Signal processing device, signal processing method, and signal processing program

Abstract

To provide a signal processing device, a signal processing method, and a signal processing program that can avoid divergence of acoustic distribution.SOLUTION: A signal processing device 30 generates an impulse response signal corrected for air absorption attenuation from a measured impulse response signal, and includes a derivation unit 30A2 that derives the corrected impulse response signal using the following equation, where h^(ν) is the Fourier transform of the impulse response signal before correction, t is time, γh is the propagation constant before correction, γr is the propagation constant after correction, n is an index specifying a poles, a is a value determined in advance according to the required accuracy, ωnp(ν) are the poles of the integrand, and r(t) is the corrected impulse response signal.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] The present invention relates to a signal processing device, a signal processing method, and a signal processing program. [Background technology]

[0002] In acoustic model experiments conventionally conducted during the design stage of a concert hall, theater, or other building, the impulse response signal inside a scale model is measured to predict the acoustic characteristics of the sound field of the completed hall or other building (hereinafter also referred to as the "real thing"). If the impulse response signal of the real thing can be estimated from the impulse response signal inside the scale model, the estimated impulse response signal can be convolved with musical sounds or vocals recorded in an anechoic chamber, allowing the acoustic state of the real thing to be confirmed by ear. In this specification, this process is referred to as auralization of the sound field inside a scale model.

[0003] The fundamental principle of acoustic model experiments is the law of similarity. If the scale of the model is 1 / S (one over S), then measurements must be performed while maintaining the relationship that frequency is S times that of the actual object and time is 1 / S times that of the actual object. Because the typical scale of a scale model is approximately 1 / 10 to 1 / 20, maintaining this relationship while covering the human audible frequency range (approximately 20 Hz to 20 kHz) requires measurement of a wide frequency range, including the ultrasonic range. To measure impulse response signals, a sound source is driven by an impulse signal, and the impulse response signal picked up by a microphone is discretized using an A / D (analog-to-digital) converter and input into a computer. Here, if the discretization period is converted to a value of 1 / S, the time axis is extended by S times, and the frequency band is shifted to a band 1 / S times larger, resulting in a signal with the same scale as the actual object. However, this impulse response signal does not fully reflect the physical conditions of the actual object.

[0004] This is because as sound waves propagate through the air, some of their energy is lost due to acoustic absorption by the air. This phenomenon is called air absorption attenuation, and it is known to be more pronounced at higher frequencies (see, for example, HE Bass, LC Sutherland, AJ Zuckerwar, DT Blackstock, and DM Hester, “Atmospheric absorption of sound: Further developments,” J. Acoust. Soc. Am. 97 (1), 680-683 (1995)).

[0005] However, conventional acoustic model experiments that auralize the sound field inside a scale model do not take into account air absorption attenuation, and therefore, sound cannot always be reproduced faithfully.

[0006] To solve this problem, there is a technique disclosed in Patent Document 1 by the proposers of the technique of the present disclosure.

[0007] That is, Patent Document 1 discloses a signal correction device that aims to reproduce sounds more faithfully.

[0008] This signal correction device comprises a separation means for separating an impulse response signal obtained from a scale model of the sound field to be reproduced into band response signals for each predetermined frequency band, a correction means for correcting energy attenuation due to air for each of the band response signals separated by the separation means, and a synthesis means for synthesizing the band response signals corrected by the correction means.

[0009] However, the technology disclosed in Patent Document 1 corrects the band-divided impulse response on the time axis using the absorption coefficient of the representative frequency of the band, which has the problem of not being able to accurately represent the frequency characteristics.

[0010] Therefore, the proposers of the technology of the present disclosure have proposed the technology relating to impulse responses disclosed in Non-Patent Document 1 and Non-Patent Document 2.

[0011] This technology performs signal processing by calculating the sound source distribution function from the impulse response and replacing only the characteristics of the medium. [Prior art documents] [Patent documents]

[0012] [Patent Document 1] Japanese Patent Application Laid-Open No. 2012-252147 [Non-patent literature]

[0013] [Non-Patent Document 1] Shinichiro Koyanagi, Yusei Yamada, Takayuki Hidaka, "Proposal of a method for correcting air absorption and dispersion for impulse responses," Proceedings of the Autumn 2013 Conference of the Acoustical Society of Japan, September 2013, pp. 1021-1022 [Non-patent document 2] Shinichiro Koyanagi, "Verification of air absorption and dispersion correction method for impulse response," Proceedings of the Autumn 2016 Conference of the Acoustical Society of Japan, September 2016, pp. 857-858 Summary of the Invention [Problem to be solved by the invention]

[0014] However, the techniques disclosed in Non-Patent Documents 1 and 2 have a problem in numerical calculations, in that the acoustic distribution diverges depending on the conditions. This problem can occur not only when the target impulse response signal is measured using a scale model of the sound field to be reproduced, but also when the target impulse response signal is obtained by acoustic measurement in a real-scale environment such as an actual concert hall or theater, or when the target signal is measured using a reverberation room sound absorption coefficient test.

