Reverberation adding device and program

Abstract

To provide a reverberation adding device capable of adding reverberation corresponding to features of respective reflected sounds in a high frequency band and a low frequency band and suppressing an increase in calculation load of reverberation adding processing, and a program for designing the same.SOLUTION: A reverberation device is an IIR reverberation device that simulates a closed space sound field, and an IIR filter that generates a reverberation sound is configured on the basis of an extended state space model obtained by integrating a first state space model obtained by modeling an IIR filter that generates a reverberation sound in a frequency band equal to or lower than a Schrader frequency of the closed space sound field and a second state space model obtained by modeling an IIR filter that generates a reverberation sound in a frequency band exceeding the Schrader frequency.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a reverberation device and a program for designing the same. [Background technology]

[0002] In the production of audio content, multiple audio signals recorded at different times and locations are mixed together to create a single program. However, for example, mixing an audio signal recorded in a location with abundant reverberation with an audio signal recorded in a location with little reverberation can result in content that sounds unnatural to the ear. To avoid this problem, adding reverberation to an audio signal with little reverberation makes it possible to create audio content that sounds natural to the ear.

[0003] Furthermore, when converting audio signals with a relatively small number of channels, such as mono or 2-channel stereo, into a multi-channel sound format such as 5.1-channel surround or 22.2-channel surround, a channel conversion with a sense of spatial expansion can be performed by assigning signals to which reverberation, which simulates indirect sound caused by reflections from walls, has been added to channels installed to the side or rear of the listening point.

[0004] However, when adding reverberation to audio signals or converting the number of channels, the actual amount of reverberation depends largely on the experience and intuition of the program producer. For example, if an audio signal recorded in a space with short reverberation is mixed with an audio signal to which extremely long reverberation has been added, the content will sound unnatural.

[0005] Typically, an acoustic signal recorded in a given space (original sound field) consists of direct sound, which arrives directly at the listening point, and indirect sound (reverberant sound), which arrives after reflecting off walls and other surfaces. This indirect sound is a signal obtained by convolving the direct sound with the impulse response between the sound source and the listening point. Therefore, if the direct sound and impulse response are known, reverberant sound can be generated by convolving the direct sound with the impulse response. Therefore, by estimating the impulse response from a sound signal containing reverberation (original sound) and convolving this impulse response with another sound signal (direct sound) that does not contain reverberation, it is possible to add reverberation that is close to the reverberation of the original sound field. This method allows for natural reverberation even when mixing multiple sounds recorded at different times and locations.

[0006] Known reverberation systems include IIR (Infinite Impulse Response) reverberation systems, commonly known as FDN (Feedback Delay Network) or Schroeder reverberation, which simply simulate a closed sound field, and sampling reverberation (convolution reverberation), which adds reverberation by measuring the impulse response in an actual sound field in advance and convolving that impulse response with the audio signal as an FIR (Finite Impulse Response) filter. Hereinafter, sampling reverberation systems will be referred to as FIR reverberation systems in contrast to IIR reverberation systems.

[0007] IIR reverberators can be constructed with a relatively simple configuration, and the reverberation can be easily adjusted by changing the circuit parameters. On the other hand, FIR reverberators use the impulse response measured in an actual closed sound field, making it difficult to adjust the reverberation, but they have the advantage of being able to impart a more realistic and vivid reverberation than IIR reverberators. However, FIR reverberators are subject to the influence of the characteristics of the speakers and microphones used in measuring the impulse response, as well as background noise, which can result in unintended changes in sound quality by content creators. Furthermore, they do not allow for parametric adjustment of the reverberation. Therefore, traditional, simple IIR reverberators tend to remain popular (Non-Patent Document 1).

