Mechanical music creation table integrating AI composition assistance and entity phonetic sequence feedback

Abstract

The invention discloses a mechanical music creation platform integrating AI composition assistance and entity phonetic sequence feedback. The mechanical music creation platform comprises a mechanical supporting platform, an AI composition assistance module, an entity phonetic sequence feedback module, a mechanical operation execution module and a data interaction bus. The mechanical supporting platform is of an integrated frame structure, an operation panel is arranged on the top, a function module mounting cavity is formed in the middle, and a damping base is arranged at the bottom. The AI composition auxiliary module is embedded into the functional module mounting cavity and comprises a music data processing unit, a style model library, a real-time generation unit and an instruction receiving unit, and the music data processing unit is used for analyzing melody fragments, harmony trend or style instructions input by a user; through an AI composition auxiliary module and an entity phonetic sequence feedback module, after a user inputs an instruction, the AI generates a music score in real time and drives a machine to produce sound, and meanwhile, the user visually perceives a creation effect through multi-mode feedback (visual sense, tactile sense and audio frequency).

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of music creation, in particular to a mechanical music creation table integrating AI composition assistance and physical sequencer feedback. BACKGROUND

[0002] With the development of digital music technology, AI composition tools and mechanical music devices have gradually become important auxiliary means for music creation. In the prior art, AI composition tools mostly exist in the form of software, generating scores or melodies through algorithms, but lack physical operation feedback, and users cannot intuitively adjust the creation parameters through mechanical interaction, resulting in insufficient creation immersion. Traditional mechanical music devices (such as electronic keyboards and sequencers) have physical operation interfaces, but have low intelligence and cannot provide active assistance functions such as style adaptation and harmony generation. Moreover, the sequencer feedback form is single, mostly relying on visual or audio feedback, lacking multi-modal feedback mechanisms such as tactile feedback.

[0003] At the same time, the prior art has the following defects: 1) AI composition and physical operation are disconnected, making it difficult to adjust the generated score data in real time through mechanical devices, and the creation process is fragmented; 2) The parameter adjustment structure of the physical sequencer is not reasonably designed, mostly being single-function knobs, with low operation efficiency and lacking personalized feedback mode configuration. Therefore, a mechanical music creation table integrating AI composition assistance and physical sequencer feedback is needed to solve the above problems. SUMMARY

[0004] The present application aims to provide a mechanical music creation table integrating AI composition assistance and physical sequencer feedback to solve the problems raised in the background art.

[0005] To achieve the above object, the present application provides the following technical scheme: an integrated AI composition assistance and entity sequencer feedback mechanical music creation platform, characterized in that: comprising a mechanical support platform, an AI composition assistance module, an entity sequencer feedback module, a mechanical operation execution module and a data interaction bus; the mechanical support platform is an integrated frame structure, provided with an operation panel at the top, a functional module installation cavity in the middle and a shock-absorbing base at the bottom; the AI composition assistance module is embedded in the functional module installation cavity and comprises a music data processing unit, a style model library, a real-time generation unit and an instruction receiving unit, the music data processing unit is used for analyzing the melody segment, the harmonic direction or the style instruction input by the user, the style model library pre-stores training models of at least 8 types of music styles such as classical, popular and electronic, the real-time generation unit calls the corresponding style model based on the analysis result to generate multi-track music score data and sequencer control signals; the entity sequencer feedback module is provided on the operation panel and comprises a mechanical knob array, an LED sequencer indicator light group and a tactile feedback motor, the mechanical knob array corresponds to the sequencer track one by one and can manually adjust the sequencer speed, the beat strength and the timbre parameters, the LED sequencer indicator light group synchronously displays the sequencer progress of the current playback track according to the time axis, the tactile feedback motor produces vibration feedback synchronously with the sequencer beat, and the vibration frequency is positively correlated with the beat strength; the mechanical operation execution module comprises a liftable key unit, a percussion trigger mechanical arm and a stringed instrument plucking mechanism, corresponding to the mechanical sound generation simulation of keyboard, percussion and stringed instruments respectively; the data interaction bus is electrically connected with the AI composition assistance module, the entity sequencer feedback module and the mechanical operation execution module respectively to realize the real-time transmission of music score data, control signals and operation feedback data; the music score data generated by the AI composition assistance module is synchronously transmitted to the entity sequencer feedback module through the data interaction bus to drive the LED sequencer indicator light group and the tactile feedback motor to work, while controlling the mechanical operation execution module to perform the corresponding sound generation action, after the user manually adjusts the parameters through the mechanical knob array, the feedback data is transmitted back to the AI composition assistance module through the data interaction bus to optimize the music score generation result in real time.

[0006] Preferably, the AI composition assistance module further comprises a music style recognition unit, a harmony adaptation unit, and a real-time modification unit; the music style recognition unit receives user humming melodies or external instrument input signals through an audio acquisition module, extracts audio features using a convolutional neural network (CNN), and matches the optimal style model in the style model library; the harmony adaptation unit automatically adapts corresponding harmony solutions based on generated melody data, and provides at least three harmony variants for user selection; the real-time modification unit supports user modification of melody fluctuations, rhythm patterns, and arrangement proportions of generated scores through touch instructions, voice instructions, or mechanical knob operations; the modified parameters are synchronized in real time to the physical sequencer feedback module and the mechanical operation execution module; the AI composition assistance module also has a built-in external instrument database interface that can import timbre sampling data of at least 15 types of instruments such as guitars, pianos, and violins, and combine with the sound simulation structure of the mechanical operation execution module to achieve personalized arrangement combinations.