[0015] The present disclosure has been made in consideration of the above circumstances, and aims to provide a signal processing device, a signal processing method, and a signal processing program that can avoid divergence of acoustic distribution. [Means for solving the problem]

[0016] The signal processing device according to the present invention as set forth in claim 1 is a signal processing device that generates an impulse response signal in which air absorption attenuation has been corrected from a measured impulse response signal, wherein ĥ(ν) is a Fourier transform of the impulse response signal before the correction, t is time, and γ h is the propagation constant before the correction, and γ r is the propagation constant after the correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω n p a derivation unit that derives the impulse response signal after the correction using the following formula, where (ν) is the pole of the integrand and r(t) is the impulse response signal after the correction.

[0017]

number

[0018] According to the signal processing device of the present invention as set forth in claim 1, the corrected impulse response signal is derived using the above formula, thereby making it possible to avoid the divergence of the acoustic distribution that occurs in the prior art.

[0019] The signal processing device of the present invention as set forth in claim 2 is the signal processing device as set forth in claim 1, further comprising an acoustic signal generation unit that generates an acoustic signal by convolving the corrected impulse response signal derived by the derivation unit with a dry source obtained by recording a predetermined sound in an anechoic chamber.

[0020] According to the signal processing device of the present invention as set forth in claim 2, an acoustic signal is generated by convolving the corrected impulse response signal derived by the derivation unit with a dry source obtained by recording a predetermined sound in an anechoic chamber, and by reproducing the acoustic signal, it is possible to reproduce the sound more faithfully.

[0021] A signal processing device according to the present invention as set forth in claim 3 is the signal processing device as set forth in claim 1 or claim 2, further comprising a noise compensation processing unit that, prior to the derivation unit deriving the corrected impulse response signal, separates the uncorrected impulse response signal into band response signals for each predetermined frequency band, and compensates for noise by transforming the separated band response signals so that a reverberation sound region on the rear end side of the band response signal corresponds to the attenuation gradient of a front end side region excluding the reverberation sound region.

[0022] According to the signal processing device of the present invention as set forth in claim 3, before the derivation unit derives the corrected impulse response signal, the uncorrected impulse response signal is separated into band response signals for each predetermined frequency band, and the separated band response signals are subjected to noise compensation processing by transforming the reverberation sound region on the rear end side of the band response signal so that it corresponds to the attenuation gradient of the front end side region excluding the reverberation sound region, thereby achieving a more natural auditory experience.

[0023] The signal processing device of the present invention as set forth in claim 4 is the signal processing device as set forth in claim 1 or claim 2, wherein the impulse response signal before the correction is a signal measured using a scale model of the sound field to be reproduced.

[0024] According to the signal processing device of the present invention as set forth in claim 4, the impulse response signal before correction is a signal measured using a scale model of the sound field to be reproduced, so that divergence of acoustic distribution can be avoided when reproducing the sound of the sound field using the impulse response signal obtained from the scale model.

[0025] A signal processing method according to the present invention as set forth in claim 5 is a signal processing method for generating an impulse response signal in which air absorption attenuation has been corrected from a measured impulse response signal, wherein ĥ(ν) is a Fourier transform of the impulse response signal before the correction, t is time, and γ h is the propagation constant before the correction, and γ r is the propagation constant after the correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω n p The computer executes the process of deriving the impulse response signal after the correction using the above formula, where (ν) is the pole of the integrand and r(t) is the impulse response signal after the correction.

[0026] According to the signal processing method of the present invention as set forth in claim 5, the corrected impulse response signal is derived using the above formula, thereby making it possible to avoid the divergence of the acoustic distribution that occurs in the prior art.

[0027] A signal processing program according to the present invention as set forth in claim 6 is a signal processing program for generating an impulse response signal in which air absorption attenuation has been corrected from a measured impulse response signal, wherein ĥ(ν) is a Fourier transform of the impulse response signal before the correction, t is time, and γ h is the propagation constant before the correction, and γ r is the propagation constant after the correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω n p The computer is caused to execute a process in which (ν) is the pole of the integrand, r(t) is the impulse response signal after the correction, and the impulse response signal after the correction is derived using the above formula.

[0028] According to the signal processing program of the present invention as set forth in claim 6, the corrected impulse response signal is derived using the above formula, thereby making it possible to avoid the divergence of the acoustic distribution that occurs in the prior art. [Effects of the Invention]

[0029] As described above, according to the present invention, divergence of the sound distribution can be avoided. [Brief explanation of the drawings]