[0008] In order to overcome the above drawbacks, a device has been proposed that identifies the characteristics of the sound field measured from the measured impulse response, and generates and adds reverberation based on the identified characteristics (Non-Patent Document 1, Non-Patent Document 2). [Prior art documents] [Non-patent literature]

[0009] [Non-Patent Document 1] Masataka Nakahara et al., "Sound Field Reproduction Technology and Highly Realistic Content Production," Journal of the Acoustical Society of Japan, Vol. 78, No. 3, 2022, pp. 135-142 [Non-patent document 2] Masataka Nakahara, et al. “Development of a 4-pi sampling reverberator, VSVerb. - Application to VR production.” AES 149th Convention, (2020), e-Brief:634 [Non-patent document 3] Karolina Prawda, et al. “IMPROVED REVERBERATION TIME CONTROL FOR FEEDBACK DELAY NETWORKS.” Proceedings of the 22nd International Conference on Digital Audio Effects (DAFx-19), Birmingham, UK, September 2-6, 2019 [Non-patent document 4] Akira Sasaki et al., "Multiple-input, Multiple-output System Identification Based on Experimental Modal Analysis of Sound Fields," Journal of the Acoustical Society of Japan, Vol. 66, No. 2, 2010, pp. 56-64 [Non-Patent Document 5] Y. Haneda,et al. “Common acoustical pole and zero modeling of room transfer functions”, IEEE Transactions on Speech and Audio Processing, vol. 2, no. 2, (1994), pp. 320-328 [Non-patent document 6] Heinrich Kuttorff, translated by Kyoji Fujiwara and Takayuki Hidaka, "Room Acoustics: Architectural Resonance and Its Theory", Ichigaya Publishing, 2003 Summary of the Invention [Problem to be solved by the invention]

[0010] Most conventional IIR reverberation systems are designed by analyzing the sound field using statistical or geometric acoustics across the entire audible frequency range. However, in actual enclosed sound fields, while it is appropriate to analyze the high frequency range where reflected sound has high diffusion, in the low frequency range where diffusion is low, the reflected sound behaves like wave acoustics. Therefore, reverberation systems designed and analyzed using statistical or geometric acoustics may not be able to generate appropriate reverberation in the low frequency range.

[0011] In addition, FIR reverberation systems, which use the measured impulse response as an FIR filter, can accurately reproduce reverberant sounds regardless of the frequency band, but in general, in an acoustic vibration field, the higher the frequency, the faster the decay, while the lower the frequency, the slower the decay and the more likely it is to remain as reverberation, which tends to result in a long impulse response.As a result, the longer the impulse response, the greater the computational load of the convolution operation.

[0012] Therefore, in consideration of the above-mentioned problems, an object of the present invention is to provide a reverberation device that can add reverberation according to the characteristics of reflected sounds in both high and low frequency bands, while suppressing the computational load of the reverberation processing, and a program for designing such a device. [Means for solving the problem]

[0013] In order to solve the above problems, the reverberation device according to the present invention comprises: (1) An IIR type reverberation device that simulates a closed sound field, characterized in that the IIR filter that generates reverberation sound is configured based on an extended state space model that integrates a first state space model that models an IIR filter that generates reverberation sound in a frequency band below the Schroeder frequency of the closed sound field, and a second state space model that models an IIR filter that generates reverberation sound in a frequency band above the Schroeder frequency.

[0014] (2) In the reverberation device of (1) above, it is further preferable that the first state space model is a state space model that reflects the wave acoustic characteristics of the sound field corresponding to a frequency band below the Schroeder frequency of the impulse response, and the second state space model is a state space model that models the geometrical acoustic or statistical acoustic reflection sound structure of the sound field corresponding to a frequency band above the Schroeder frequency of the impulse response.

[0015] (3) The reverberation device of (1) or (2) above can further reduce the load on the computer by using an IIR filter that generates reverberant sound using a dimension-reduced augmented state-space model obtained by deleting eigenvalues ​​that can approximate 0 in the eigenvalue matrix of the state matrix from the augmented state-space model that is an integration of the first state-space model and the second state-space model.

