Simulation method for audio device, simulation apparatus for audio device, and simulation system for audio device

Abstract

The simulation method of the audio device obtains a standard model that models the input-output characteristics of the audio device, sets at least one parameter of the target audio device of the same type as the audio device, measures the input-output characteristics of the target audio device, and corrects the standard model using the measured input-output characteristics to create an individual model that models the input-output characteristics of the target audio device.

Description

Technical Field

[0001] One embodiment of the present invention relates to a simulation method for a sound machine, a simulation device for a sound machine, and a simulation system for a sound machine. Background Art

[0002] A simulator for an analog sound machine that can vary the amount of harmonic distortion according to the frequency of each harmonic order is disclosed in Patent Document 1.

[0003] Patent Document 1: Japanese Unexamined Patent Application Publication No. 2006-94153 Summary of the Invention

[0004] Analog sound machines produce sounds with different timbres even if they are of the same type because of individual differences. Therefore, even if a certain sound machine is modeled, it will be a different model from the sounds of other music machines of the same type.

[0005] An object of one embodiment of the present invention is to provide a simulation method for a sound machine, a simulation device for a sound machine, and a simulation system for a sound machine that can model the inherent sound of each machine.

[0006] The simulation method for a sound machine according to one embodiment of the present invention obtains a standard model that models the input-output characteristics of the sound machine, sets at least one parameter of an object sound machine of the same type as the sound machine, measures the input-output characteristics of the object sound machine, and corrects the standard model using the measured input-output characteristics to create an individual model that models the input-output characteristics of the object sound machine.

[0007] Effects of the Invention

[0008] One embodiment of the present invention can model the inherent sound of each machine. Brief Description of the Drawings

[0009] Figure 1 It is a block diagram showing the structure of the simulation system 1.

[0010] Figure 2 It is a schematic external view of the case where the measuring device 14 is attached to the analog sound machine 15.

[0011] Figure 3 It is a block diagram showing the structure of the measuring device 14.

[0012] Figure 4 It is a block diagram showing the structure of the information processing terminal 11.

[0013] Figure 5It is a flowchart showing the operations of the information processing terminal 11, the server 12, and the measuring device 14.

[0014] Figure 6 A is a conceptual diagram showing the signal processing module of the standard model 900. Figure 6 B is a conceptual diagram showing the signal processing module of the individual model 950. Detailed implementation

[0015] Figure 1 It is a block diagram showing the structure of the simulation system 1 of the audio device. The simulation system 1 of the audio device in this embodiment includes an information processing terminal 11, a server 12, a measuring device 14, and a simulated audio device 15. The information processing terminal 11 is connected to the server 12 via the Internet 13.

[0016] The information processing terminal 11 is composed of an information processing device such as a personal computer or a smart phone used by the user. The simulated audio device 15 is, for example, an audio device such as a simulated amplifier or a simulated effect unit. The user installs the measuring device 14 on the simulated audio device 15. The measuring device 14 is a device for measuring the input / output characteristics of the simulated audio device 15.

[0017] Figure 2 It is a schematic external view of the case where the measuring device 14 is installed on the simulated audio device 15. The simulated audio device 15 in this example is a simulated effect unit that distorts the audio signal. The simulated audio device 15 has an input terminal (IN), an output terminal (OUT), a knob 151, and a slider 152. The input terminal and the output terminal are analog audio terminals. The knob 151 corresponds to the parameter (Drive) of the distortion intensity in this example. The slider 152 corresponds to the parameter (Vol.) of the volume.

[0018] The measuring device 14 has an output terminal (OUT) 101, an input terminal (IN) 102, a servo motor 141, and a servo motor 142. The measuring device 14 outputs an analog audio signal to the input terminal of the simulated audio device 15 via the output terminal 101. The measuring device 14 inputs an analog audio signal from the output terminal of the simulated audio device 15 via the input terminal 102.