[0030] [Figure 1] 1 is a partial block diagram showing the overall configuration of a sound reproduction system according to an embodiment. [Figure 2] 1 is a block diagram showing a configuration of a main part of an electrical system of a signal processing device according to an embodiment; [Figure 3] 2 is a schematic diagram showing main storage contents of a secondary storage unit provided in the signal processing device according to the embodiment; FIG. [Figure 4] FIG. 2 is a schematic diagram illustrating a configuration of an impulse response signal database according to the embodiment. [Figure 5] FIG. 2 is a schematic diagram illustrating a configuration of a dry source database according to an embodiment. [Figure 6] 1 is a block diagram showing an example of a functional configuration of a signal processing device according to an embodiment. [Figure 7] 4 is a flowchart showing a processing flow of a signal processing program according to the embodiment. [Figure 8] 10 is a graph illustrating a noise compensation process according to an embodiment. [Figure 9A] 10 is another graph illustrating the noise compensation process according to the embodiment. [Figure 9B] 10 is another graph illustrating the noise compensation process according to the embodiment. [Figure 10] 10 is another graph illustrating the noise compensation process according to the embodiment. [Figure 11] FIG. 10 is a diagram for explaining an example according to an embodiment, and is a graph showing an example of an impulse response in which sound waves arrive at 1 meter intervals as a propagation distance when the temperature is 10°C and the humidity is 20%. [Figure 12]FIG. 12 is a diagram for explaining an example according to an embodiment, and is a graph showing an example of a corrected impulse response when the impulse response of FIG. 11 is regarded as data for a 1 / 10 scale model, and the temperature is 20°C and the humidity is 50%. [Figure 13] FIG. 13 is a diagram for explaining an example according to the embodiment, and is a partially enlarged view of the graph shown in FIG. 12. DETAILED DESCRIPTION OF THE INVENTION

[0031] Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings.

[0032] First, the overall configuration of a sound reproduction system 10 to which a signal processing device and a signal processing method according to the present invention are applied will be described with reference to Fig. 1. Fig. 1 is a partial block diagram showing the overall configuration of the sound reproduction system 10 according to the present embodiment.

[0033] As shown in the figure, in sound reproduction system 10 according to this embodiment, speaker 50 is installed at a predetermined sound source position (in this embodiment, a position approximately in the center of the stage of the concert hall in a plan view) with respect to scale model 14, which is a 1 / S (in this embodiment, S=10) scale of the sound field to be reproduced (in this embodiment, a concert hall). Also, in sound reproduction system 10 according to this embodiment, microphone 52 is installed at a predetermined sound collection position (in this embodiment, any position in the audience seats of the concert hall) with respect to scale model 14.

[0034] In this state, in the sound reproduction system 10 according to this embodiment, a predetermined impulse signal 12 is input to the speaker 50, causing the speaker 50 to emit a sound corresponding to the impulse signal 12, and the microphone 52 picks up the sound emitted by the speaker 50. As a result, an impulse response signal 16 indicating the signal level of the sound is output from the microphone 52, and this impulse response signal 16 is stored in a secondary storage unit 30D (see FIG. 2) described later.

[0035] On the other hand, the sound reproduction system 10 according to this embodiment is provided with a digital signal processing unit 18 that receives an impulse response signal 16 and performs digital signal processing on the impulse response signal 16 to generate an impulse response signal 20 that has been subjected to noise compensation processing and air absorption attenuation correction processing, which will be described later.

[0036] On the other hand, in the sound reproduction system 10 according to this embodiment, multiple types of sounds (musical sounds and singing in this embodiment) to which the sound reproduction system 10 is applied are reproduced in an anechoic chamber 22, and the sound signals obtained by recording these sounds (hereinafter referred to as "dry sources") are stored in advance.

[0037] The sound reproduction system 10 according to this embodiment is provided with a convolution unit 24 that generates an acoustic signal representing the sound to be reproduced by convolving the impulse response signal 20 with the dry source. The acoustic signal generated by the convolution unit 24 is reproduced by speakers in a listening room 26 or the like.

[0038] In the sound reproduction system 10 according to this embodiment, the processing of the digital signal processing unit 18 and the convolution unit 24 is realized by software processing using a signal processing device 30 (see FIG. 2) described later.

[0039] Next, the configuration of the main parts of the electrical system of the signal processing device 30 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing the configuration of the main parts of the electrical system of the signal processing device 30 according to this embodiment.

[0040] As shown in the figure, the signal processing device 30 according to this embodiment includes a CPU (Central Processing Unit) 30A as a computer that controls the operation of the entire signal processing device 30, a RAM (Random Access Memory) 30B that is used as a work area when the CPU 30A executes various processing programs, a ROM (Read Only Memory) 30C in which various parameters and the like are pre-stored, a secondary storage unit (here, a hard disk drive) 30D that functions as a storage means used to store various information, a keyboard 30E that is used to input various information, a display 30F that is used to display various information, and an external interface (I / F) 30G that exchanges various information with external devices, and these units are electrically connected to each other by a system bus BUS.

[0041] Therefore, the CPU 30A can access the RAM 30B, the ROM 30C, and the secondary memory unit 30D, obtain various input information via the keyboard 30E, display various information on the display 30F, and send and receive various information to and from external devices, etc. via the external interface 30G.

[0042] The external interface 30G is electrically connected to the aforementioned speaker 50 and microphone 52, and the CPU 30A can play sound using the speaker 50 and acquire acoustic signals obtained by sound pickup using the microphone 52.

[0043] Meanwhile, FIG. 3 schematically shows the main storage contents of the secondary storage unit 30D provided in the signal processing device 30. As shown in FIG.

