * IRss) into a level X _t that is each square root. Similarly, with regard to the RMS waveform RMS (UpsideIR), the energy at each time t is converted into a level X _0t that is each square root. Then, the CPU 25 calculates the RMS value RMS ((UpsideIR ′ × D ₁ ) * IRss), RMS ((UpsideIR ′ × D ₂ ) * IRss), RMS ((UpsideIR ′ × D ₃ ) * IRss), and RMS ( For each (UpsideIR ′ × D ₄ ) * IRss), at each time t from the start time t = 0 to the end time t = tmax of the RMS waveform RMS ((UpsideIR ′ × D _i ) * IRss). a level _{X t,} by substituting the level _{X 0t} at each time t of the effective value waveform RMS (UpsideIR) the following expression, the level of the effective value waveform _{RMS ((UpsideIR '× D i} ) * IRss) X t And the square sum MinValue of the error of the level X _0t in the RMS value RMS (UpsideIR) is calculated.

The CPU 25 calculates the RMS value RMS ((UpsideIR ′ × D ₁ ) * IRss), RMS ((UpsideIR ′ × D ₂ ) * IRss), RMS ((UpsideIR ′ × D ₃ ) * IRss), and RMS ((UpsideIR selecting a '× _D 4) * IRss error sum of squares MinValue determined by among formula (1)) is the effective value waveform RMS was minimal _{((UpsideIR' × D i)} * IRss). The CPU 25 acquires an impulse response UpsideIR ′ × D _{i in} which the impulse response UpsideIR ′ is attenuated by the attenuation rate D _i in the selected RMS waveform RMS ((UpsideIR ′ × D _i ) * IRss), and this impulse response UpsideIR Filter coefficient h _m (the amplitude of time t = 0, t = 1, t = 2... T = M in '× D _i is used for the convolution operation of the setting target FIR filter 16L-k (or 16R-k). m = 0, 1, 2,... M).

FIG. 7 shows that the attenuation rate D ₁ is D ₁ = 30 (dB), the attenuation rate D ₂ is D ₂ = 40 (dB), the attenuation rate D ₃ is D ₃ = 50 (dB), and the attenuation rate D ₄ is D _4. = RMS ((UpsideIR ′ × D ₁ ) * IRss), RMS ((UpsideIR ′ × D ₂ ) * IRss), RMS ((UpsideIR ′ × D ₃ ) * IRss) , And RMS ((UpsideIR ′ × D ₄ ) * IRss) and the impulse response UpsideIR ′ * IRss which is the result of convolution of the impulse response UpsideIR ′ and the impulse response IRss, the RMS waveform RMS (UpsideIR ′ * IRss ) And the RMS waveform RMS (UpsideIR) obtained for the impulse response UpsideIR. In the example of FIG. 7, 'although the attenuated by the attenuation factor D ₄ rms waveform RMS ((UpsideIR' impulse response UpsideIR the inclination of × D ₄₎ * _IRss), the inclination of the effective value waveform RMS (UpsideIR) Is closest. Thus, in this example, in step S134, the filter coefficient simulating the impulse response _{UpsideIR '× D 4 h m (} m = 0,1,2 ... M) is calculated.

CPU25, after calculating the filter coefficients _{h m (m = 0,1,2 ... M} ) at step S134, the calculated filter coefficients _{h m (m = 0,1,2 ... M} ) the setting target of the FIR filter 16L -K (or 16R-k) is set (S135). Next, the CPU 25 determines whether any of the FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8) are not yet set (S136). If there is one or more FIR filters 16L-k and 16R-k that have not yet been set (S135: Yes), the CPU 25 returns to step S133 and sets one of them as a new setting target. And repeat the subsequent processing.

If the CPU 25 determines in step S136 that there is no setting target yet (S136: No), it performs a PEQ parameter determination process (S140). The PEQ parameter determination process is a process of determining the PEQ parameter p _s (s = 1 to 8) and setting the determined PEQ parameter p _s (s = 1 to 8) in the PEQs 12L and 12R.

  In the PEQ parameter determination process, the CPU 25 performs the following process. First, the CPU 25 turns off the switches 14L and 14R, and then supplies a test sound signal from the noise generator 30 to the adders 15L and 15R. Adders 15L and 15R → FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8) → PEQ17L-k (k = 1 to 8) and PEQ17R-k (k = 1) 8) → Delay unit 18L-k (k = 1-8) and delay unit 18R-k (k = 1-8) → Power amplifier units 19L and 19R → Speakers 2L-k (k = 1-8) and 2R -K (k = 1 to 8) → acoustic space 80 → microphones 1L-1, 1L-2, 1R-1, and 1R-2 → addition units 11L and 11R → PEQs 12L and 12R via signal sampling points 13L and The response sound signal propagated to 13R is captured.