[0019] The servo motor 141 is mounted on the knob 151. The servo motor 141 adjusts the distorted parameter to an arbitrary value by rotating the knob 151. The servo motor 142 is mounted on the slider 152 via the rack 143. The servo motor 142 has a pinion gear. The servo motor 142 linearly moves the rack 143 by means of the pinion gear. The servo motor 142 moves the slider 152 via the rack 143, thereby adjusting the parameter of the volume to an arbitrary value. In addition, the motor 142 may also move the slider 152 by converting the rotational motion into a linear motion by means of a crank mechanism.

[0020] The information processing terminal 11 is connected to the measuring device 14 via a communication line such as USB. The information processing terminal 11 sends a measurement signal to the measuring device 14. The information processing terminal 11 inputs a sound signal for measurement into the analog audio device 15 via the measuring device 14 and receives the sound signal after signal processing. The information processing terminal 11 measures the input / output characteristics of the analog audio device 15 based on the sound signal sent to the measuring device 14 and the sound signal after signal processing received.

[0021] Figure 3 It is a block diagram showing the structure of the measuring device 14. The measuring device 14 has an output terminal 101, an input terminal 102, a CPU 103, a USB I / F 104, a flash memory 105, a RAM 106, a motor controller 107, a servo motor 141, and a servo motor 142.

[0022] The CPU 103 is a control unit that controls the operation of the measuring device 14. The CPU 103 performs various operations by reading a prescribed program stored in the flash memory 105, which is a storage medium, into the RAM 106 and executing it. For example, the CPU 103 controls the servo motor 141 and the servo motor 142 via the motor controller 107 to adjust the parameters of the analog audio device 15 to arbitrary values.

[0023] The USB I / F 104 is connected to the information processing terminal 11. The USB I / F 104 receives the first digital sound signal from the information processing terminal 11. The first digital sound signal is a sound signal for measurement. The sound signal for measurement is a measurement signal such as white noise, TSP (TimeStretched Pulse), or tone burst, for example. In addition, the sound signal for measurement may also be a music signal.

[0024] The CPU 103 converts the first digital audio signal into a first analog audio signal and outputs it to the analog audio device 15 via the output terminal 101. The CPU 103 receives the second analog audio signal from the analog audio device 15 via the input terminal 102. The CPU 103 converts the received second analog audio signal into a second digital audio signal. The CPU 103 sends the second digital audio signal to the information processing terminal 11 via the USB I / F 104.

[0025] Figure 4 is a block diagram showing the configuration of the information processing terminal 11. The information processing terminal 11 includes a display 301, a user I / F 302, a USB I / F 303, a flash memory 304, a RAM 305, a communication I / F 306, and a CPU 307.

[0026] The display 301 displays various information to the user. The user I / F 302 accepts operations from the user. In addition, the user I / F 302 can be stacked on the display 301 as a touch panel. The USB I / F 303 sends the first digital audio signal to the measuring device 14. In addition, the USB I / F 303 receives the second digital audio signal from the measuring device 14. The communication I / F 306 communicates with the server 12 via the network. The CPU 307 reads out the program stored in the storage medium, i.e., the flash memory 304, to the RAM 305 to implement a specified function. As Figure 4 shown, the CPU 307 functionally constitutes a standard model acquisition unit 171, a measurement unit 172, and an individual model creation unit 173. These configurations are implemented as functional configurations of the application program read by the CPU 307. The standard model acquisition unit 171 acquires the standard model 900 from the server 12. The measurement unit 172 measures the input / output characteristics of the analog audio device 15 via the measuring device 14. The individual model creation unit 173 creates an individual model that models the input / output characteristics of the analog audio device 15.

[0027] Figure 5 is a flowchart showing the operations of the information processing terminal 11, the server 12, and the measuring device 14. First, the user selects the model of the analog audio device 15 via the user I / F 302 of the information processing terminal 11 and requests a measurement (S11). For example, the CPU 307 displays a list of amplifier model names on the display 301 through an application program. The user selects the model that they are using from the displayed list. Alternatively, the CPU 307 can also display the names of representative effectors such as distortion, equalizer, or compressor on the display 301. The user selects the name of the effector that they are using from the displayed names of effectors.

