Music synthesizer using resonators

Abstract

To provide a resonator-based audio synthesizer and a method for generating an audio output.SOLUTION: The audio synthesizer generates audio signals using a resonator array 110 that includes thousands or tens of thousands of resonator circuits based on the analysis of any acoustic space, such as acoustic instruments, rooms, studios, or concert halls. The AI network 145 is trained to learn characteristics of the sound of music that may be pleasing to the human ear, and generates audio effects 140 that are applied to selected frequencies in the spectrum. The input 130 or excitation signal 120 applied to each resonator circuit is provided to the resonator array, which configures the set of resonator circuits to process through the trained model of the target audio source to generate the output audio signal 150. The set of resonant circuits creates a new impulse response that creates a sound quality and timbre that is unique to the existing audio source.SELECTED DRAWING: Figure 1

Description

[Background technology]

[0001] Audio signal processing involves the electronic manipulation of audio signals. Audio signals are electronic representations of sound waves traveling through the air. Audio signals can be represented in analog or digital form. Signal processing or manipulation can be performed in either the analog or digital domain. Due to the electronic nature of the represented signals, computer-based devices can be used to synthesize audio signals by generating electronic signals that can be interpreted as sound waves. Analog signals are continuous, varying voltages that resemble sound waves traveling through the air. Digital signals, on the other hand, represent waveforms as a sequence of discrete values ​​or symbols (e.g., binary digits). The more values ​​per unit time in a digital signal, the more the output resembles an analog signal used to generate output sound waves. The ability to manipulate audio signals allows for the enhancement of source input signals and opens the door to synthesizing other sounds by creating audio signals from other data. Summary of the Invention

[0002] An array of resonator circuits may be provided with an excitation signal that is applied to each resonator circuit in the array. The resonator circuits may be tuned to emit different frequencies based on an input signal. The excitation signal may be a noise signal, such as pink noise. The excited resonator circuits generate a raw output signal having different frequency amplitudes across the frequency domain. A user may provide an input, such as a signal representing a particular note or combination of notes to be applied to the resonator array. The user's input is reflected in the resonator array's output signal as a higher frequency amplitude relative to the frequency corresponding to the user input.

[0003] Sound effects such as amplitude, attenuation, and phase advance may be applied to the output of the resonator array. Sound effects may be applied to selected frequencies within a frequency band. Sound effects may be applied to selected frequencies depending on the relationship between the user input and the sound effect. For example, frequencies corresponding to sounds close to the user input or an octave or other interval away from the user input. Sound effects may have the effect of generating sound waves from the modified output of the resonator array. Effects may be used to create new, previously unknown sounds and timbres. In this way, new and interesting sounds never before heard may be created. Another application may apply sound effects to capture the acoustic signature of a particular space, such as a concert hall or recording studio. In another application, sound effects may be applied to recreate the sound and timbre of a particular instrument.

[0004] According to some aspects of the present disclosure, artificial intelligence may be applied to create sound effects, such that any frequency in the frequency spectrum will have a sound effect applied to it. [Brief explanation of the drawings]

[0005] [Figure 1] FIG. 1 is a diagram of a resonator-based audio synthesizer according to an aspect of the present disclosure. [Figure 2] FIG. 1 is a diagram of a training process in a neural network for generating an audio signal from a set of resonators according to an aspect of the present disclosure. [Figure 3] FIG. 1 is a diagram of a system for generating an impulse response from a set of resonators according to an aspect of the present disclosure. [Figure 4] FIG. 1 is a diagram of the reproduction of an acoustic source in a set of resonators according to an aspect of the present disclosure. [Figure 5] FIG. 1 is a diagram of the operation of a musical synthesizer according to an aspect of the present disclosure. [Figure 6]FIG. 1 is a diagram of a computing device for implementing impulse response generation according to aspects of the present disclosure. [Figure 7] FIG. 10 is a process flow diagram for training a neural network to generate an impulse response from a set of resonators according to an aspect of the present disclosure. [Figure 8] FIG. 1 is a process flow diagram for operating a music synthesizer according to an aspect of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0006] The reverberation of an acoustic space can be defined by its impulse response or by a set of resonances called modes. The resonances define a room, a chamber, a musical instrument body, or any acoustic body. If these resonances can be reproduced, it is possible to simulate instruments, concert halls, recording studios, etc., even without access to the original source space.