[0044] As shown in the same figure, the secondary memory unit 30D has a database area DB for storing various databases, and a program area PG for storing control programs for controlling each part of the signal processing device 30 and application programs for performing various processing.

[0045] The database area DB includes an impulse response signal database DB1 and a dry source database DB2. The configuration of each database will be described in detail below.

[0046] In the acoustic reproduction system 10 according to this embodiment, microphones 52 are installed at a plurality of different positions in the audience seating of the scale model 14, such as the center of the first floor, the center of the second floor, etc., and the above-mentioned impulse response signals 16 are obtained at each position, and these plurality of impulse response signals 16 are registered in association with information indicating the installation positions of the microphones 52.

[0047] The impulse response signal database DB1 is configured to store the plurality of impulse response signals 16, and stores information on installation positions and impulse response signals, as shown in Fig. 4. Fig. 4 is a schematic diagram showing the configuration of the impulse response signal database DB1 according to this embodiment.

[0048] The installation position is information indicating the installation position of the microphone 52, and the impulse response signal is information indicating the impulse response signal 16 obtained when the microphone 52 is installed at the corresponding installation position.

[0049] On the other hand, in the acoustic reproduction system 10 according to the present embodiment, multiple types of the aforementioned dry sources are obtained in advance by recording in the anechoic chamber 22, and each dry source is registered in association with information indicating its type.

[0050] The dry sauce database DB2 is where these multiple dry sauces are registered, and is configured to store information on the type and dry sauce, as shown in Fig. 5. Fig. 5 is a schematic diagram showing the configuration of the dry sauce database DB2 according to this embodiment.

[0051] The type is information indicating the type of sound, such as musical sound, singing, etc., and the dry source is information indicating the dry source obtained using the corresponding type of sound.

[0052] Next, a functional configuration of the signal processing device 30 according to this embodiment will be described with reference to Fig. 6. Fig. 6 is a block diagram showing an example of the functional configuration of the signal processing device 30 according to this embodiment.

[0053] 6, the signal processing device 30 according to this embodiment includes a noise compensation processing unit 30A1, a derivation unit 30A2, and an acoustic signal generation unit 30A3. The CPU 30A of the signal processing device 30 executes a signal processing program (described later) stored in the program area PG, whereby the CPU 30A functions as the noise compensation processing unit 30A1, the derivation unit 30A2, and the acoustic signal generation unit 30A3.

[0054] The derivation unit 30A2 according to this embodiment defines ĥ(ν) as the Fourier transform of the impulse response signal before correction, t as time, and γ h is the propagation constant before correction, and γ r is the propagation constant after correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω n p With (ν) as the pole of the integrand and r(t) as the corrected impulse response signal, the impulse response signal with air absorption attenuation corrected is derived using the following formula (Equation (1)). Hereinafter, the process of correcting air absorption attenuation performed by the derivation unit 30A2 will be referred to as the "air absorption attenuation correction process." Note that in this embodiment, the medium is air and there are four poles, one of which is a non-physical quantity, so the remaining three, i.e., a, are set to 3, but this is not limited to this. When applying the technology of the present disclosure to other media, a may be set to 1, 2, or a value greater than or equal to 4.

[0055]

number

[0056] Here, the principle of the air absorption attenuation correction process according to this embodiment will be described in detail.

[0057] The calculation formulas used for correcting the air absorption attenuation disclosed in the above-mentioned Non-Patent Document 2 are the following formulas (2) and (3).

[0058]

number

[0059]

number

[0060] In equations (2) and (3), h represents the measured impulse response signal, r represents the impulse response signal after correction for air absorption attenuation, and ^ represents the Fourier transform. h (ω) is the wave number of the air when the impulse response signal is measured, and k r (ω) represents the wave number of the air to be corrected for air absorption attenuation, and each wave number and the air absorption coefficient α(ω) are related by the following equation (4).

[0061]

number

[0062] In equation (4), c is the speed of sound. The technology of Non-Patent Document 2 mentioned above does not take into account the frequency dependence of the speed of sound c, but this is taken into account in the correction of air absorption attenuation according to this embodiment. At this stage, the speed of sound c is considered a constant, and frequency dependence will be introduced later. φ(x) is the sound source distribution function as a function of the propagation distance x. Also, ω is the angular frequency. By eliminating the sound distribution φ and limiting the integral of the propagation distance x to an integral over a finite interval up to the integral interval X with a small negative value, the following equation (5) is obtained.

[0063]

number

[0064] In equation (5), the integral interval X is an arbitrary value that is large in the negative direction, and must be set to negative infinity to obtain the true value. However, the exponent coefficient is the wave number k at the time of measurement. h (ν) and the wave number k to be corrected r (ω), and the real part can be either positive or negative. If it is negative, it diverges, so this component cannot be implemented. It can be implemented numerically using a frequency filter, but the results are sensitive to the filter design, and the design is difficult, and it is an approximate method.

[0065] Therefore, we integrate the integral in the range where it converges, and use analytic continuation in the condition where it diverges.

[0066]

number

[0067] Furthermore, an inverse Fourier transform is performed, and the following equation (7) is calculated.