[0016] In order to solve the above problems, the program according to the present invention comprises: (4) A program for designing an IIR type reverberation device simulating a closed sound field, the program causing a computer to execute the following steps: acquire an impulse response; divide the impulse response into bands at the Schroeder frequency of the closed sound field; estimate parameters based on the wave acoustic characteristics of the sound field for the impulse response in a frequency band below the Schroeder frequency by experimental mode analysis to create a first state-space model; estimate parameters of the sound field for the impulse response in a frequency band above the Schroeder frequency by geometric acoustics or statistical acoustics analysis to create a second state-space model; integrate the first state-space model and the second state-space model to create an extended state-space model; and generate an IIR filter based on the extended state-space model to configure the reverberation device.

[0017] (5) In the program of (4) above, it is preferable that estimating parameters based on the wave acoustic characteristics of the sound field by the experimental modal analysis includes estimating modal parameters including the resonant frequency of the closed-space sound field and the modal damping ratio for each eigenmode, and that estimating parameters of the sound field by the geometrical acoustic or statistical acoustic analysis includes estimating at least one parameter of the closed-space sound field's reverberation time, virtual sound source distribution, early reflection sound, and reflection sound density.

[0018] (6) Preferably, the program of (4) or (5) above further causes the computer to execute the following steps: calculating an eigenvalue matrix of a state matrix for the augmented state-space model obtained by integrating the first state-space model and the second state-space model; deleting eigenvalues ​​that can be approximated to 0 from the eigenvalue matrix; and reconstructing the state-space model to generate a dimension-reduced augmented state-space model in which redundancy is eliminated; and generating an IIR filter based on the dimension-reduced augmented state-space model to configure a reverberation device. [Effects of the Invention]

[0019] The reverberation device and program of the present invention can add reverberation according to the characteristics of reflected sounds in both the high and low frequency bands, and by using an IIR filter, it is possible to prevent the computational load of the reverberation process from becoming too large. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is an example of a flowchart for designing a reverberation apparatus of the present invention. [Figure 2] 1 is a conceptual diagram illustrating a design process of a reverberation apparatus according to an embodiment of the present invention. [Figure 3] 1 is a conceptual diagram of a reverberation device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0021] The outline of the present invention will be explained. In the design of the reverberation system of the present invention, the Schroeder frequency f s The impulse response is divided into bands (described later), and the sound field in the low frequency band is modeled as an IIR filter using wave acoustic analysis, while the sound field in the high frequency band is modeled as an IIR filter using statistical acoustic or geometric acoustic analysis, and these are integrated to create an IIR reverberator.This allows us to obtain an IIR reverberator that captures the characteristics of the sound field more accurately without changing the structure of conventional IIR reverberators.

[0022] First, we will briefly explain the design of a reverberation device that utilizes the statistical or geometric acoustic characteristics of a sound field, and the design of a reverberation device that takes into account the wave acoustic characteristics of a sound field, both of which are used in the present invention.

[0023] In the design of reverberation systems that utilize the statistical or geometric acoustic characteristics of a sound field, the sound field is described by treating sound statistically and geometrically, ignoring the wave nature of sound. Reverberation in a sound field (especially in the high-frequency range) is predicted using a statistical acoustic model, which statistically treats the reflection structure assuming that reflected sound arrives from all directions with equal probability, or a geometric acoustic model, which models the propagation of sound waves as a finite number of sound rays and the reflection structure assuming specular reflections at walls (Non-Patent Document 3). From this model, an IIR filter simulating a closed-space sound field, known as FDN or Schroeder reverb, is designed to construct a reverberation system.