[0007] Preferably, the mechanical knob array of the physical sequencer feedback module adopts a layered structure design, including a track selection layer, a parameter adjustment layer, and a feedback mode layer; the track selection layer has 8-16 independent knobs corresponding to 8-16 sequencer tracks, and rotating the knob can switch the current operation track; the parameter adjustment layer includes a speed adjustment knob (range 60-240 BPM), a volume adjustment knob (range 0-100 dB), and a timbre filter knob (frequency range 20 Hz-20 kHz), which uses an optical encoder to achieve stepless adjustment with an adjustment accuracy of ≤0.1 unit; the feedback mode layer has three function knobs to control the brightness of the LED indicator light (3-level adjustable), the vibration intensity of the tactile feedback motor (5-level adjustable), and the synchronization delay of the sequencer feedback (0-50 ms adjustable); the physical sequencer feedback module also has a sequencer visual display screen using an OLED touch screen to display the waveform diagram, beat distribution, and parameter values of the current sequencer in real time, and supports user direct drag-and-drop modification of sequencer nodes through touch.

[0008] Preferably, the liftable key unit of the mechanical operation execution module comprises a 25-88 key adjustable key group, an electromagnetic drive assembly, and a damping adjustment mechanism; the key group adopts a counterweight design, the key stroke range is adjustable from 2-10 mm, the electromagnetic drive assembly receives the score data of the AI composition assistance module, drives the keys to strike and sound at a preset force and speed, the striking force range is 20-127 velocity; the damping adjustment mechanism changes the movement resistance of the keys by rotating the adjustment rod to adapt to the playing habits of different users; the percussion trigger mechanical arm is provided with 3-6 independent mechanical arms, each mechanical arm is provided with replaceable striking heads (rubber, wool, and metal materials can be selected), is driven by a stepping motor, and the striking force and frequency are controlled by the parameters of the physical sound sequence feedback module; the string plucking mechanism comprises a string fixing frame, an electromagnetic plectrum, and a tension adjustment assembly, the string is made of nylon or metal material, the plucking angle (0-45°) and speed (0-1 m / s) of the electromagnetic plectrum can be adjusted by the AI module or a manual knob, and the tension adjustment assembly can change the vibration frequency of the string to realize fine tuning of the pitch.

[0009] Preferably, the interactive control module is further provided, and the interactive control module comprises a touch display screen, a voice instruction unit, a gesture recognition camera, and a foot controller interface; the touch display screen is arranged in the middle of the operation panel, has a size of ≥10.1 inches, a resolution of ≥1920x1200, supports multi-point touch control, and is used for displaying an AI composition interface, a sound sequence editing interface, and a device parameter setting interface; the voice instruction unit is provided with a microphone array and a voice recognition module, supports recognition of at least 50 specific instructions such as “generate a popular style melody”, “increase the sound sequence speed”, and “modify the harmonic direction”, has an identification accuracy of ≥95%, and has a response time of ≤1 s; the gesture recognition camera is arranged on the top of the mechanical support platform, adopts a 3D depth camera, supports 6 preset gesture controls such as hand waving to switch tracks, clenching to pause / play, and sliding to adjust the volume; the foot controller interface is a USB-C expansion interface, can be connected with an external foot pedal, realizes hands-free operation of functions such as sound sequence starting and stopping and recording triggering, and the like; the interactive control module and the AI composition assistance module realize bidirectional communication, and the instructions input by the user through any interactive mode are synchronously transmitted to the AI module in real time to drive the execution of related functions.

[0010] Preferably, the data interaction bus adopts a CAN bus and Ethernet dual bus architecture, the CAN bus is used for transmitting real-time control signals (audio sequence parameters, mechanical operation instructions), the transmission delay is ≤10 ms, and the Ethernet is used for transmitting large-capacity music data (music score files, tone sampling, style models), the transmission rate is ≥1 Gbps; the mechanical music creation platform is also provided with a data processing and storage module, including an edge computing unit, an encrypted storage unit, and a historical creation track backtracking unit; the edge computing unit adopts an FPGA chip, is used for accelerating the inference operation of an AI composition model, reduces the dependence on the cloud, and the response time of locally generating a music score is ≤3 s; the encrypted storage unit has a capacity of ≥1 TB, supports AES-256 encryption, stores music score files, custom style models, and operation parameter configurations created by a user; the historical creation track backtracking unit can record the whole-process operation data of a user from initial instruction input to final work generation, including each version of a music score generated by AI, parameter adjustment records of an entity audio sequence, and execution logs of mechanical operation, supports time node backtracking and viewing, and can automatically generate creation suggestions based on historical tracks.

[0011] Preferably, the mechanical support platform adopts a modular assembly design, including an operation table panel, a functional module support, a damping base, and foldable side arms; the operation table panel is made of carbon fiber material, is provided with an anti-slip wear-resistant coating on the surface, and is provided with mounting grooves for mechanical knobs, a touch display screen, and a mechanical operation execution module, groove sizes can be adapted to different specifications of components through adjusting shims; the functional module support adopts an aluminum alloy frame, is provided with heat dissipation and ventilation holes, and is provided with a cable storage channel inside, so as to avoid cable winding; the damping base is internally provided with a rubber damping pad and a spring damping mechanism, can reduce vibration conduction when the mechanical operation execution module works, and the vibration attenuation rate is ≥80%; the foldable side arms are arranged on both sides of the operation table panel, can be folded upwards to be parallel to the panel (to save storage space) or unfolded downwards to be horizontal (to increase the area of the operation table), and the side arms are provided with USB interfaces, audio output interfaces, and external instrument connection interfaces on the surface.