[0028] Server 12 receives a request (S21). Server 12 obtains a standard model 900 corresponding to the information indicating the model type included in the request (S22). The standard model 900 is a model in which the standardized input / output characteristics of an analog audio device of a certain model type are modeled by a digital signal processing module.

[0029] Figure 6 A is a conceptual diagram of the signal processing module of the standard model 900. The standard model 900 has a standard filter module 901 and an adaptive filter module 902. The standard filter module 901 and the adaptive filter module 902 are respectively signal processing modules that simulate the electrical characteristics of an analog circuit (a circuit composed of electronic components such as resistors, diodes, capacitors, vacuum tubes, or coils) with digital filters. In Figure 6 In A, for the sake of easy explanation, only one standard filter module 901 and one adaptive filter module 902 are shown, and an example of two filter modules connected in series is shown. However, the standard model 900 actually has multiple signal processing modules and is a digital filter circuit with various connection methods.

[0030] These signal processing modules are pre-produced by simulating the electrical characteristics of the actual analog circuit at the manufacturer of the analog audio device. Alternatively, the standard model can also be produced by measuring the input / output characteristics (such as impulse response) of the analog audio device under multiple measurement conditions. Such standard models are stored in the database of server 12.

[0031] The standard filter module 901 is a digital filter that does not depend on the changes in parameters such as the knobs and sliders of the analog audio device, and includes, for example, an envelope extraction filter (envelope follower), etc. The adaptive filter module 902 is a digital filter whose filter coefficients change corresponding to the changes in the parameters of the analog audio device. The standard filter module 901 and the adaptive filter module 902 can be non-linear filters or linear filters.

[0032] Server 12 sends the standard model 900 obtained from the database to the information processing terminal 11 (S23). The information processing terminal 11 receives the standard model 900 (S12). Thereby, the standard model acquisition unit 171 of the information processing terminal 11 acquires the standard model 900.

[0033] The measurement unit 172 of the information processing terminal 11 sends a measurement audio signal to the measurement device 14 to indicate the measurement (S13). If the measurement device 14 receives the measurement instruction (S31), it sets the parameters of the analog audio device 15 (S32).

[0034] The measurement device 14 sets the values of the knob 151 and the slider 152 when the measurement device 14 is installed on the analog audio device 15 as reference values. For example, after the user sets the parameter values to the most frequently used parameter values and installs the measurement device 14, the user instructs the start of measurement via the user I / F 302 of the information processing terminal 11.

[0035] The measurement device 14 inputs the measurement audio signal to the analog audio device 15 and receives the audio signal after signal processing from the analog audio device 15 (S33). As described above, the measurement audio signal is, for example, a measurement signal such as white noise or a music signal. In the case of a measurement signal, multiple measurement signals with different levels are measured respectively. In the case of a music signal, multiple music signals with different contents are measured respectively.

[0036] The measurement device determines whether all the parameter values of the analog audio device 15 have been measured (S34). If not all the parameter values have been measured, it returns to S32 to set the parameters.

[0037] The measurement device 14 can set each parameter of the analog audio device 15 to each value with the minimum resolution from the minimum value to the maximum value in sequence by controlling the servo motor 141 and the servo motor 142. The relationship between the rotational positions of the servo motor 141 and the servo motor 142 and the rotational position of the knob 151 and the sliding position of the slider 152 is obtained, for example, in the following manner.

[0038] The manufacturer of the audio device registers information such as the minimum value, maximum value, and resolution of Drive and Vol. in the database of the server 12. The information processing terminal 11 obtains the information such as the minimum value, maximum value, and resolution of Drive and Vol. from the server 12. Alternatively, for example, the user can also input the information such as the minimum value, maximum value, and resolution of Drive via the user I / F 302 of the information processing terminal 11. In addition, for example, the user can also use the camera (not shown) of the information processing terminal 11 to take pictures of the knob 151 and the slider 152. The information processing terminal 11 can also identify the maximum value, minimum value, resolution, and current position of the knob 151 and the slider 152 through image processing.