[0068]

number

[0069] In equation (7), t is time, γ is the propagation constant, and the relationship is γ = -ik. However, the integral path is not limited to the real axis in order to pick up an appropriate solution branch. Therefore, the integrand is considered to be a function of the complex number ω. Since the representation of complex wave numbers is difficult to understand, the propagation constant γ is used to represent them.

[0070] First, the propagation constant of air is identified. By rearranging the sound absorption coefficient of air (real part of the propagation constant) by Bass, which is published in ISO (International Organization for Standardization), and the imaginary part of the Alvaletz propagation constant, which is obtained by the Hilbert transform of the real part, the following equation (8) is obtained. Note that μ in equation (8) C represents the viscosity coefficient of air, and μ O , ω Orepresents the coefficient related to oxygen, and μ N , ω N represents the coefficient for nitrogen.

[0071]

number

[0072] This function is regular in most domains where ω is complex, and the above integral defined by analytic continuation can be calculated using the residue theorem. n p (ν) is γ r (ν)-γ h It is a solution of (ω)=0 and is obtained as a solution of a quartic equation. The exponential function e of the numerator -iωt From the convergence region of the equation, we consider that three of the four solutions of the quartic equation are meaningful poles, and the integral is finally evaluated using the formula above (1) with a set to 3.

[0073] Here, the inventors of the present invention will explain how they introduced the idea of ​​analytic continuation and ultimately came to perform calculations using the residue theorem, as well as the innovations they made.

[0074] Although the problem of the integral of the propagation distance x diverging could not be solved from the beginning, it was clear that the calculation result was a corrected result and would not diverge. Although a finite integral interval X was introduced, the result should not depend on the integral interval X. Furthermore, since it does not always diverge but converges under certain conditions, it was thought that it would be okay to drop the term related to the integral interval X.

[0075] Even after dropping the terms in the integration interval X, 1 / (ik h (ν)-ik r Numerical integration of (ω) with respect to ν suffers from numerical instability and divergence in regions with very small denominators, and is inefficient in regions with large denominators, since it contributes very little to the integral.

[0076] Therefore, in order to prevent the denominator from becoming unstable, we thought that we could perform the integration using the residue theorem, and converted it to the time domain.

[0077] Propagation constant γ h , γ r is written in the form -α+iω / c, where the air absorption coefficient α of the real part is frequency dependent, while the speed of sound c of the imaginary part is treated as a constant. Since the Cauchy-Riemann relation does not hold between the real and imaginary parts, it is not analytic and the validity of the analytic continuation is not guaranteed.

[0078] Therefore, we considered a method of setting the sound speed c from the air absorption coefficient α so that it satisfies the Cauchy-Riemann relation. This is not an unreasonable attempt to create something mathematically convenient, but is the same as the condition that the propagation constant is causal, and was adopted because it has physical validity. As a result, by creating an analytic propagation constant from the air absorption coefficient α, it became possible to evaluate the inverse Fourier transform using the residue theorem.

[0079] On the other hand, prior to the derivation unit 30A2 deriving the corrected impulse response signal, the noise compensation processing unit 30A1 according to this embodiment performs a process of compensating for noise (hereinafter referred to as "noise compensation process") by separating the uncorrected impulse response signal into band response signals for each predetermined frequency band, and transforming the reverberation sound region on the rear end side of the separated band response signal so that it corresponds to the attenuation gradient of the front end side region excluding the reverberation sound region.

[0080] Then, the acoustic signal generating unit 30A3 according to this embodiment performs a process of generating an acoustic signal by convolving the corrected impulse response signal derived by the derivation unit 30A2 with a dry source obtained by recording a predetermined sound in the anechoic chamber 22 (hereinafter referred to as "acoustic signal generating process").

[0081] Next, the operation of the sound reproduction system 10 according to this embodiment will be described with reference to FIG. 7. FIG. 7 is a flowchart showing the flow of processing of a signal processing program executed by the CPU 30A of the signal processing device 30 when an execution instruction is received via the keyboard 30E or the like of the signal processing device 30. This program is pre-stored in the program area PG of the secondary storage unit 30D. To avoid confusion, the following description will be given assuming that an impulse response signal database DB1 and a dry source database DB2 are pre-established and that the impulse response signal 16 and dry source to be applied are pre-specified. Furthermore, the following description will be given assuming that the viscosity coefficient of air, the temperature and humidity at the time the impulse response signal 16 is measured, and the temperature and humidity assumed for the corrected impulse response signal 20 are pre-specified.

[0082] In step 100 in the figure, the CPU 30A reads out the specified impulse response signal 16 (hereinafter referred to as "impulse response signal h(t)") from the impulse response signal database DB1, and also reads out the specified dry source from the dry source database DB2.

[0083] In the next step 102, the CPU 30A decomposes the read impulse response signal h(t) into band response signals hi(t) (i=1 to N (N is the number of frequency bands)) for each of a plurality of predetermined frequency bands.

[0084] In the next step 104, the CPU 30A executes noise compensation processing, the details of which will be described later, on one of the band response signals hi(t) (hereinafter referred to as the "processing target band response signal").