[0012] Preferably, the AI composition assistance module further comprises a custom style training interface and a collaborative creation unit; the custom style training interface supports users to import personally created music score files (formats including MIDI, MusicXML, WAV) and train a dedicated style model through a transfer learning algorithm. The training process can set the number of iterations (100-1000 adjustable) and the learning rate (0.001-0.01 adjustable). The trained model can be saved to the style model library for subsequent creation calls. The collaborative creation unit supports multiple creation stations to be connected through a local area network or a cloud server, realizing real-time collaboration of multiple people. One of the creation stations serves as a host, responsible for unified style model calling and audio sequence synchronization. The other creation stations serve as slaves and can edit different audio tracks respectively. The edited content is synchronized to the host and other slaves in real time. The host can integrate and optimize all audio tracks to generate the final work. The collaborative creation unit also supports permission settings. The host can assign editing permissions, viewing permissions, or read-only permissions to different slave users.

[0013] Preferably, the physical audio sequence feedback module further comprises an audio feedback unit and an error prompt unit. The audio feedback unit is built-in with a full-frequency loudspeaker (frequency response range 50Hz-20kHz) and a headphone interface, which can play the sound effects of the mechanical operation execution module and the AI-generated music score audio in real time, support stereo output, and synchronize the loudspeaker volume with the volume parameter of the physical audio sequence. The error prompt unit includes a sound and light alarm component. When the AI composition assistance module cannot recognize user instructions, data transmission abnormalities occur, or mechanical components fail, the LED audio sequence indicator light group flashes red light, the loudspeaker emits a prompt sound (frequency 1kHz, duration 0.5s), and the touch display screen displays the fault type and troubleshooting suggestions. The physical audio sequence feedback module also supports audio sequence export function, which can export the edited audio sequence data as MIDI, MP3 or WAV format files and transmit them to external devices through USB interface or wireless communication module.

[0014] Preferably, it also includes a power supply and heat dissipation optimization module; the power supply module supports dual power supply modes, including mains power supply (AC 100-240V) and lithium battery power supply (capacity ≥20000mAh, endurance time ≥8 hours), and is provided with a power management unit that can automatically switch power supply modes, and the lithium battery supports fast charging (30 minutes to 60%) and overcharge and overdischarge protection; the heat dissipation optimization module includes a heat dissipation fan (rotating speed 1000-3000rpm adjustable) in the functional module installation cavity, a heat dissipation channel of the mechanical operation execution module, and a heat-conducting silica gel pad of the AI module; the heat dissipation fan is linked with a temperature sensor, and is automatically started when the temperature in the cavity is ≥45℃ and is automatically turned off when the temperature is ≤30℃; the heat dissipation channel adopts a labyrinth design, which ensures heat dissipation efficiency and prevents dust from entering; the heat-conducting silica gel pad is attached to the core chip of the AI composition assistance module, and has a heat conductivity coefficient ≥2.0W / (m·K), ensuring temperature stability during AI model operation; the overall size of the mechanical music creation table is ≤120cm×60cm×80cm, and the weight is ≤30kg, which is suitable for various scenes such as home studio and mobile creation.

[0015] Compared with the prior art, the present scheme designs a mechanical music creation table integrating AI composition assistance and entity sound sequence feedback, which has the following beneficial effects: (1) Realize closed-loop cooperation of AI composition and entity interaction: through deep linkage of the AI composition assistance module, the entity sound sequence feedback module and the mechanical operation execution module, after the user inputs an instruction, the AI generates a score in real time and drives the mechanical sound generation, at the same time, through multi-modal feedback (visual, tactile, audio), the user can intuitively perceive the creation effect, and after manually adjusting the parameters, the generation result is optimized in real time, solving the problem of the separation of intelligence and entity interaction in the prior art, making the creation process more coherent and improving the efficiency by more than 40%; (2) The entity sound sequence feedback function is more flexible: the layered mechanical knob array realizes efficient adjustment of multiple tracks and multiple parameters, cooperates with the LED sound sequence indicator light group, the tactile feedback motor and the visual display screen, provides a multi-modal feedback mechanism, supports personalized feedback mode configuration, meets the operation habits and creation scene needs of different users, and the operation efficiency is improved by 2 times compared with the traditional single knob design. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0017] The application provides a technical solution: a mechanical music creation table integrating AI composition assistance and entity sequencer feedback, characterized by comprising a mechanical support platform, an AI composition assistance module, an entity sequencer feedback module, a mechanical operation execution module and a data interaction bus; the mechanical support platform is an integrated frame structure, provided with an operation panel at the top, a functional module installation cavity in the middle and a shock-absorbing base at the bottom; the AI composition assistance module is embedded in the functional module installation cavity and comprises a music data processing unit, a style model library, a real-time generation unit and an instruction receiving unit; the music data processing unit is used for analyzing a melody segment, a harmonic trend or a style instruction input by a user; the style model library pre-stores training models of at least 8 types of music styles such as classical, popular and electronic; the real-time generation unit calls a corresponding style model based on an analysis result to generate multi-track music score data and sequencer control signals; the entity sequencer feedback module is arranged on the operation panel and comprises a mechanical knob array, an LED sequencer indicator light group and a tactile feedback motor; the mechanical knob array corresponds to a sequencer track one by one and can manually adjust sequencer speed, beat strength and timbre parameters; the LED sequencer indicator light group synchronously displays a sequencer progress of a current playing track according to a time axis; the tactile feedback motor produces vibration feedback synchronously with a sequencer beat, and the vibration frequency is positively correlated with the beat strength; the mechanical operation execution module comprises a liftable key unit, a percussion trigger mechanical arm and a stringed instrument plucking mechanism, which correspond to mechanical sound production simulation of keyboard, percussion and stringed instruments respectively; the data interaction bus is electrically connected with the AI composition assistance module, the entity sequencer feedback module and the mechanical operation execution module respectively to realize real-time transmission of music score data, control signals and operation feedback data; the music score data generated by the AI composition assistance module is synchronously transmitted to the entity sequencer feedback module through the data interaction bus to drive the LED sequencer indicator light group and the tactile feedback motor to work and simultaneously control the mechanical operation execution module to perform corresponding sound production actions; after a user manually adjusts parameters through the mechanical knob array, feedback data is transmitted back to the AI composition assistance module through the data interaction bus to optimize music score generation results in real time.