[0024] When designing a reverberation system that takes into account the wave-acoustic characteristics of a sound field, the sound field can be analyzed by solving the wave equation that describes the sound field. Experimental modal analysis is known as a method for obtaining the wave-acoustic characteristics of a sound field using measured impulse responses (Non-Patent Documents 4 and 5). The behavior of a steady-state enclosed sound field can be described in wave terms as a superposition of multiple independent eigenmodes. Experimental modal analysis uses single or multiple impulse responses measured in an enclosed sound field to estimate modal parameters, such as the room's natural frequency and modal damping ratios, which indicate the degree of damping for each eigenmode, determined by the shape and boundary conditions of the sound field. These parameters are then modeled using an Auto Regressive Moving Average (ARMA) model or a state-space model. A state-space model is known as a multi-input, multi-output model of a linear system. Its specific input-output relationships can be converted to and from an ARMA model, making it essentially the same as an ARMA model. Furthermore, these models are equivalent to approximating the room acoustic transfer function with an IIR filter, and can be used as a reverberator by adding reverberation to a speech signal using the approximated IIR filter as an estimated reverberation.

[0025] However, compared to the IIR filter described above, which is an IIR filter that simply simulates a closed-space sound field, called the FDN or Schroeder model, although the IIR filter structure is the same, its characteristics more closely approximate the wave acoustic characteristics of the sound field contained in the measured impulse response.

[0026] The behavior of sound waves in a closed sound field exhibits wave acoustic characteristics in the low frequency band due to low diffusion, but statistical and geometric characteristics in the high frequency band due to high diffusion. Therefore, it is appropriate to analyze an actual closed sound field using wave acoustics in the low frequency band where diffusion is low, and using statistical or geometric acoustics in the high frequency band where reflected sound has high diffusion (Non-Patent Document 6). Therefore, when adding reverberation using a reverberation device, it is desirable to separate the reverberation process according to the frequency band.

[0027] It is known that the frequency that serves as a guide for dividing a sound field into low and high frequency bands is given by the Schroeder frequency as shown in the following equation (1), where T is the reverberation time and V is the volume of the closed sound field.

[0028]

number

[0029] (Design of reverberation equipment) A method for designing a reverberation device will now be described in detail. Fig. 1 is an example of a flowchart for designing a reverberation device of the present invention. Fig. 2 is a conceptual diagram of the design process of a reverberation device according to an embodiment of the present invention. The flowchart in Fig. 1 is written as a process performed by, for example, a computer (information processing device). Each step will be explained below.

[0030] Step S1: The computer acquires an impulse response. The impulse response is preferably an actual impulse response measured from the reverberation of a closed sound field realized by a reverberation device.

[0031] Step S2: The computer calculates the Schroeder frequency f s Specifically, first, the Schroeder frequency f in equation (1) is calculated based on the volume V and reverberation time T of the target closed space sound field. s Then, as shown in Figure 2, the Schroeder frequency fs The impulse response is divided into bands using a low-pass filter (LPF) and a high-pass filter (HPF) with cutoff frequencies of . After this, IIR modeling is performed for each band of the impulse response.

[0032] Step S3: The low frequency band of the impulse response is modeled in wave acoustics using an IIR, i.e., the Schroeder frequency f s In the following frequency bands, parameters based on the wave acoustic characteristics of the sound field are estimated by experimental modal analysis, and a first state space model is created. Note that this state space model for the low frequency band is sometimes called the "low-frequency IIR model." A wave acoustic state space model for the low frequency band can be created, for example, by the method described in Non-Patent Document 4. Specifically, modal parameters such as the resonant frequency of a closed sound field and the modal damping ratio for each eigenmode are estimated by wave acoustic analysis, and the sound field is created as a state space model.

[0033] Step S4: The high frequency band of the impulse response is IIR modeled geometrically or statistically, i.e., the Schroeder frequency f s In the frequency band above 100 kHz, the sound field parameters are estimated by geometric or statistical acoustic analysis, and a second state space model is created. This high-frequency band state space model is sometimes called a "high-frequency IIR model." A geometric or statistical acoustic state space model in the high-frequency band can be created, for example, by the method described in Non-Patent Document 3, and the same conventional method can be used in the present invention. Specifically, various parameters used in FDN reverberation, such as the reverberation time of a closed-space sound field, virtual sound source distribution, early reflections, and reflection density, are estimated by statistical or geometric acoustic analysis, and an IIR filter is modeled. In other words, the sound field is created as a state space model.