[0085] In the next step 106, the CPU 30A determines whether or not the processing of step 104 has been completed for all band response signals hi(t), and if the determination is negative, the process returns to step 104, whereas if the determination is positive, the process proceeds to step 108. When repeatedly executing the processing of steps 104 to 106, the CPU 30A sets the band response signals hi(t) that have not been processed up to that point as the band response signals to be processed.

[0086] In step 108, the CPU 30A combines the band response signals hi(t) obtained by the above processing.

[0087] In step 110, the CPU 30A applies the impulse response signal synthesized in the process of step 108 as the impulse response signal h to perform air absorption attenuation correction processing to derive the impulse response signal 20 after correction for air absorption attenuation. At this time, the CPU 30A calculates the temperature and humidity at the time when the above-mentioned impulse response signal h(t) is measured and the viscosity coefficient of air as a propagation constant γ h The temperature and humidity assumed for the corrected impulse response signal 20 and the viscosity coefficient of air are used as the propagation constant γ r applies to.

[0088] In this case, to convert the scale of the acoustic model experiment using the scale model 14 to the same scale as the real thing, it is sufficient to multiply the propagation distance x of the sound source distribution function by the scale factor σ. For example, if the scale model 14 is a 1 / 10 scale model, the scale factor σ is 1 / 10. This is equivalent to multiplying the propagation constant by σ, so the propagation constant γ h σγ h The scale conversion of the scale model 14 is completed by converting it into

[0089] Here, the noise compensation process according to this embodiment will be described.

[0090] Normally, the impulse response signal 16 actually measured contains superimposed noise from the surroundings of the scale model 14 and self-noise of the measuring device. Therefore, when air absorption attenuation correction processing is performed, the tail part of the impulse response signal, which is covered by this noise, is amplified, and the waveform of the impulse response signal may diverge, as shown in Figure 8, for example.

[0091] To solve this problem, the sound reproduction system 10 according to this embodiment executes the following noise compensation process, which takes into account the time structure of reflected sounds in a room and the human perception of reflected sounds.

[0092] That is, first, the band response signal hi(t) is smoothed. Specifically, the moving average or root mean square is calculated. FIG. 9A shows an example of the smoothed band response signal hi(t). As shown in the figure, the smoothed waveform is expressed as a level value (logarithmic amplitude).

[0093] The flat portion at the end of the smooth waveform in the figure corresponds to noise, and in a relatively large space such as a concert hall, the level values ​​of the band response signal decay approximately linearly. Next, as shown in Figure 9B as an example, we determine a regression line I for this linearly decaying portion and a regression line II for the noise portion. The regression lines can be determined using conventional techniques such as the least squares method.

[0094] The overall shape of the band response signal with noise superimposed on it can be represented by curve C, which is obtained by superimposing these two regression lines.

[0095] The reflected sound contained in the rear of a band response signal is called reverberation, and is composed of numerous sound waves that have been reflected multiple times by walls, ceilings, etc. It is known that the phase of these sound waves becomes randomized during the repeated reflection process. In other words, the reverberation sound at the rear of a noise-covered response signal itself has properties similar to Gaussian noise (see M.R. Schroeder, "Statistical Parameters of the Frequency Response Curves of Large Rooms," J. Audio. Eng. Soc., 35 (5), 299-305 (1987)).

[0096] Human perception of reflected sound is such that we cannot hear the individual reflected sounds that make up reverberation, nor can we distinguish between phase differences (see H. Kuttruff, “On the audibility of Phase Distortion in Rooms and its Significance for sound Reproduction and Digital Simulation in Room Acoustics,” Acustica, 74, 3-7 (1991)).

[0097] Based on the above findings, it can be concluded that the waveform at the end of the response signal, which is obscured by noise, can be replaced with a waveform that converges with the gradient of the initial decay of the response without affecting the auditory perception. This can be achieved by weighting the band response signal by the amplitude ratio (=I / C) between the regression line I and curve C.

[0098] Therefore, in the sound reproduction system 10 according to the present embodiment, the above-mentioned weighting is applied to the band response signal hi(t) as the noise compensation process before the air absorption attenuation correction process is performed.

[0099] FIG. 10 shows examples of an impulse response signal when neither noise compensation processing nor air absorption attenuation correction processing is performed, an impulse response signal when only noise compensation processing is performed, and an impulse response signal when both noise compensation processing and air absorption attenuation correction processing are performed.

[0100] As shown in the figure, by performing noise compensation processing, the roughly flat portion at the end of the impulse response signal (the region where the reverberation sound is covered with noise) is made to have the same gradient as the gradient in the region including the front end of the impulse response signal, and by further performing air absorption attenuation correction processing, air absorption attenuation is corrected. Note that the figures shown in Figures 8 to 10 were obtained using a 40 kHz band response signal measured with a 1 / 10 scale model.

[0101] In the next step 112, the CPU 30A performs an acoustic signal generation process to generate an acoustic signal to be reproduced in the listening room 26 by convolving this impulse response signal 20 with the dry source read out in the process of step 100 above.