[0018] The AI composition assistance module further comprises a music style recognition unit, a harmony adaptation unit, and a real-time modification unit; the music style recognition unit receives a user humming melody or an external instrument input signal through an audio acquisition module, extracts audio features using a convolutional neural network (CNN), and matches an optimal style model in a style model library; the harmony adaptation unit automatically adapts corresponding harmony generation schemes based on generated melody data, and provides at least three harmony variants for user selection; the real-time modification unit supports user modification of melody fluctuations, rhythm patterns, and arrangement proportions of generated music scores through touch instructions, voice instructions, or mechanical knob operations, and the modified parameters are synchronized in real time to an entity sequencer feedback module and a mechanical operation execution module; the AI composition assistance module further comprises an external instrument database interface, which can import timbre sampling data of at least 15 types of instruments such as guitars, pianos, and violins, and realize personalized arrangement combination in combination with the sound simulation structure of the mechanical operation execution module.

[0019] The mechanical knob array of the entity sequencer feedback module adopts a layered structure design, including a track selection layer, a parameter adjustment layer, and a feedback mode layer; the track selection layer is provided with 8-16 independent knobs corresponding to 8-16 sequencer tracks, and the current operation track can be switched by rotating the knob; the parameter adjustment layer includes a speed adjustment knob (range 60-240 BPM), a volume adjustment knob (range 0-100 dB), and a timbre filter knob (frequency range 20 Hz-20 kHz), which adopts an optical encoder to realize stepless adjustment, and the adjustment accuracy is ≤0.1 unit; the feedback mode layer is provided with three function knobs to control the brightness of the LED indicator light (3-level adjustable), the vibration intensity of the tactile feedback motor (5-level adjustable), and the synchronization delay of the sequencer feedback (0-50 ms adjustable); the entity sequencer feedback module is further provided with a sequencer visual display screen, which adopts an OLED touch screen to display the waveform diagram, beat distribution, and parameter values of the current sequencer in real time, and supports user direct drag modification of the sequencer node through touch.

[0020] The liftable key unit of the mechanical operation execution module comprises a 25-88 key adjustable key group, an electromagnetic driving assembly and a damping adjustment mechanism; the key group adopts a counterweight design, the key stroke range is adjustable from 2-10 mm, the electromagnetic driving assembly receives the music score data of the AI composition auxiliary module, drives the key to strike and sound at a preset force and speed, the striking force range is 20-127 velocity; the damping adjustment mechanism changes the key movement resistance by rotating the adjusting rod, and adapts to the playing habits of different users; the percussion trigger mechanical arm is provided with 3-6 independent mechanical arms, a replaceable striking head (rubber, wool, metal material can be selected) is arranged at the end of each mechanical arm, the striking force and frequency are controlled by the parameters of the physical sound sequence feedback module through the stepping motor driving; the string plucking mechanism comprises a string fixing frame, an electromagnetic plectrum and a tension adjustment assembly, the string is made of nylon or metal material, the plucking angle (0-45°) and speed (0.1-1 m / s) of the electromagnetic plectrum can be adjusted through the AI module or the manual knob, and the tension adjustment assembly can change the string vibration frequency to realize fine tuning of the pitch.

[0021] The interactive control module is also included, and the interactive control module comprises a touch display screen, a voice instruction unit, a gesture recognition camera and a foot pedal controller interface; the touch display screen is arranged in the middle part of the operation panel, the size is ≥10.1 inches, the resolution is ≥1920×1200, multi-point touch is supported, and the touch display screen is used for displaying an AI composition interface, a sound sequence editing interface and a device parameter setting interface; the voice instruction unit is provided with a microphone array and a voice recognition module, supports identification of at least 50 specific instructions such as “generating a popular style melody”, “increasing the sound sequence speed” and “modifying the harmonic direction”, the identification accuracy is ≥95%, and the response time is ≤1s; the gesture recognition camera is arranged on the top of the mechanical support platform, adopts a 3D depth camera, supports 6 kinds of preset gesture controls such as hand waving to switch tracks, clenching to pause / play, and sliding to adjust the volume; the foot pedal controller interface is a USB-C expansion interface, can be connected with an external foot pedal, realizes double-hand free operation of functions such as sound sequence start / stop and recording triggering; the interactive control module and the AI composition auxiliary module realize bidirectional communication, and the instructions input by the user through any interactive mode are synchronized to the AI module in real time, and drive the related functions to be executed.