[0034] It is possible to perform either step S3 or step S4 first, or to perform both steps simultaneously.

[0035] Step S5: The low-pass IIR model and the high-pass IIR model are integrated to generate an extended state space model. The IIR model can be equivalently expressed as a time series model that expresses the relationship between input and output, called a state space model, via latent variables called state variables. In this step, the low-pass IIR model (first state space model) obtained in step S3 and the high-pass IIR model (second state space model) obtained in step S4 are integrated. The procedure is shown below.

[0036] x L (n) and x H (n) are the state variables of the low-pass IIR model and the high-pass IIR model, respectively, and [A L ,B L ,C L ,D L ] and [A H ,B H ,C H ,D H ] are the state matrix, input matrix, output matrix, and feedthrough matrix of the low-pass IIR model and the high-pass IIR model, respectively, u(n) is the input signal, y L (n) and y H If (n) is the output signal of the low-pass IIR model and the high-pass IIR model, respectively, the respective state-space models can be expressed as the following equations (2) and (3). Note that here, it is assumed that the state-space model is a discrete system, and n represents the time sample.

[0037]

number

[0038]

number

[0039] By adding the outputs of the two band-divided models and integrating the models, the two models (2) and (3) described above can be expressed as an augmented state space model as shown in the following equation (4).

[0040]

number

[0041] Furthermore, the expanded state space model of the above equation (4) can be collectively expressed as the following equation (5).

[0042]

number

[0043] Step S6: Generate an IIR filter from the extended state space model and configure a reverberator. Since the state space model is equivalent to an approximation of the acoustic transfer function of a closed space using an IIR filter, an IIR filter can be generated from the extended state space model using conventional techniques.

[0044] The extended state space model in equation (5) above includes a low-frequency IIR model and a high-frequency IIR model. Therefore, an IIR reverberation system based on this extended state space model can add reverberation that reflects the wave acoustic characteristics of the sound field in the low frequency band and the geometric and statistical acoustic characteristics in the high frequency band.

[0045] (Reverberation device) 3 is a conceptual diagram of a reverberation device according to an embodiment of the present invention. The reverberation device of the present invention is an IIR type reverberation device 10 that simulates a closed sound field, and has a Schroeder frequency f s The low-frequency IIR model 11 generates reverberation in the following frequency range, and the Schroeder frequency f s It is based on an extended state space model that integrates a high-frequency IIR model 12 that generates reverberation sounds in a frequency band exceeding 100 kHz.

[0046] 3 is a state space model that models the wave acoustic characteristics of the sound field corresponding to the low frequency band of the impulse response. The high frequency IIR model that generates high frequency reverberation is a state space model that models the geometric or statistical acoustic characteristics of the sound field corresponding to the high frequency band of the impulse response. Note that the actual IIR reverberation device 10 may be configured as an IIR filter that integrates the low frequency IIR model 11 and the high frequency IIR model 12 by further performing coordinate transformation, as necessary, on the expanded state space model that integrates the two models.

[0047] The low-pass IIR model 11 and the high-pass IIR model 12 are state space models that extract different acoustic features and model them as parameters, and therefore have different characteristics. In other words, the low-pass IIR model 11 and the high-pass IIR model 12 act differently on the same input signal (original sound) and generate different reverberation sounds.

[0048] (redundancy elimination) The augmented state space model in equation (5) is simply a parallel connection of a low-pass IIR model 11 and a high-pass IIR model 12 expressed as state space models, and there is a possibility that there may be redundancy in the state variables. Therefore, the redundancy is eliminated by reducing the order based on the eigenvalue decomposition of the state matrix.