[0007] FIG. 1 is a diagram of an audio synthesizer according to an embodiment of the present disclosure. The resonator array 110 includes multiple resonator circuits 111 capable of receiving an electrical input signal and converting it into a wave signal having a specific frequency. Each resonator circuit 111 has the property of oscillating at and around its resonant frequency with a larger amplitude than at other frequencies. The resonator circuits 111 can include inductors and capacitors, and the inductance and capacitance levels cause the current through the resonator circuit 111 to oscillate at the specific resonant frequency. The resonator circuits 111 may also include resistive elements that can affect the peak resonant frequency of the resonator circuit 111. Each resonator circuit 111 can be tuned to a specific resonant frequency. An acoustic resonator uses its resonant frequency to generate sound waves with a specific sound quality. The resonators can be created in the digital domain, implemented as digital filters. The resonator array 110 can include thousands or tens of thousands of resonator circuits 111, each tuned to a specific resonant frequency. When a voltage is applied to the resonator circuit 111, the components of the circuit conduct current and interact with each other to generate a frequency.

[0008] The resonator array 110 can receive an excitation signal 120 that is applied to each resonator circuit 111 in the resonator array 111. In response to the excitation signal 120, each resonator circuit 111 oscillates and generates a frequency that becomes a component of the raw output signal 115 of the resonator array 110. The excitation signal 120 can be selected to generate a particular baseline for the raw output signal 115 of the resonator array 110. In one non-limiting example, pink noise can be used as the excitation signal 110. Pink noise is a signal with a frequency spectrum in which the power of each frequency interval is inversely proportional to the frequency of the signal. Pink noise is commonly observed in nature and is commonly used to tune audio systems. Due to the nature of pink noise occurring in nature, audio systems can use it to process, filter, and / or add effects to generate desired sounds.

[0009] In addition to the excitation signal 120, the resonator array 110 can receive additional user input 130. A user can transmit musical input to the resonator array 110 using a musical keyboard, a MIDI controller, a computer interface, or other means. The user input 130 can represent a particular musical note or a group of notes, such as a chord. For example, a user can press the middle A key on a keyboard, which will amplify the output at and around 440 Hz. The represented note(s) will be applied to selected resonator circuits 111 corresponding to the note(s), generating an increased energy level at the note(s)'s corresponding frequency. The increased energy at the frequency corresponding to the user input 130 will appear in the raw output signal 115 of the resonator array 110.

[0010] The raw output signal 115 can be represented in the frequency domain 115a as an individual signal at each frequency. Each frequency may have a level of energy. Frequencies representing the user input 130 may have increased amplitude 116 relative to other frequencies that may have energy generated from the excitation signal 120.

[0011] The raw output signal 115 may be further processed to generate a processed output signal 150. One or more sound effects 140 may be applied to the raw output signal 115 to enhance or alter the processed output signal 150. Additionally, selected frequencies 141 may be identified, and the sound effects 140 are applied only to the selected frequencies 141. The selected frequencies may include frequencies occurring near the user input 116 on the frequency spectrum, or may be selected to generate the sound effects 140 at other frequencies in the spectrum, such as octaves, harmonics, selected intervals, or other modes corresponding to the user input 116.

[0012] The sound effects 140 that may be applied to the selected frequencies 141 may include attenuation, phase advance / retard, and / or amplitude modification. The sound effects 140 may be applied in combination with one another and may be strategically applied to the selected frequencies 141 to generate sound effects that work with user input 130 to generate a desired sound. The desired sound may be an effect that recreates a physical acoustic space, such as a concert hall or recording studio. In some cases, the selection of effects may recreate the sound of a particular instrument. Additionally, new sounds never before perceived may be created to generate new and interesting instrument sounds.

[0013] The resonator array 110 may include thousands of resonator circuits 111 covering a wide frequency spectrum across many frequencies. The combinations of frequencies and one or more sound effects 140 that can be applied to any number of these frequencies or frequency combinations represent a vast number of options available for generating new and exciting sounds. To aid in the discovery of new sounds and instruments, artificial intelligence (AI) 145 may be applied to apply sound effects 140 to selected frequencies 141 in the raw output signal 115. The AI ​​145 can be trained to recognize sounds and effects that are pleasing to the ear. Furthermore, the AI ​​can analyze a signal to determine the characteristics of the signal that produce a pleasing result. Using this knowledge, the AI ​​145 can select a sound effect 140 and instruct a synthesizer system to apply a particular sound effect 140 to a specific number of selected frequencies 141. The result is a processed output signal 150 that will produce a pleasing sound when processed through an audio speaker 160 or other sound-generating device.