[0022] The data interaction bus adopts CAN bus and Ethernet dual bus architecture, CAN bus is used for transmitting real-time control signals (audio sequence parameters, mechanical operation instructions), transmission delay ≤10ms, Ethernet is used for transmitting large capacity music data (music score files, tone sampling, style model), transmission rate ≥1Gbps; The mechanical music creation platform also has a data processing and storage module, including an edge computing unit, an encrypted storage unit and a historical creation trajectory backtracking unit; The edge computing unit adopts FPGA chip, which is used to accelerate the inference operation of AI composition model, reduce the dependence on cloud, and the response time of local music score generation is ≤3s; The encrypted storage unit capacity ≥1TB, supports AES-256 encryption, stores user created music score files, custom style models and operation parameter configurations; The historical creation trajectory backtracking unit can record the whole process operation data of the user from the initial instruction input to the final work generation, including each version of the music score generated by AI, the parameter adjustment record of the entity audio sequence, and the execution log of the mechanical operation, supports time node backtracking and viewing, and can automatically generate creation suggestions based on the historical trajectory.

[0023] The mechanical support platform adopts modular assembly design, including an operation table panel, a function module support, a damping base and a foldable side arm; The operation table panel is made of carbon fiber material, and the surface is provided with an anti-slip wear-resistant coating, and the installation groove of the mechanical knob array, the touch display screen and the mechanical operation execution module is reserved, and the groove size can be adapted to different specifications of components through adjusting gaskets; The function module support adopts an aluminum alloy frame, is provided with heat dissipation vents, and is internally provided with a cable storage channel to avoid cable entanglement; The damping base is internally provided with a rubber damping pad and a spring damping mechanism, which can reduce the vibration conduction of the mechanical operation execution module during work, and the vibration attenuation rate is ≥80%; The foldable side arm is arranged on both sides of the operation table panel, and can be folded upwards to be parallel to the panel (to save storage space) or unfolded downwards to be horizontal (to increase the area of the operation table), and the surface of the side arm is provided with a USB interface, an audio output interface and an external musical instrument connection interface.

[0024] The AI composition assistance module further comprises a custom style training interface and a collaborative creation unit; the custom style training interface supports users to import personal creation music score files (formats including MIDI, MusicXML, WAV), trains a dedicated style model through a transfer learning algorithm, the training process can set the number of iterations (100-1000 adjustable), the learning rate (0.001-0.01 adjustable), the trained model can be saved to the style model library for subsequent creation calling; the collaborative creation unit supports multiple creation stations to be connected through a local area network or a cloud server, realizing real-time collaboration of multiple people, one of the creation stations acts as a host, responsible for unified style model calling and audio sequence synchronization, the other creation stations act as slaves, can respectively edit different audio tracks, the editing content is synchronized to the host and other slaves in real time, the host can integrate and optimize all audio tracks to generate a final work; the collaborative creation unit also supports permission setting, the host can assign editing permission, viewing permission or read-only permission to different slave users.

[0025] The entity audio sequence feedback module further comprises an audio feedback unit and an error prompt unit; the audio feedback unit is built-in with a full-frequency loudspeaker (frequency response range 50Hz-20kHz) and a headphone interface, can play the sound effect of the mechanical operation execution module and the AI generated music score audio in real time, supports stereo output, the loudspeaker volume is synchronously adjusted with the volume parameter of the entity audio sequence; the error prompt unit comprises a sound and light alarm component, when the AI composition assistance module cannot recognize user instructions, data transmission abnormity occurs or mechanical components fail, the LED audio sequence indicator light group flashes red light, the loudspeaker emits a prompt sound (frequency 1kHz, duration 0.5s), and the touch display screen displays the fault type and troubleshooting suggestions; the entity audio sequence feedback module also supports an audio sequence export function, can export the edited audio sequence data into a MIDI, MP3 or WAV format file, and transmit it to external equipment through a USB interface or a wireless communication module.

[0026] It also includes a power supply and heat dissipation optimization module; the power supply module supports dual power supply mode, including mains power supply (AC 100-240V) and lithium battery power supply (capacity ≥20000mAh, endurance time ≥8 hours), and is provided with a power management unit, which can automatically switch the power supply mode, and the lithium battery supports fast charging (30 minutes to 60%) and overcharge and overdischarge protection; the heat dissipation optimization module includes a heat dissipation fan (rotating speed 1000-3000rpm adjustable) in the functional module installation cavity, a heat dissipation channel of the mechanical operation execution module and a heat-conducting silica gel pad of the AI module; the heat dissipation fan is linked with the temperature sensor, and is automatically started when the temperature in the cavity is ≥45℃ and is automatically turned off when the temperature is ≤30℃; the heat dissipation channel adopts a labyrinth design, which not only ensures the heat dissipation efficiency, but also prevents dust from entering; the heat-conducting silica gel pad is attached to the core chip of the AI composition auxiliary module, and the heat conductivity coefficient is ≥2.0W / (m・K), which ensures the temperature stability during AI model operation; the overall size of the mechanical music creation platform is ≤120cm×60cm×80cm, and the weight is ≤30kg, which is suitable for various scenes such as home studio, mobile creation, etc.