[0102] In the next step 114, the CPU 30A stores the acoustic signal generated by the above processing in a predetermined area of ​​the secondary storage unit 30D in association with information that can identify each of the applied impulse response signal 16 and dry source, and then terminates the signal processing program.

[0103] The acoustic signals stored in the secondary storage unit 30D by this signal processing program are read out at any timing and reproduced in the listening room .

[0104] 11 to 13 are diagrams illustrating examples according to this embodiment. Fig. 11 is a graph showing an example of an impulse response in which sound waves arrive at 1 m intervals as a propagation distance x when the temperature is 10°C and the humidity is 20%. Fig. 12 is a graph showing an example of an impulse response corrected when the impulse response in Fig. 11 is regarded as data for a 1 / 10 scale model and the temperature is 20°C and the humidity is 50%. Fig. 13 is a partially enlarged view of the graph shown in Fig. 12.

[0105] As shown in FIGS. 11 to 13, the impulse response signal deriving method according to this embodiment was able to obtain an impulse response signal that almost perfectly matches the exact solution that was separately synthesized.

[0106] As described above in detail, in the signal processing device 30 according to this embodiment, the corrected impulse response signal is derived using equation (1). Therefore, it is possible to avoid the divergence of the acoustic distribution that occurs in the conventional technology.

[0107] Furthermore, in the signal processing device 30 according to this embodiment, the acoustic signal is generated by convolving the corrected impulse response signal derived by the derivation unit 30A2 with a dry source obtained by recording a predetermined sound in an anechoic chamber. Therefore, by reproducing the acoustic signal, it is possible to reproduce the sound more faithfully.

[0108] Furthermore, in the signal processing device 30 according to this embodiment, before the derivation unit 30A2 derives the corrected impulse response signal, the uncorrected impulse response signal is separated into band response signals for each predetermined frequency band, and the separated band response signals are subjected to noise compensation processing by transforming the reverberation sound region on the rear end side of the band response signal so that it corresponds to the attenuation gradient of the front end side region excluding the reverberation sound region, thereby providing a more natural listening experience.

[0109] Furthermore, in the signal processing device 30 according to this embodiment, the impulse response signals before correction are signals measured using a scale model of the sound field to be reproduced. Therefore, when reproducing the sound of a sound field using the impulse response signals obtained from the scale model, divergence of the acoustic distribution can be avoided.

[0110] Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. Various modifications and improvements can be made to the above embodiments without departing from the gist of the invention, and such modifications and improvements are also included in the technical scope of the present invention.

[0111] Furthermore, the above-described embodiments do not limit the inventions described in the claims, and not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention. The above-described embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining the multiple disclosed constituent elements. Even if some constituent elements are deleted from all of the constituent elements shown in the embodiments, as long as the effect is obtained, the configuration from which these constituent elements are deleted can be extracted as an invention.

[0112] For example, in the above embodiment, a case has been described in which an impulse response signal is obtained using one speaker 50 and one microphone 52, but the present invention is not limited to this, and for example, an impulse response signal may be obtained using multiple speakers 50 and / or multiple microphones 52.

[0113] Furthermore, in the above embodiment, the case where the processing of the digital signal processing unit 18 and the convolution unit 24 is performed by software processing has been described, but the present invention is not limited to this. For example, these processes may be performed by hardware processing, or these processes may be performed by a combination of software processing and hardware processing.

[0114] Furthermore, in the above embodiment, a regression line or the like is used to determine the attenuation gradient of a band response signal. However, the present invention is not limited to this. For example, a region of the band response signal that is equal to or greater than a predetermined level value may be extracted, the gradient of the extracted band response signal may be derived based on differences in level values ​​at multiple positions of the band response signal, and the signal with the derived gradient may be replaced with a signal of a region below the predetermined level. In this way, the region corresponding to the reverberation sound at the rear end of the impulse response signal may be transformed to correspond to the attenuation gradient of the front end region excluding the extracted region.

[0115] Furthermore, in the above embodiment, an explanation has been given of a case where an impulse signal is applied as a signal to be input to the speaker 50 when collecting an impulse response signal, but the present invention is not limited to this, and for example, a sweep signal, a noise signal, or the like may be applied instead of an impulse signal. In this case, too, the same effects as those of the above embodiment can be achieved.

[0116] Furthermore, the configurations of the sound reproduction system 10 and the signal processing device 30 described in the above embodiment (see Figures 1 to 3 and 6) are merely examples, and it goes without saying that unnecessary components may be deleted or new components may be added within the scope of the present invention.

[0117] Furthermore, the processing flow of the signal processing program shown in the above embodiment (see Figure 7) is also one example, and it goes without saying that unnecessary processing steps may be deleted, new processing steps may be added, or the order of processing steps may be changed, within the scope of the gist of the present invention.

[0118] For example, immediately after the processing of step 110 of the signal processing program, a process of individually multiplying the amplitude of the impulse response signal 20 by a desired magnification may be added. Note that this process is equivalent to adding an equalizer that corrects the frequency characteristics of the sound source or microphone.