[0014] The AI ​​145 may be in the form of a neural network. A neural network is a machine learning (ML) model that includes one or more layers of nonlinear operations to predict an output for a received input. In addition to an input layer and an output layer, some neural networks include one or more hidden layers. The output of each hidden layer can be input to another hidden layer or output layer of the neural network. Each layer of the neural network can generate a respective output from the received input depending on the value of one or more model parameters for that layer. The model parameters can be weights or biases that cause the neural network to generate an accurate output, and the weights or biases are determined through a training algorithm. In aspects of the present disclosure, the input to the ML model can be audio input, including streamed audio, pre-recorded audio, or audio as part of a video or other source or media. A machine learning model in an audio context can include separating components of an input signal, such as different voices, instruments, reverberations, harmonics, and other characteristics of the input. The model may separate different features of the audio input and enhance certain characteristics of the components to make them more or less perceptible to the ear, or the model may create new and previously unknown audio sources using information contained in the input signal. During training, the model is provided with audio samples, which may be associated with other inputs, such as audio pleasantness, based on metadata obtained from human perception of audio signals and human impressions of the input as pleasant or desirable. The model's accurate output will correspond to what the model's training indicated was desirable.

[0015] 2 illustrates a system for training an artificial intelligence 145 to create an audio signal using a set of resonators according to an embodiment of the present disclosure. Audio sources 210 are examples of sounds, tones, notes, or timbres, among other characteristics that define sound. Various audio samples 210 are provided as training data 240 for the AI ​​neural network 145. The audio samples 210 are also provided to a human 215 (or a group of humans) to determine whether the content of the audio input source 210 is pleasant 216 to the human 215 or whether the human 215 finds the audio source 210 unpleasant 217. This human feedback is stored as ground truth 230, which represents the real-world desirability of the given input audio source 210 as perceived by the actual human 215.

[0016] The AI ​​network 145 can generate outputs that are applied to the resonators 110. The set of resonators 110 may all be identical or may be controlled by parameters and inputs provided to the resonator circuits. The AI ​​network can determine the settings and parameters to apply to some of the resonators 110 to generate a desired frequency. In some cases, an impulse response may be used to establish characteristics such as the relative level, damping, and phase of the set of resonators 110. A user or the AI ​​network 145 can modify the parameters in the set of resonators 110 based on the impulse response to generate a note or other sound. The AI ​​network 145 can be a neural network 201 or similar machine learning mechanism. The neural network 201 generates a model output 202 that includes a set of resonator parameters, including audio effects 140, that, when provided to the resonator array 110, control the resonators in the array of resonators 110. When the audio effect 140 is applied to the resonator 110, the resonator 110 generates a generated audio signal 150. As an example, consider Audio Source 3 2103. Ground truth representing the desirability 216 or undesirability 217 of Audio Source 3 2103 is compared to the generated audio signal 150 to determine the difference between the ground truth 230 and the model output (generated audio signal 150). Based on the comparison, the generated audio signal 150 is characterized as a pleasant or unpleasant sounding signal. This information is provided as additional training data 240 and used to further adjust the weights and biases of the AI ​​network 145. The trained AI network 145 learns what is pleasant or unpleasant to a human listener and sends the data through the AI ​​network 145 to generate model output 202 that defines audio effects 140 to apply to selected frequencies in a frequency spectrum as discussed above with respect to FIG. 1.

[0017] Models can be trained for any number of input sources or purposes. The trained model 202 may comprise a library of pre-trained models. In the library, a user selects a desired audio source 210 and generates an input (503 in FIG. 5) that will be converted into an audio signal 150 with similar qualities to the modeled input source. In some cases, the output may represent a specific instrument. Additionally, the output may represent a specific location or landmark (audio space) from which the original input source 210 was generated. The array of resonators 110 generates frequencies across the audio spectrum. Particular resonators 110 can be centered around frequencies corresponding to notes of any musical scale. The model may specify a subset of resonators 130 that correspond to specific notes based on the input provided to the model. When considering a modal representation of an acoustic space, the selection of resonators is guided by the impulse response of the space, such that instruments contribute to the room's character. In some embodiments, this does not require the set of resonators to be generic. In these embodiments, the resonators may be perceived as thousands of small and large wind instruments, bells, string instruments, or any other source that, when excited, vibrates at audio frequencies to recreate an audio signal that matches the original source 210.