[0027] Example 1: Single-person creation process Device startup: connect the mains or start the lithium battery power supply, the device automatically enters the standby state, the touch display screen displays the main interface, including "AI composition", "audio sequence editing", "device setting" three function entrances; Creation instruction input: the user selects the "AI composition" entrance, inputs the voice instruction "generate a 4-bar pop style melody, speed 120BPM", the voice instruction unit recognizes and transmits the instruction to the AI composition auxiliary module; AI music score generation: the music data processing unit of the AI composition auxiliary module analyzes the instruction, calls the pop style model in the style model library, generates 4-bar melody data and corresponding 3 kinds of harmonic variants within 3 seconds through FPGA accelerated inference, and synchronously transmits to the physical audio sequence feedback module and the mechanical operation execution module; Physical feedback and mechanical sound production: the LED audio sequence indicator light group of the physical audio sequence feedback module flashes synchronously at 120BPM rhythm, the tactile feedback motor produces corresponding beat vibration, and the audio feedback unit plays the sound production effect of the mechanical operation execution module (the ascending and descending piano key unit simulates the piano melody, and the percussion trigger mechanical arm simulates the drum points); the user selects the second harmonic variant through the touch screen, and rotates the parameter adjustment layer of the physical rotary knob array to adjust the volume to 80dB and the timbre filter frequency to 5kHz; Local modification and optimization: the user switches to the melody track by waving his hand through the gesture recognition camera, selects the melody track by rotating the track selection layer knob, modifies the melody pitch of the 3rd bar through the touch screen, and the real-time modification unit synchronously transmits the adjusted parameters to each module, and the mechanical operation execution module immediately updates the sound production effect; Work export: after the creation is completed, the user selects the "export" function through the touch screen, exports the sequencer data into a MIDI format file, and transmits it to the computer through the USB interface or uploads it to the cloud storage through WiFi6.

[0028] Embodiment 2: Multi-person collaborative creation process Device networking: three creation tables are connected through a local area network, one of which is set as a host, and the other two are set as slaves. The host user enters the "collaborative creation" mode through the touch screen and assigns permissions: the host has editing + integration permissions, slave 1 has melody track editing permissions, and slave 2 has rhythm track editing permissions. Style uniformity setting: the host user selects a jazz style model, sets the global speed to 140 BPM, and sets the harmonic progression to II-V-I, and configures the parameters to be synchronized to all slaves. Track editing: slave 1 connects a guitar through an external connection device and plays a melody. The audio acquisition module receives the signal and extracts the features, and the AI composition assistance module optimizes the melody details to generate melody track data. Slave 2 adjusts the parameters of the percussion instrument through the mechanical knob array, sets the bass drum frequency to 80 Hz, and sets the snare drum intensity to 60 velocity to generate rhythm track data. The edited content at both ends is synchronized to the host in real time. Host integration and optimization: the host user views the synchronized melody track and rhythm track, adds bass track harmonies through the harmonic adaptation unit, adjusts the volume ratio of each track (melody track 40%, rhythm track 30%, bass track 30%), and listens to the overall effect through the physical sequencer feedback module. Work saving and sharing: after integration, the host user saves the work to the local encrypted storage unit, exports it as a WAV format file, and shares it with the slave users through the local area network, realizing a multi-person collaborative creation closed loop.

[0029] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims, not the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any mark in the claims should not be considered as limiting the involved claims.

[0030] In addition, it should be understood that although the present specification is described in terms of embodiments, not every embodiment contains only one independent technical solution, and the description of the specification is only for the sake of clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. A mechanical music creation station integrating AI composition assistance and physical sequencer feedback, characterized in that: It comprises a mechanical support platform, an AI composition auxiliary module, an entity sequencer feedback module, a mechanical operation execution module and a data interaction bus; the mechanical support platform is an integrated frame structure, provided with an operation panel at the top, a functional module installation cavity in the middle and a shock-absorbing base at the bottom; The AI composition auxiliary module is embedded in the functional module installation cavity and comprises a music data processing unit, a style model library, a real-time generation unit and an instruction receiving unit; the music data processing unit is used for analyzing the melody segment, the chord progression or the style instruction input by the user; the style model library pre-stores training models of at least 8 types of music styles such as classical, popular and electronic; the real-time generation unit calls the corresponding style model based on the analysis result to generate multi-track music score data and sequencer control signals; the entity sequencer feedback module is arranged on the operation panel and comprises a mechanical knob array, an LED sequencer indicator light group and a tactile feedback motor; the mechanical knob array corresponds to the sequencer track one by one and can manually adjust the sequencer speed, the beat strength and the timbre parameters; the LED sequencer indicator light group synchronously displays the sequencer progress of the current playing track according to the time axis; the tactile feedback motor produces vibration feedback synchronously with the sequencer beat, and the vibration frequency is positively correlated with the beat strength; the mechanical operation execution module comprises a liftable key unit, a percussion instrument triggering mechanical arm and a string instrument plucking mechanism, which respectively correspond to the mechanical sound generation simulation of keyboard, percussion and stringed instruments; the data interaction bus is electrically connected with the AI composition auxiliary module, the entity sequencer feedback module and the mechanical operation execution module to realize the real-time transmission of music score data, control signals and operation feedback data; the music score data generated by the AI composition auxiliary module is synchronously transmitted to the entity sequencer feedback module through the data interaction bus to drive the LED sequencer indicator light group and the tactile feedback motor to work, while controlling the mechanical operation execution module to execute the corresponding sound generation action; after the user manually adjusts the parameters through the mechanical knob array, the feedback data is transmitted back to the AI composition auxiliary module through the data interaction bus to optimize the music score generation result in real time.