[0119] Furthermore, the formula (1) shown in the above embodiment is also an example, and it goes without saying that unnecessary parameters may be deleted, new parameters may be added, or parameters may be changed within the scope of the present invention.

[0120] Furthermore, while the above embodiments are based on acoustic model experiments, it is also possible to correct numerical variance that often appears in full-scale acoustic measurements and numerical simulation results. Examples of the former include temperature and humidity correction for a full-scale concert hall and air absorption attenuation correction (removal) in reverberation chamber sound absorption tests. In the latter case of numerical simulation, if the propagation constants of the simulation system are roughly similar to those of the system being reproduced but there are small differences, correction can be performed using a method similar to that of the above embodiments, provided that the propagation constants of the simulation system and the system being reproduced are known. The technology disclosed herein can be applied as a technology for correcting the air absorption portion of impulse responses in general.

[0121] Furthermore, in the above-described embodiment, for example, the various processors listed below can be used as the hardware structure of the processing units that execute the processes of the noise compensation processing unit 30A1, the derivation unit 30A2, and the acoustic signal generation unit 30A3. As described above, the various processors include a CPU, which is a general-purpose processor that executes software (programs) and functions as a processing unit, as well as dedicated electrical circuits that are processors having a circuit configuration specifically designed to execute specific processes, such as a programmable logic device (PLD), a processor whose circuit configuration can be changed after manufacture, such as an FPGA (Field-Programmable Gate Array), and an ASIC (Application Specific Integrated Circuit).

[0122] The processing unit may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA).The processing unit may also be configured with a single processor.

[0123] Examples of configuring a processing unit with a single processor include, first, a form in which one processor is configured with a combination of one or more CPUs and software, and this processor functions as the processing unit, as typified by computers such as client and server. Second, a form in which a processor is used to realize the functions of the entire system, including the processing unit, on a single IC (Integrated Circuit) chip, as typified by systems on chips (SoCs). In this way, the processing unit is configured using one or more of the above-mentioned various processors as a hardware structure.

[0124] Furthermore, more specifically, the hardware structure of these various processors can be an electric circuit that combines circuit elements such as semiconductor elements. [Explanation of symbols]

[0125] 10. Acoustic reproduction system 12 Impulse Signal 14 Scale Model 16 Impulse response signals 18 Digital signal processing section 20 Impulse response signal 22 Anechoic chamber 24 Convolution section 26 Listening Room 30 Signal Processing Device 30A CPU 30A1 Noise compensation processing section 30A2 Derivation part 30A3 Acoustic signal generation section 30D Secondary storage 50 speakers 52 Microphone DB1 Impulse Response Signal Database DB2 dry source database

Claims

1. A signal processing device that generates an impulse response signal corrected for air absorption attenuation from a measured impulse response signal, Let h^(ν) be the Fourier transform of the impulse response signal before the correction, t be time, and γ ｈ is the propagation constant before the correction, and γ ｒ is the propagation constant after the correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω ｎ ｐ a derivation unit for deriving the corrected impulse response signal by the following formula, where (ν) is the pole of the integrand and r(t) is the corrected impulse response signal; [Equation 1] A signal processing device comprising:

2. an acoustic signal generating unit that generates an acoustic signal by convolving the corrected impulse response signal derived by the derivation unit with a dry source obtained by recording a predetermined sound in an anechoic chamber; The signal processing device according to claim 1 , further comprising:

3. a noise compensation processing unit that, prior to the derivation of the corrected impulse response signal by the derivation unit, classifies the impulse response signal before the correction into band response signals for each predetermined frequency band, and performs processing to compensate for noise by transforming the band response signals so that a reverberation sound region on the rear end side of the band response signal corresponds to an attenuation gradient of a front end side region excluding the reverberation sound region; 3. The signal processing device according to claim 1, further comprising:

4. The impulse response signal before correction is a signal measured using a scale model of a sound field to be reproduced.

3. The signal processing device according to claim 1 or 2.

5. A signal processing method for generating an impulse response signal corrected for air absorption attenuation from a measured impulse response signal, comprising: Let h^(ν) be the Fourier transform of the impulse response signal before the correction, t be time, and γ ｈ is the propagation constant before the correction, and γ ｒ is the propagation constant after the correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω ｎ ｐ Let (ν) be the poles of the integrand, and r(t) be the impulse response signal after the correction, and derive the impulse response signal after the correction using the following formula: [Equation 2] A signal processing method in which processing is performed by a computer.

6. A signal processing program for generating an impulse response signal corrected for air absorption attenuation from a measured impulse response signal, Let h^(ν) be the Fourier transform of the impulse response signal before the correction, t be time, and γ ｈ is the propagation constant before the correction, and γ ｒ is the propagation constant after the correction, n is an index specifying a poles, a is a value predetermined according to the required accuracy, and ω ｎ ｐ Let (ν) be the poles of the integrand, and r(t) be the impulse response signal after the correction, and derive the impulse response signal after the correction using the following formula: [Equation 3] A signal processing program that causes a computer to execute the processing.

Images