[0018] Referring to FIG. 3, an example of the use of a resonator-based synthesizer is shown. A selected input, such as a guitar 310, is associated with a given impulse response 320. The input (305 in FIG. 5) is presented to a trained model 301, which generates a model output 202. The output 302 may comprise a set of audio effects 140, including parameters for controlling the operation of one or more resonators 110. The resonators 110 generate an audio signal output 150 resulting from the application of the audio effects 140. The audio signal output 150 replicates 350 the audio signal output 150 to sound like it was produced by the guitar 310. Software may further be configured to control the resonators to "speak" or "sing." Some models may be trained to produce speech sounds. In this case, using resonators that represent specific acoustic spaces can add personality to the synthesized voice.

[0019] FIG. 4 illustrates an example of a user for an AI-enhanced resonator-based music synthesizer according to an embodiment of the present disclosure. A raw audio source, such as a musical instrument 401 or a particular acoustic space 403, produces unique sound signatures 405 that characterize the quality, timbre, or timbre of the instrument 401 or space 403. The sound signatures 405 may be formatted into a format that can be consumed by the AI ​​network 145. The model included in the AI ​​network 145 may have been trained, for example, by the training process described in FIG. 2. The AI ​​network 145 will have knowledge of human preferences regarding the pleasantness of a given sound. The AI ​​network 145 can apply this knowledge to the provided sound signatures 405 and generate audio effects to generate enhanced or additional pleasant features by applying the audio effects to selected resonators in the resonator array 110. The resonator array will generate an audio output signal 150 that can be provided to a speaker 160 of another audio device to create a perceptible sound from the audio signal 150.

[0020] Using the example of FIG. 4 , an output audio signal 150 can be generated that emulates the inherent qualities of input sources 401, 403. For example, audio space 403 can be a well-known, highly acclaimed space that has produced successful music in the past, such as a recording studio like Muscle Shoals, Motown's Hitsville USA, or Abbey Road. While the success of music produced in these spaces depends largely on the talent and creativity of the performers, the space itself has its own unique acoustic signature that contributes to the overall impression of the music. The dimensions and structural acoustics of a space create reverberations and modes of vibration that create the heart and timbre of the music produced therein. Sound characteristics 405 contain a representation of these inherent qualities and can be used to create an audio output signal 150 that sounds like it was produced in a famous space, even though in reality, sound characteristics 405 were produced in a distant location.

[0021] FIG. 5 illustrates an example using a resonator-based music synthesizer according to an embodiment of the present disclosure. An input signal may be provided to an AI network 145 by any means, including, but not limited to, a keyboard 501 or a computing device such as a MIDI controller 503. The input signal represents an audio signal, such as one or more musical notes. The input signal is processed according to a trained AI model 510 to generate a trained model output 202 that includes a set of resonator parameters that generate an audio effect 140. The audio effect is applied to selected resonators in the array of resonators 110 to generate an audio signal 150 based on the input signal, with the output having the quality of the original source on which the selected model 510 was based.

[0022] The resonator-based synthesizer may provide a user interface that presents a library of sound models 510 to the user. The library of sound models 510 may model musical instruments 511 or may emulate sounds originating from a particular acoustic space. Furthermore, the models 510 may be trained to recognize characteristics of a piece of music that are pleasing to the human ear. The synthesizer may include an input device, such as a keyboard 501 or a MIDI controller 503, or may include an input port for accepting an input device. The AI ​​network 145 receives input from the input device along with the user-selected model 510. The AI ​​network receives the user input 501, 503 and processes the input according to the selected model 510. The model output 202 includes information necessary to create the audio effect 140 to apply to selected resonators in the resonator array 110. The model output 202 may include a selection of specified frequencies corresponding to the user input 501, 503. The frequencies may include the frequency of the note entered by the user and may further include additional frequencies surrounding the user input. The additional frequencies may be notes that supplement the user input. Other effects, such as phase advance, attenuation, and amplitude, may be applied to some or all of the selected frequencies in any combination. The effects are applied as parameters to selected resonators in the resonator array 110 to generate the audio signal output 150.