[0039] The measurement device 14 receives information such as the minimum value, maximum value, and resolution of each parameter from the information processing terminal 11. The measurement device 14 rotates the servo motor 141 to the left and right, associates the position where it stops when rotating to the right with the minimum value, and associates the position where it stops when rotating to the left with the maximum value. Similarly, the measurement device 14 rotates the servo motor 142 to the left and right, associates the position where it stops when rotating to the right with the maximum value, and associates the position where it stops when rotating to the left with the minimum value. Then, the measurement device 14 associates the information of the resolution with the rotation angle. Thus, the measurement device 14 can set each value with the minimum resolution from the minimum value to the maximum value for each parameter by rotating the servo motor 141 and the servo motor 142.

[0040] In addition, for example, the user can also use a camera (not shown) of the information processing terminal 11 to take pictures of the knob 151 and the slider 152 in the state where the measurement device 14 is installed. The information processing terminal 11 identifies the maximum value, minimum value, resolution, and current position of the knob 151 and the slider 152 through image processing. The measurement device 14 obtains the maximum value, minimum value, resolution, and position information of the knob 151 and the slider 152 from the information processing terminal 11, and obtains the relationship between the rotation angles of the servo motor 141 and the servo motor 142 and the positions of the knob 151 and the slider 152.

[0041] In addition, the value of the parameter can also be changed manually by the user. After the measurement of the value of a certain parameter is completed, the information processing terminal 11 displays a reminder of the change of the parameter value on the display 301. The user operates the knob 151 and the slider 152 to change the value of the parameter. Thereafter, the user can also operate the user I / F 302 of the information processing terminal 11 to give an instruction to measure the value of the next parameter. Alternatively, the information processing terminal 11 can also display a guide for the value of the next parameter overlapping on the images of the knob 151 and the slider 152 captured by a camera (not shown).

[0042] The measurement device 14 repeatedly measures the values of each parameter using multiple measurement signals with different volumes or multiple music signals with different types. When the measurement device 14 determines that the values of all parameters have been measured, it sends the sound signal (measurement result) received from the analog audio device 15 to the information processing terminal 11 (S35). However, it is not necessary to send the measurement result after all parameter values have been measured. The measurement result can also be sent sequentially after measurement using, for example, a certain measurement signal or music signal.

[0043] The information processing terminal 11 receives the measurement result (S14). The individual model creation unit 173 of the information processing terminal 11 corrects the standard model based on the measurement result, and creates an individual model that models the input / output characteristics of the analog audio device 15 (S15).

[0044] The individual model is a model obtained by correcting the standard model in a manner that represents the input / output characteristics of the measured target audio device (analog audio device 15). As Figure 6 shown in B, the individual model 950 is a model obtained by correcting the output of the standard model 900 by the correction filter module 951, for example. The individual model creation unit 173 separately obtains the difference between the output result when each of a plurality of measurement tone signals with different volumes is input to the individual model 950 and the measurement result measured via the measurement device 14, and sets the filter coefficient that minimizes this difference to the correction filter module 951. Thereby, the correction filter module 951 represents the frequency characteristics depending on the volume. The input / output characteristics of the analog audio device 15 change non-linearly according to the change of parameters. The correction filter module 951 is provided for each value of the combination of a plurality of parameters (knob 151 and slider 152) of the analog audio device 15. Therefore, the correction filter module 951 becomes a filter that corrects the difference between the standard model 900 and the input / output characteristics of the analog audio device 15.

[0045] Alternatively, the individual model creation unit 173 can also correct the filter coefficient of the adaptive filter module 902 shown in Figure 6 A, thereby correct the output of the standard model 900, and use the corrected standard model 900 as the individual model 950. In this case, the correction filter module 951 is not required. The individual model creation unit 173 separately obtains the difference between the output result when each of a plurality of measurement tone signals with different volumes is input to the individual model 950 and the result measured via the measurement device 14. The individual model creation unit 173 obtains the filter coefficient that minimizes this difference by a prescribed adaptive algorithm, and corrects the filter coefficient of the adaptive filter module 902.

[0046] Thereby, the individual model creation unit 173 creates the individual model 950 that represents the inherent sound of the analog audio device 15 owned by the user.