  The CPU 25 calculates the difference between the power spectra of the test sound signal and the response sound signal by adding the adders 15L and 15R → FIR filters 16L-k (k = 1-8) and 16R-k (k = 1-8) → PEQ 17L. -K (k = 1-8) and PEQ17R-k (k = 1-8) → delay unit 18L-k (k = 1-8) and delay unit 18R-k (k = 1-8) → power amplifier unit 19L and 19R → speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) → acoustic space 80 → microphones 1L-1, 1L-2, 1R-1, and 1R-2 → addition It is assumed that the amplitude characteristic R (ω) of the transfer function (frequency response) of the system of the parts 11L and 11R → PEQ 12L and 12R.

The CPU 25 obtains the target amplitude characteristic C (ω) necessary for setting the level of the entire band in the amplitude characteristic R (ω) to be equal to or lower than the target level LEV (for example, LEV = 6 (dB)), and transmits the PEQs 12L and 12R. A PEQ parameter p _s (s = 1 to 8) necessary for making the amplitude characteristic G (ω) of the function (frequency response) closest to the target amplitude characteristic C (ω) is calculated, and this PEQ parameter p _s (s = 1 to 8) are set in PEQs 12L and 12R. The specific procedure for calculating the PEQ parameter p _s (s = 1 to 8) necessary to make the amplitude characteristic G (ω) most approximate to the target amplitude characteristic C (ω) is not related to the present invention. Therefore, this explanation is omitted. For details, refer to Patent Document 3.

In the present embodiment described above, the slope of the effective value waveform of the impulse response, which is the result of convolution of the filter coefficient h _m (m = 0, 1, 2... M) and the impulse response IRss, is the effective of the impulse response UpsideIR. A filter coefficient h _m (m = 0, 1, 2,... M) approximating the slope of the value waveform is obtained, and this filter coefficient h _m (m = 0, 1, 2,... M) is obtained from the FIR filters 16L-k and 16R-. to use the convolution operation of k and filter coefficient _{h m (m = 0,1,2 ... M} ). As a result, a spatial feeling similar to that in the case where there is no balcony 86 in the listening area 91 under the balcony 86 in the acoustic space 80 is simulated. That is, in this embodiment, the waveform portion of the direct sound is obtained from the impulse response UpsideIR between the sound source 90 and the speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) on the balcony 86. The removed impulse response UpsideIR ′ is obtained, and the impulse response UpsideIR ′ × D _i obtained by attenuating the impulse response UpsideIR ′ by the attenuation factor D _i is used as a filter coefficient h _{m for the} convolution operation of the FIR filters 16L-k and 16R-k. (M = 0, 1, 2,... M). This corrects the macroscopic structure (envelope structure) of the reflected sound group of the impulse response UpsideIR while preserving the microscopic structure of the reflected sound group of the original impulse response UpsideIR (relative relationship of the reflected sound group). Filter coefficients h _m (m = 0, 1, 2,... M) are obtained. Then, the filter coefficients generated in this way _{h m (m = 0,1,2 ... M} ), the filter coefficients _h m (m = 0 to be used for the convolution operation of the FIR filter 16L-k and 16R-k, 1, 2, ... M). Therefore, according to the present embodiment, unlike a simple reverberation adding device, a person in the listening area 91 can feel the same spatial feeling as when the balcony 86 is not provided.

Although one embodiment of the present invention has been described above, the present invention may have other embodiments. For example, it is as follows.
(1) In the above embodiment, the number of microphones 1L-1, 1L-2, 1R-1, 1R-2 in the vicinity of the sound source 90 may be 1 to 3, or 5 or more. . The microphones 3L-k and 3R-k and the speakers 2L-k and 2R-k provided on the surfaces 79 and 78 of the balcony 86 may be 1 to 7 pairs, or may be 9 pairs or more. . Further, the number of FIR filters 16L-k (k = 1 to 8) and 16R-k (k = 1 to 8) in the sound field support device 10 may be 17 or more, or 15 or less. Good. The number of speakers 2L-k (k = 1 to 8) and 2R-k (k = 1 to 8) may be 17 or more, or 15 or less. Further, the number of bands in the PEQs 12L and 12R may be 9 bands or more, or 7 bands or less.