[0023] FIG. 6 illustrates an example system 600 for performing source impulse response reconstruction using a resonator as described in this disclosure. System 600 may include one or more processing devices 610 configured to execute a set of instructions or executable programs. Processor 610 may be a dedicated component such as a general-purpose CPU, or an application-specific integrated circuit (ASIC), or other hardware-based processor. Although not required, special-purpose hardware components may be included to perform particular computing processes faster or more efficiently. For example, the operations of this disclosure may be performed in parallel on a computer architecture having multiple cores with parallel processing capabilities.

[0024] 7 and 8 describe various instructions in more detail along with flow diagrams. The system may further include one or more storage devices or memories 620 for storing instructions 630 and programs executed by the one or more processors 610. Additionally, the memory 620 may be configured to store data 640, such as one or more trained models 644 of the original audio source and impulse responses 642.

[0025] System 600 may further include an interface 650 for input and output of data. For example, a model may be selected in response to input to system 600 via interface 650, and an audio signal output based on the selected model and the user input may be generated as output via interface 650.

[0026] In some examples, system 600 may include a user's personal computer, laptop, tablet, or other computing device having both processor 610 and memory 620 housed therein. The operations performed by system 600 are described in more detail in the accompanying figures and description.

[0027] Other parameters and instructions may be provided to and from system 600 via interface 650. For example, parameters for controlling the collection of resonators may be identified by input provided by a user.

[0028] FIG. 7 is a flow diagram of a method for training a model of an audio source according to an embodiment of the present disclosure. An audio source is provided to a neural network (710). The audio source may be an impulse response corresponding to a particular instrument or an impulse response corresponding to a particular acoustic space. The audio source may be an audio sample labeled as to whether the audio sample is pleasant to the human ear. The audio sample may be listened to by a human, who indicates whether the audio sample is pleasant to the human. The human provides an indication that is stored and associated with the audio sample as a label. The input audio source is processed by the neural network (720) to generate an output including a set of sound effects. The sound effects may be in the form of a set of parameters for a selected number of resonator circuits in an array of resonator circuits. The resonator circuit parameters are applied to selected resonator circuits to generate a generated audio signal based on the parameters (730). The generated impulse response is compared to the source impulse response to determine a difference between the audio source and the generated audio signal (740). Based on the comparison, weights are adjusted in the neural network to more closely approximate the audio source (750).

[0029] FIG. 8 is a process flow diagram for generating an audio signal with a resonator-based synthesizer according to an embodiment of the present disclosure. A user selects (810) a model representing an audio source they wish to emulate. User input is provided to the model (820). The user input may be provided by any suitable input device, including, but not limited to, a musical keyboard or a computing device such as a MIDI controller. The model processes the input and generates (830) a set of acoustic effects in the form of resonator parameters based on the selected model and the input signal. The acoustic effects can be specified for application to a selected number of resonator circuits in the array of resonator circuits (840). In response to an excitation signal, the affected resonators in the resonator array generate frequencies that form an audio signal based on the input signal and the selected model (850).

[0030] The disclosed system allows the user to control and manipulate the set of resonators, including the amplitude / level of each resonator. Typically, this is controlled by the keyboard dynamics. Additionally, the user can control the decay time of each resonator. This can be controlled in a variety of ways, such as by a keyboard foot pedal.

[0031] To reproduce a note, a range of resonators centered around that note can sound. For example, if the A key on a keyboard corresponding to 440 Hz is pressed, a single resonator at 440 Hz can sound, or a range of resonators centered around 440 Hz can sound. The levels of the various resonators within this range can be constant, or the levels of the various resonators can be modulated by various means. The user may select a single note corresponding to the pressed key, or may select several notes spaced an octave apart. In other words, if note A on the keyboard is pressed, the instrument can output resonators at the frequency corresponding to A on the keyboard, or at frequencies corresponding to all (or any combination of) A's (55 Hz, 110 Hz, 220 Hz, 440 Hz, 880 Hz, 1760 Hz, 3520 Hz, 7040 Hz, 14080 Hz). Resonators at frequencies not associated with note A can also contribute to the synthesized note by adding tonal elements through control of the shape or envelope of the additional resonators.