[0047] The information processing terminal 11 sends the created individual model 950 to the server 12 (S16). The server 12 receives the individual model 950 (S24) and registers it in the database (S25).

[0048] The user can download the individual model 950 registered in the database of the server 12 at any time by using an information processing device such as the information processing terminal 11, and thus use it. Therefore, the user does not need to carry the analog audio device 15, and can use virtual amplifiers and effectors with the same input / output characteristics as the analog audio device 15 at the required place and time.

[0049] The standard model and the individual model can also be filters using a specified algorithm such as a deep neural network (hereinafter referred to as DNN). The filter of the DNN is pre-constructed. For example, the manufacturer of the audio device inputs a large number of different types of music signals as measurement audio signals into the standard audio device, and uses the output signal as the correct solution to train the input / output characteristics of the deep learning audio device.

[0050] First, the manufacturer of the audio device fixes the parameters and inputs a large number of different types of music signals to train the input / output characteristics of the deep learning audio device. After that, when the manufacturer of the audio device changes the parameter values, restricted deep learning is performed to restrict the learning of the filter processing module without affecting the parameter changes. The manufacturer of the audio device manufactures the standard model based on DNN in the above manner.

[0051] The individual model production unit 173 causes the correction filter module 951 that corrects the output of the standard model 900 to learn as shown in Figure 6 B. In this case, the correction filter module 951 is also a filter of the DNN. The individual model production unit 173 manufactures the individual model 950 by causing the correction filter module 951 based on DNN to perform deep learning. The individual model production unit 173 obtains the difference between the output result when a plurality of different types of music signals are input to the individual model 950 and the measurement result measured via the measurement device 14, and causes the correction filter module 951 to learn in such a way that this difference becomes the minimum. In addition, when performing learning based on DNN, it is preferable to input a large number of different types of music signals and use the measurement result as the correct solution for learning. However, since the learning of the correction filter module 951 is a process of correcting the standard model 900 to the individual model 950, the computational load is significantly reduced compared to the learning in the case of manufacturing the standard model 900.

[0052] Alternatively, the individual model production unit 173 can also correct the output of the standard model 900 by causing the adaptive filter module 902 shown in Figure 6 A to perform deep learning again, and use the corrected standard model 900 as the individual model 950. In this case, restricted learning is performed to restrict the learning except for the adaptive filter module 902 related to the parameters of the knob 151 and the slider 152.

[0053] As described above, the individual model 950 can also be created through deep learning. In particular, the individual model creation unit 173 preferably corrects the standard model 900 created in advance by the manufacturer of the audio device according to the first deep learning through the second deep learning with restrictions related to parameters, and creates the individual model 950. Thereby, the simulation system 1 of the present embodiment can reduce the computational load when creating the individual model 950.

[0054] In the present embodiment, an example is shown in which the information processing terminal 11 communicates with the server 12, obtains the standard model, and creates the individual model. That is, in the present embodiment, the information processing terminal 11 is shown as an example of the simulation device of the audio device. However, for example, the measurement device 14 may have a communication function, communicate with the server 12, obtain the standard model, and create the individual model. In this case, the measurement device 14 also functions as a simulation device of the audio device. In addition, the server 12 may send a sound signal for measurement to the measurement device 14 and receive the measurement result, and create the individual model based on the measurement result. In this case, the server 12 functions as a simulation device of the audio device.

[0055] The description of the present embodiment is illustrative in all aspects and not restrictive. The scope of the present invention is not represented by the above embodiments, but by the claims. Further, within the scope of the present invention, it is intended to include meanings equivalent to the claims and all changes within the scope.