[0049] First, the state matrix A arg can be decomposed into eigenvalues ​​as follows: where Λ is an N-th order diagonal matrix with eigenvalues ​​as diagonal components, and P is an N-th order square matrix with eigenvectors arranged in the column direction.

[0050]

number

[0051] At this time, by replacing the state variables as in equation (7), the expanded state space model can be transformed into the following equation (8).

[0052]

number

[0053]

number

[0054] Furthermore, this state variable model can be abbreviated as the following equation (9).

[0055]

number

[0056] where A arg The eigenvalue matrix Λ of is a diagonal matrix whose elements are eigenvalues ​​as shown in the following equation (10). However, it is assumed that the eigenvalues ​​are arranged in descending order of their absolute values.

[0057]

number

[0058] If the state variables of the extended state space model are redundant, some of the absolute values ​​of the eigenvalues ​​will be very small or will be 0. For example, as shown in the following equation (11), if the Mth and subsequent eigenvalues ​​can approximate to 0, the state variables thereon are redundant and can be ignored. The criterion for determining whether an eigenvalue can approximate to 0 can be set appropriately depending on the spread of the absolute values ​​of the eigenvalues. For example, with respect to the eigenvalue with the largest absolute value, an eigenvalue whose absolute value is below a predetermined level (for example, 20 dB or less, or 40 dB or less) can be approximated to 0.

[0059]

number

[0060] Therefore, by ignoring the last element of the eigenvalue matrix Λ and reconstructing the state space model, a state space model that eliminates redundancy can be constructed.

[0061] Here, the new eigenvalue matrix obtained by deleting the last eigenvalues ​​(M+1th and later) is as follows:

[0062]

number

[0063] Furthermore, it is assumed that the eigenvector P can be decomposed into submatrices as follows: P = [P1P2]. Here, P1 and P2 are matrices with dimensions of N × M and N × (NM), respectively, and are eigenvectors corresponding to the first M eigenvalues ​​and the M+1 eigenvalues ​​onward. In this case, the state variables are transformed as shown in the following equation (13).

[0064]

number

[0065] Here, P1 -1 =P1 T The following relationship holds. The T on the right shoulder represents the transpose of the matrix. By adopting the state variables in equation (13), the N-dimensional augmented state space model can be reduced to M dimensions. By converting the state matrix, input matrix, output matrix, and feedthrough matrix as in equation (14), equation (15) is derived, and an augmented state space model that eliminates redundancy can be constructed.

[0066]

number

[0067]

number

[0068] By using the dimension-reduced augmented state space model in equation (15) above as an IIR reverberation device, it is possible to add reverberation that reflects the wave acoustic characteristics of the sound field in the low frequency band and the geometric and statistical acoustic characteristics in the high frequency band.Furthermore, dimension reduction makes it possible to make the IIR filter smaller, thereby further reducing the calculation load of the reverberation processing.

[0069] <Program> The design of the reverberation apparatus of the present invention can be executed by a computer according to the flowchart in Fig. 1. Also, a program for designing the reverberation apparatus of the present invention causes a computer to execute the steps of the flowchart in Fig. 1.

[0070] In this embodiment, the computer temporarily stores a program recorded on a portable recording medium or a program transmitted from a server in a main storage device. The computer then reads the program stored in the main storage device with a processor and executes processing in accordance with the read program with the processor. The computer may also read the program directly from a portable recording medium and execute processing in accordance with the program. Programs include information used for processing by an electronic computer that is equivalent to a program. For example, data that is not a direct instruction to a computer but has properties that define computer processing falls under the category of "something equivalent to a program."

[0071] In this embodiment, the program can be recorded on a computer-readable recording medium. The computer-readable recording medium includes non-transitory computer-readable media, such as a magnetic recording device, an optical disc, a magneto-optical recording medium, or a semiconductor memory. The program can be distributed, for example, by selling, transferring, or lending a portable recording medium, such as a DVD (digital versatile disc) or a CD-ROM (compact disc read only memory), on which the program is recorded. The program can also be distributed by storing the program in the storage of an external server and transmitting the program from the external server to another computer. The program can also be provided as a program product.