(2) In the above embodiment, the present invention is applied to the reinforcement of the spatial feeling of the person in the listening area 91 under the balcony 86. However, another reflected sound shielding at a position where the direct sound from the sound source 90 to the listening area 91 is not hindered and the reflected sound from the sound source 90 to the listening area 91 via the inner surface of the acoustic space 80 is hindered. When there is a body, the present invention may be applied to reinforce the sense of space of the person behind the reflection sound shielding body. FIG. 8 is a plan view of the acoustic space 80 ″ to which the present invention is applied as viewed from the ceiling. In this acoustic space 80 ″, the partition plate 87 is directed from the wall 83B ″ facing the stage 81 ″ to the stage 81 ″. Therefore, the feeling of space felt by the person in the listening area 91 behind the partition plate 87 when viewed from the sound source 90 on the stage 81 ″ is poor and uncomfortable. On the other hand, a person in the listening area 91 feels by outputting the sound signal Z generated by the sound field support apparatus 10 of the above embodiment from a speaker (not shown) provided on the listening area 91 side of the partition plate 87. It is possible to reinforce the sense of space to the same extent as when the partition plate 87 is not provided. This embodiment is also suitable for reinforcing the sense of space in an area called “stair area (commonly known as dug out)” that goes up from the rear of the hall to the front passenger seat.

(3) in the filter coefficient sequence determination process in the above embodiment, the four candidates for attenuation rate D _i attenuate the impulse response UpsideIR ', and select the optimal value of the four damping factor D _i. However, to the candidate of the attenuation factor D _i for attenuating impulse response UpsideIR 'may be two or three, may be 5 or more. In short, the number of attenuation rate D _i candidates may be any finite value.

DESCRIPTION OF SYMBOLS 10 ... Sound field assistance apparatus, 1L, 1R, 3L, 3R ... Microphone, 2L, 2R ... Speaker, 11L, 11R ... Adder, 12L, 12R, 17L, 17R ... PEQ, 13L, 13R ... Signal sampling point, 14L, 14R ... switch, 15L, 15R ... adder, 16L, 16R ... FIR filter, 18L, 18R ... delay unit, 19L, 19R ... power amplifier unit, 30 ... noise generator, 31 ... operation unit, 99 ... sound signal generator.

Claims (2)

A convolution operation is performed to convolve the sound signal of the sound collected by the sound collection means in the vicinity of the sound source in the acoustic space and the filter coefficient sequence, and the result of the convolution operation is calculated from the sound source in the acoustic space. Digital filter means for outputting a sound from the reflected sound shield on the propagation path of the reflected sound to the listening area to the listening area;
A first impulse response that is an impulse response in a section between the sound source and the reflection sound shielding body and a second impulse response that is an impulse response in a section between the sound source and the sound collecting means are obtained. Impulse response calculating means;
Means for determining a filter coefficient sequence used for the convolution operation of the digital filter means, and means for obtaining a third impulse response which is an impulse response obtained by removing a waveform portion corresponding to a direct sound from the first impulse response. An effective value of an impulse response that is a convolution operation result of the impulse response and the second impulse response among a plurality of types of impulse responses attenuated by a plurality of types of attenuation rates A filter coefficient sequence that selects an impulse response whose waveform slope is closest to the slope of the effective value waveform of the first impulse response, and uses a filter coefficient sequence that simulates the selected impulse response for the convolution operation of the digital filter means And a filter coefficient sequence determination means. On the computer,
A convolution operation is performed to convolve the sound signal of the sound collected by the sound collection means in the vicinity of the sound source in the acoustic space and the filter coefficient sequence, and the result of the convolution operation is calculated from the sound source in the acoustic space. Digital filter means for outputting a sound from the reflected sound shield on the propagation path of the reflected sound to the listening area to the listening area;
A first impulse response that is an impulse response in a section between the sound source and the reflection sound shielding body and a second impulse response that is an impulse response in a section between the sound source and the sound collecting means are obtained. Impulse response calculating means;
Means for determining a filter coefficient sequence used for the convolution operation of the digital filter means, and means for obtaining a third impulse response which is an impulse response obtained by removing a waveform portion corresponding to a direct sound from the first impulse response. An effective value of an impulse response that is a convolution operation result of the impulse response and the second impulse response among a plurality of types of impulse responses attenuated by a plurality of types of attenuation rates A filter coefficient sequence that selects an impulse response whose waveform slope is closest to the slope of the effective value waveform of the first impulse response, and uses a filter coefficient sequence that simulates the selected impulse response for the convolution operation of the digital filter means And a filter coefficient sequence determining means
A program that realizes

Images