[0056] Description of reference numerals

[0057] 1... Simulation system

[0058] 11... Information processing terminal

[0059] 12... Server

[0060] 13... Internet

[0061] 14... Measurement device

[0062] 15... Simulated audio device

[0063] 101... Output terminal

[0064] 102... Input terminal

[0065] 103... CPU

[0066] 104... USB I / F

[0067] 105... Flash memory

[0068] 106... RAM

[0069] 107…Motor controller

[0070] 141…Servo motor

[0071] 142…Servo motor

[0072] 152…Slider

[0073] 171…Standard model acquisition unit

[0074] 172…Measurement unit

[0075] 173…Individual model production unit

[0076] 301…Display

[0077] 302…User I / F

[0078] 303…USB I / F

[0079] 304…Flash memory

[0080] 305…RAM

[0081] 306…Communication I / F

[0082] 307…CPU

[0083] 900…Standard model

[0084] 901…Standard filter module

[0085] 902…Adaptive filter module

[0086] 950…Individual model

[0087] 951…Calibration filter module

Claims

1. A simulation method for an audio machine, wherein, obtain a standard model that models the input-output characteristics of the audio machine, which performs signal processing on the input audio signal and outputs it; set at least one parameter of an object audio machine of the same type as the audio machine, input a measurement audio signal into the object audio machine, receive the measurement audio signal that has undergone signal processing through the parameter set in the object audio machine, and measure the input-output characteristics of the object audio machine based on the audio signal input into the object audio machine and the audio signal received from the object audio machine; correct the standard model using the measured input-output characteristics to produce an individual model that models the input-output characteristics of the object audio machine.

2. The simulation method for an audio machine according to claim 1, wherein, the at least one parameter includes a plurality of parameters, and the measurement is performed separately for each of the plurality of parameters.

3. The simulation method for an audio machine according to claim 1 or 2, wherein, the individual model is produced by deep learning using the measured input-output characteristics.

4. The simulation method for an audio machine according to claim 3, wherein, the standard model is pre-produced by a first deep learning, and the individual model is produced by correcting the standard model produced by the first deep learning through a second deep learning.

5. The simulation method for an audio machine according to claim 1 or 2, wherein, the measurement uses a music signal.

6. The simulation method for an audio machine according to claim 1 or 2, wherein, the measurement uses a measurement signal.

7. The simulation method for an audio machine according to claim 1 or 2, wherein, a motor is used to move an operating member provided in the object audio machine for accepting changes in the parameter.

8. An audio machine simulation device, which has: a standard model acquisition unit that acquires a standard model that models the input-output characteristics of the audio machine, which performs signal processing on the input audio signal and outputs it; a measurement unit that sets at least one parameter of an object audio machine of the same type as the audio machine, inputs a measurement audio signal into the object audio machine, receives the measurement audio signal that has undergone signal processing through the parameter set in the object audio machine, and measures the input-output characteristics of the object audio machine based on the audio signal input into the object audio machine and the audio signal received from the object audio machine; and an individual model production unit that corrects the standard model using the measured input-output characteristics to produce an individual model that models the input-output characteristics of the object audio machine.

9. The audio machine simulation device according to claim 8, wherein, the at least one parameter includes a plurality of parameters, and the measurement is performed separately for each of the plurality of parameters.

10. The audio machine simulation device according to claim 8 or 9, wherein, The individual model is created by deep learning using the measured input-output characteristics.

11. The simulation device for an audio device according to claim 10, wherein, the standard model is pre-created by a first deep learning, and the individual model is created by correcting the standard model created by the first deep learning through a second deep learning.

12. The simulation device for an audio device according to claim 8 or 9, wherein, the measurement uses a music signal.

13. The simulation device for an audio device according to claim 8 or 9, wherein, the measurement uses a measurement signal.

14. The simulation device for an audio device according to claim 8 or 9, wherein, it further includes a motor that moves an operating member provided in the target audio device for receiving a change in the parameter.

15. An audio device simulation system, comprising a simulation device for an audio device and a measurement device. In this audio device simulation system, the simulation device obtains a standard model that models the input-output characteristics of the audio device, which performs signal processing on an input audio signal and outputs it. The measurement device sets at least one parameter of a target audio device of the same type as the audio device, inputs a measurement audio signal to the target audio device, receives the measurement audio signal that has undergone signal processing based on the parameter set in the target audio device, and measures the input-output characteristics of the target audio device based on the audio signal input to the target audio device and the audio signal received from the target audio device. The simulation device corrects the standard model using the measured input-output characteristics and creates an individual model that models the input-output characteristics of the target audio device.

Images