[0072] The IIR reverberation apparatus 10 can also be implemented by a computer. Such a computer can be implemented by storing a program describing the processing steps for implementing each function of the IIR reverberation apparatus 10 in a memory of the computer, and having the central processing unit (CPU) of the computer read and execute the program.

[0073] Although the above-described embodiments have been described as typical examples, it will be apparent to those skilled in the art that many modifications and substitutions can be made within the spirit and scope of the present invention. Therefore, the present invention should not be construed as being limited by the above-described embodiments, and various modifications or alterations are possible without departing from the scope of the claims. For example, the functions included in each block, step, etc. described in the embodiments can be rearranged so as not to be logically inconsistent, and multiple constituent blocks, steps, etc. can be combined or divided into one. [Explanation of symbols]

[0074] 10 IIR type reverberation device 11 Low-pass IIR model 12 High-Frequency IIR Model

Claims

1. An IIR type reverberation device that simulates a closed space sound field, The IIR filter that generates the reverberation sound is The sound field is configured based on an expanded state space model obtained by integrating a first state space model that models an IIR filter that generates reverberation sound in a frequency band below the Schroeder frequency of the closed-space sound field, and a second state space model that models an IIR filter that generates reverberation sound in a frequency band above the Schroeder frequency. Reverberation device.

2. 2. The reverberation apparatus according to claim 1, the first state space model is a state space model that reflects wave acoustic characteristics of a sound field corresponding to a frequency band equal to or lower than the Schroeder frequency of an impulse response, The reverberation device, wherein the second state space model is a state space model that models a geometric acoustic or statistical acoustic reflection sound structure of a sound field corresponding to a frequency band exceeding the Schroeder frequency of the impulse response.

3. 3. The reverberation apparatus according to claim 2, The IIR filter that generates the reverberation sound is a reverberation adding device configured based on a dimension-reduced augmented state space model obtained by deleting eigenvalues ​​that can approximate zero in an eigenvalue matrix of a state matrix from the augmented state space model obtained by integrating the first state space model and the second state space model.

4. A program for designing an IIR type reverberation device that simulates a closed space sound field, On the computer, Obtaining an impulse response; Dividing the impulse response into bands at Schroeder frequencies of the closed space sound field; Estimating parameters based on wave acoustic characteristics of a sound field by experimental modal analysis for the impulse response in a frequency band equal to or lower than the Schroeder frequency, and creating a first state space model; Estimating parameters of a sound field by performing geometric acoustics or statistical acoustics analysis on the impulse response in a frequency band exceeding the Schroeder frequency, and creating a second state space model; merging the first state space model and the second state space model to generate an augmented state space model; generating an IIR filter based on the augmented state space model to configure a reverberation device; A program to execute.

5. 5. The program according to claim 4, estimating parameters based on wave acoustic characteristics of the sound field by the experimental modal analysis includes estimating modal parameters including a resonant frequency of the closed-space sound field and a modal damping ratio for each eigenmode; The program, wherein estimating parameters of the sound field by the geometric acoustics or statistical acoustics analysis includes estimating at least one parameter of the reverberation time of the closed-space sound field, the distribution of a virtual sound source, the early reflection sound, and the density of the reflected sound.

6. 6. The program according to claim 5, On the computer, Furthermore, for the augmented state space model obtained by integrating the first state space model and the second state space model, an eigenvalue matrix of a state matrix is ​​calculated, eigenvalues ​​that can be approximated to 0 are deleted from the eigenvalue matrix, and a dimension-reduced augmented state space model is generated by reconstructing the state space model and eliminating redundancy; generating an IIR filter based on the dimension-reduced augmented state space model to configure a reverberation device; A program to execute.

Images