2. The mechanical music creation station of claim 1, wherein: The AI composition auxiliary module further comprises a music style recognition unit, a chord adaptation unit and a real-time modification unit; the music style recognition unit receives the user's humming melody or external instrument input signal through an audio acquisition module, extracts audio features by using a convolutional neural network (CNN) and matches the optimal style model in the style model library; the chord adaptation unit automatically adapts the corresponding chord generation scheme based on the generated melody data and provides at least 3 chord variants for the user to select; the real-time modification unit supports the user to modify the melody fluctuation, rhythm type and arrangement proportion of the generated music score through touch instructions, voice instructions or mechanical knob operations; the modified parameters are synchronously transmitted to the entity sequencer feedback module and the mechanical operation execution module in real time; the AI composition auxiliary module further has a built-in external instrument database interface, which can import timbre sampling data of at least 15 types of instruments such as guitar, piano and violin, and realizes personalized arrangement combination in combination with the sound generation simulation structure of the mechanical operation execution module.

3. The mechanical music creation station of claim 2, wherein: The mechanical knob array of the entity sequencer feedback module adopts a hierarchical structure design, including a track selection layer, a parameter adjustment layer, and a feedback mode layer; the track selection layer is provided with 8-16 independent knobs, corresponding to 8-16 audio sequence tracks, and the current operation track can be switched by rotating the knob; the parameter adjustment layer includes a speed adjustment knob (range 60-240 BPM), a volume adjustment knob (range 0-100 dB), and a timbre filter knob (frequency range 20 Hz-20 kHz), which adopts an optical encoder to realize stepless adjustment, and the adjustment accuracy is ≤0.1 unit; the feedback mode layer is provided with three function knobs, respectively controlling the brightness of the LED indicator light (three-stage adjustable), the vibration intensity of the tactile feedback motor (five-stage adjustable), and the synchronization delay of the audio sequence feedback (0-50 ms adjustable); the entity sequencer feedback module is also provided with a visual display screen of the audio sequence, which adopts an OLED touch screen to display the waveform diagram, the beat distribution, and the parameter values of the current audio sequence in real time, and supports the user to directly drag and modify the audio sequence nodes through the touch method.

4. The mechanical music creation station of claim 1, wherein: The liftable key unit of the mechanical operation execution module includes a 25-88 key adjustable key group, an electromagnetic drive assembly, and a damping adjustment mechanism; the key group adopts a counterweight design, and the key stroke range is 2-10 mm adjustable; the electromagnetic drive assembly receives the music score data of the AI composition assistance module, drives the key to strike and sound at a preset force and speed, and the striking force range is 20-127 velocity; the damping adjustment mechanism changes the key movement resistance by rotating the adjusting rod, and adapts to the playing habits of different users; the percussion trigger mechanical arm is provided with 3-6 independent mechanical arms, each mechanical arm is provided with a replaceable striking head (rubber, wool, and metal materials are optional) at the end, is driven by a stepping motor, and the striking force and frequency are controlled by the parameters of the entity sequencer feedback module; the string plucking mechanism includes a string fixing frame, an electromagnetic plectrum, and a tension adjustment assembly, the string is made of nylon or metal material, the plucking angle (0-45°) and speed (0.1-1 m / s) of the electromagnetic plectrum can be adjusted by the AI module or the manual knob, and the tension adjustment assembly can change the string vibration frequency to realize the fine tuning of the pitch.

5. The mechanical music creation station of claim 1, wherein: Also includes an interaction control module, the interaction control module includes a touch display screen, a voice instruction unit, a gesture recognition camera and a foot controller interface; The touch display screen is arranged in the middle of the operation panel, the size is ≥10.1 inches, the resolution is ≥1920×1200, supports multi-point touch, is used to display the AI composition interface, the sound sequence editing interface and the device parameter setting interface;The voice instruction unit is built-in microphone array and voice recognition module, supports at least 50 specific instruction recognition such as "generating popular style melody", "increasing sound sequence speed", "modifying harmonic direction", the recognition accuracy is ≥95%, the response time is ≤1s;The gesture recognition camera is arranged on the top of the mechanical support platform, adopts 3D depth camera, supports 6 kinds of preset gesture control such as hand switching track, clenched fist pause / play, sliding volume adjustment;The foot controller interface is USB-C expansion interface, can connect external foot pedal, realize double-hand liberation operation of sound sequence start / stop, recording trigger and other functions;The interaction control module and the AI composition auxiliary module communicate bidirectionally, the instructions input by the user through any interaction mode are all synchronized to the AI module in real time, drive related functions to execute.