[0032] In some embodiments, the envelope, timing, and level of the excitation signal may be controlled by the user. The user may decide whether the excitation is constant or applied only when a key is pressed. In the case of constant excitation, the resonator sounds instantly, while in the case of instant resonance, it strengthens (increases) upon key press. Other characteristics of the output audio signal may be controlled, including, but not limited to, global decay time, housing size, and / or sound quality / EQ.

[0033] Although the invention herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. It is therefore to be understood that numerous modifications can be made to the illustrative embodiments and that other arrangements can be devised without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. a first processor for receiving input from a user and generating a signal based on said input, said signal being applied to a plurality of resonator circuits, different resonator circuits being tuned to produce different output frequencies; a first processor further configured to generate an excitation signal that, when applied to the plurality of resonator circuits, causes one or more resonator circuits in the plurality of resonator circuits to output a signal at an associated frequency; a sound effects module for receiving frequency responses resulting from the excitation of the plurality of resonator circuits and applying one or more sound effects to selected frequencies from the frequency responses produced by the plurality of resonator circuits; 1. An audio synthesizer device comprising:

2. 2. The audio synthesizer device of claim 1, wherein the one or more sound effects are selected from one or more of a phase advance, an amplitude level, and a decay interval.

3. 10. The audio synthesizer device of claim 1, further comprising: the one or more sound effects including a set of parameters, the set of parameters including inputs to a resonator circuit of the plurality of resonator circuits.

4. The audio synthesizer device of claim 1 further comprising an input port for accepting a user input device.

5. 5. The audio synthesizer device of claim 4, wherein the user input device is a musical keyboard.

6. The audio synthesizer device of claim 4 , wherein the user input device is a Musical Instrument Digital Interface (MIDI) controller.

7. 5. The audio synthesizer device of claim 4, wherein a user input device receives input from a user, and wherein the sound effects module applies the one or more sound effects to selected frequencies that correspond to frequencies of the input from the user.

8. The audio synthesizer device of claim 1 , further comprising an artificial intelligence (AI) network in communication with the sound effects module.

9. 9. The audio synthesizer device of claim 8, wherein the AI ​​network stores a library of models, the models providing inputs to the sound effects module for applying sound effects to frequencies selected by the AI ​​network.

10. 10. The audio synthesizer device of claim 8, wherein the AI ​​network is trained using audio samples, the audio samples having labels indicating whether the audio samples contain pleasant sounds.

11. 10. The audio synthesizer device of claim 8, wherein the AI ​​network is trained to include a model that emulates a particular musical instrument.

12. 9. The audio synthesizer of claim 8, wherein the AI ​​network is trained to include a model that emulates a particular acoustic space.

13. 1. A method for generating an audio output from a plurality of resonator circuits, comprising: receiving an excitation signal for generating a frequency from at least one of the plurality of resonator circuits; applying at least one sound effect to a selected number of the plurality of resonator circuits in a sound effects module; generating an acoustic signal from the plurality of resonator circuits based on the excitation signal and the applied acoustic effect; A method comprising:

14. selecting one or more sound effects and the selected number of the plurality of resonator circuits in an artificial intelligence (AI) network model; providing the selected one or more audio effects and the selected number of resonator circuits to a sound effects module; 14. The method of claim 13, further comprising:

15. applying the selected one or more sound effects to the selected number of resonator circuits by the sound effects module; generating an audio signal output based on the sound effect and the selected frequency; 15. The method of claim 14, further comprising:

16. The method of claim 15 , wherein the one or more acoustic effects are selected from one or more of a phase advance, an amplitude level, and a decay interval.

17. training the AI ​​network with a plurality of audio samples, each audio sample being labeled to indicate whether the audio sample is pleasant to the human ear; 16. The method of claim 15, further comprising:

18. generating the audio signal output from a model of the AI ​​network, the model being trained to generate audio samples that are pleasing to the human ear; 20. The method of claim 17, further comprising:

19. generating the audio signal output from a model of the AI ​​network, the model being trained to generate audio samples that emulate a particular musical instrument; 16. The method of claim 15, further comprising:

20. generating the audio signal output from a model of the AI ​​network, the model being trained to generate audio samples that emulate a particular audio space; 16. The method of claim 15, further comprising:

Images