6. The mechanical music creation station of claim 1, wherein: The data interaction bus adopts CAN bus and Ethernet double-bus architecture, CAN bus is used for transmitting real-time control signals (sound sequence parameters, mechanical operation instructions), transmission delay ≤10ms, Ethernet is used for transmitting large-capacity music data (music score file, timbre sample, style model), transmission rate ≥1Gbps;The mechanical music creation table is also provided with data processing and storage module, including edge computing unit, encrypted storage unit and historical creation track backtracking unit;The edge computing unit adopts FPGA chip, is used for accelerating the inference operation of AI composition model, reduces cloud dependence, and the response time of locally generated music score is ≤3s;The encrypted storage unit capacity is ≥1TB, supports AES-256 encryption, stores the music score file, custom style model and operation parameter configuration created by the user;The historical creation track backtracking unit can record the whole process operation data of the user from the initial instruction input to the final work generation, including each version of music score generated by AI, parameter adjustment record of entity sound sequence, execution log of mechanical operation, supports time node backtracking view, and can automatically generate creation suggestion based on historical track.

7. The mechanical music creation station of claim 1, wherein: The mechanical support platform adopts a modular assembly design, including an operation table panel, a functional module support, a shock-absorbing base, and foldable side arms; the operation table panel is made of carbon fiber material, has an anti-slip and wear-resistant coating on the surface, and has installation slots for mechanical knob arrays, a touch display screen, and a mechanical operation execution module, the slot sizes being adapted to different specifications of components through adjusting shims; the functional module support is made of an aluminum alloy frame, has heat dissipation and ventilation holes, and has a cable storage channel inside to avoid cable entanglement; the shock-absorbing base has rubber shock-absorbing pads and spring shock-absorbing mechanisms built-in, can reduce vibration conduction when the mechanical operation execution module is working, and has a vibration attenuation rate of ≥80%; the foldable side arms are arranged on both sides of the operation table panel, can be folded upwards to be parallel to the panel (to save storage space) or unfolded downwards to be horizontal (to increase the area of the operation table), and have USB interfaces, audio output interfaces, and external instrument connection interfaces on the surfaces.

8. The mechanical music creation station of claim 1, wherein: The AI composition assistance module further includes a custom style training interface and a collaborative creation unit; the custom style training interface supports users to import personal creation music score files (formats including MIDI, MusicXML, and WAV), trains a dedicated style model through a transfer learning algorithm, sets the number of iterations (100-1000 adjustable) and the learning rate (0.001-0.01 adjustable) during the training process, saves the trained model to a style model library for subsequent creation, and calls the model; the collaborative creation unit supports multiple creation tables to be connected through a local area network or a cloud server, realizes real-time collaboration of multiple people, one of the creation tables serving as a host computer responsible for unified style model calling and audio sequence synchronization, and the other creation tables serving as slave computers that can respectively edit different audio tracks, the edited content being synchronized to the host computer and the other slave computers in real time, the host computer being capable of integrating and optimizing all audio tracks to generate a final work; The collaborative creation unit also supports permission setting, and the host computer can assign editing permissions, viewing permissions, or read-only permissions to different slave computer users.

9. The mechanical music creation station of claim 1, wherein: The entity audio sequence feedback module further includes an audio feedback unit and an error prompt unit; the audio feedback unit has a full-frequency loudspeaker (frequency response range 50Hz-20kHz) and a headphone interface built-in, can play the sound effect of the mechanical operation execution module and the audio of the AI-generated music score in real time, supports stereo output, and synchronously adjusts the loudspeaker volume with the volume parameter of the entity audio sequence; the error prompt unit includes a sound and light alarm component, flashes red light of an LED audio sequence indicator light group when the AI composition assistance module cannot recognize user instructions, data transmission abnormity occurs, or mechanical components fail, simultaneously issues a prompt sound (frequency 1kHz, duration 0.5s) through the loudspeaker, and displays the fault type and troubleshooting suggestions on the touch display screen; the entity audio sequence feedback module also supports an audio sequence export function, can export the edited audio sequence data as a MIDI, MP3, or WAV format file, and transmits the file to external equipment through a USB interface or a wireless communication module.

10. The mechanical music creation station of claim 1, wherein: Also includes power supply and heat dissipation optimization module; the power supply module supports double power supply mode, including mains power supply (AC 100-240V) and lithium battery power supply (capacity ≥20000mAh, endurance time ≥8 hours), set power management unit, can automatically switch power supply mode, lithium battery supports fast charging (30 minutes charging to 60%) and overcharge, overdischarge protection; the heat dissipation optimization module includes the heat dissipation fan (rotating speed 1000-3000rpm adjustable) in the functional module installation cavity, the heat dissipation channel of mechanical operation execution module and the heat conduction silica gel pad of AI module;The heat dissipation fan is linked with temperature sensor, and when the temperature in the cavity is ≥45℃, it is automatically started, and when the temperature is ≤30℃, it is automatically closed;The heat dissipation channel adopts labyrinth type design, which can guarantee the heat dissipation efficiency and prevent dust from entering;The heat conduction silica gel pad is attached to the core chip of AI composition auxiliary module, and the heat conductivity coefficient is ≥2.0W / (m・K), which ensures the temperature stability during AI model operation;The overall size of the mechanical music creation platform is ≤120cm×60cm×80cm, and the weight is ≤30kg, which is suitable for various scenes such as home studio, mobile creation, etc.