Acoustic modeling and 3D printing technology-based ancient musical instrument digital duplicating method
By combining multi-dimensional data collaborative acquisition and acoustic modeling with 3D printing technology, a high-precision digital replica of the shape, sound, and playing of ancient musical instruments has been achieved, solving the problem of the separation between morphological and acoustic data and improving the accuracy of the replica and the effect of cultural relic protection.
Patent Information
- Application Number
- CN202511707571.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-20
- Publication Date
- 2026-02-17
AI Technical Summary
Existing methods for replicating ancient musical instruments often separate morphological and acoustic data, making it difficult to achieve high-fidelity and dynamic reproduction of performance characteristics, resulting in poor digital replication effects.
By employing multi-dimensional data collaborative acquisition technology, combined with acoustic modeling and 3D printing, a parametric physical acoustic model is constructed by synchronously acquiring the geometric structure and acoustic characteristics data of musical instruments. A physical copy is then prepared through additive manufacturing, and sound-shape fusion verification and iterative calibration are performed to ensure data correlation and model accuracy.
It has achieved high-precision digital reproduction of the shape, sound, and playing of ancient musical instruments, reducing damage to cultural relics, improving the accuracy and consistency of the replication, and expanding the ways of preserving and utilizing cultural heritage.
Smart Images
Figure CN121543416A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital cultural heritage protection technology, specifically a method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology. Background Technology
[0002] Existing methods for replicating ancient musical instruments primarily rely on traditional handcrafting or digital techniques based on a single data dimension. While traditional handcrafting can recreate the instrument's form to some extent, it heavily depends on the craftsman's experience, making precise replication difficult and prone to causing irreversible damage to the artifact. Digital methods based on 3D scanning or photo modeling can acquire data on the instrument's appearance, but lack systematic collection and analysis of the instrument's acoustic characteristics and performance.
[0003] In existing technologies, acoustic data acquisition is often separated from morphological data, resulting in digital models failing to accurately reflect the intrinsic relationship between the sound and form of musical instruments. Furthermore, current acoustic modeling methods mostly employ sampling playback or simple physical simulations, making it difficult to achieve high-fidelity reproduction of dynamic performance responses and tonal details. These issues limit the practical application of digital replication in academic research, music production, and virtual presentations.
[0004] In response to the problems of fragmented morphological and acoustic data, insufficient modeling fidelity, and difficulty in dynamically restoring performance characteristics in existing technologies, this invention proposes a digital replication method and system for ancient musical instruments, aiming to achieve high-precision digital reproduction of the form, sound, and performance of ancient musical instruments. Summary of the Invention
[0005] To address the problems existing in the background technology of ancient musical instrument replication, such as the separation of morphological and acoustic data, insufficient fidelity of digital modeling, and difficulty in dynamically reproducing the real performance characteristics, this paper provides a method for digitally replicating ancient musical instruments that can achieve high-precision digital reproduction of the instrument's shape, acoustic characteristics, and performance.
[0006] To achieve the above objectives, the present invention provides a method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology, comprising the following steps: S1. Multidimensional Data Collaborative Acquisition Steps When replicating the ancient musical instrument, geometric structure data acquisition and acoustic characteristic data acquisition are performed simultaneously. The acoustic characteristic data acquisition is carried out under multiple preset excitation signals and performance states. This step is the data acquisition stage of the entire replication process. By capturing the physical form and sound characteristics of the instrument under the same spatiotemporal reference, it ensures that the subsequently constructed digital model can be built on accurate data that shows the intrinsic relationship between form and sound. This fundamentally solves the problem of data correlation loss caused by the separation of morphological scanning and acoustic measurement in traditional methods.
[0007] Synchronous acquisition refers to the high temporal alignment of the acquisition of two types of data to ensure that each acoustic measurement corresponds to a specific geometric state of the instrument. This can be achieved at the microsecond level through precise hardware synchronization triggers, or through post-event timestamp alignment algorithms for quasi-synchronization. Geometric data encompasses all information describing the physical form of the instrument, including but not limited to external 3D point clouds, surface texture images, and volumetric data of internal structures (such as those obtained through CT scans). Acoustic characteristic data, on the other hand, is the instrument's acoustic response acquired under controlled conditions. Its acquisition requires multiple preset excitation signals and performance states, aiming to systematically stimulate the instrument's acoustic behavior under different operating conditions. For example, standardized pulse and sweep signals are used to measure frequency response and attenuation characteristics, and simulated plucking and striking of strings in real performances are used to capture dynamic tonal changes.
[0008] S2. Acoustic Modeling and Digital Timbre Synthesis Steps Based on the acoustic characteristic data, a parameterized physical acoustic model is constructed. The physical acoustic model uses the geometric structure data as initial constraints, and its model parameters are solved by fitting the measured acoustic data through an optimization algorithm. This step involves creating a computational model that not only reproduces the measured sound but also predicts the acoustic output of an instrument under arbitrary playing inputs. Its purpose goes beyond simple sample playback; it is central to achieving high-fidelity, interactive digital replication.
[0009] A parametric physical acoustic model refers to a mathematical model based on fundamental physical laws of acoustics and vibration (such as wave equations and structural vibration equations), but whose specific performance is controlled by a set of parameters (such as material density, elastic modulus, and damping coefficient). Using geometric data as initial constraints means that the model's geometric topology and boundary settings are directly determined by the high-precision 3D model acquired in the first step. For example, the shape of the resonating cavity directly determines the waveguide network layout, and the thickness distribution of the panel serves as input to the finite element model. Solving for model parameters by fitting measured acoustic data using optimization algorithms means that the model parameters are determined through an automated inverse problem-solving process, ensuring that the model's output is as close as possible to the actually measured sound.
[0010] This approach deeply integrates computational physics and optimization theory. First, a computable forward model is established using physical equations. Then, leveraging the abundant acoustic data collected in the first step, inverse problem-solving methods (such as gradient descent and genetic algorithms) are employed to calibrate the model, ensuring it moves from physical correctness to data accuracy. This allows the final model to maintain both the interpretability and extrapolation capabilities of the physical model and possess data-driven precision.
[0011] S3. Solid Manufacturing Steps Using the aforementioned geometric data, a physical replica of the ancient musical instrument is prepared using additive manufacturing technology. The processed 3D digital model, based on geometric data, serves directly as the manufacturing blueprint. Additive manufacturing technology, with its layer-by-layer material accumulation properties, is used to create the complex, hollow resonating structure of ancient musical instruments. This technology includes, but is not limited to, stereolithography (SLA), selective laser sintering (SLS), and multi-jet melting (MJF). The fabrication process also includes post-processing, such as cleaning, curing, polishing, and installing accessories like strings and frets, to create a complete physical replica.
[0012] S4. Sound and Image Fusion Verification and Iterative Calibration Steps Acoustic characteristic data are collected from the physical entity replica to obtain its acoustic response; the acoustic response of the replica is compared and analyzed with the acoustic response of the original instrument body and the output of the physical acoustic model, and calibration parameters are generated based on the comparison results. The calibration parameters are fed back to the acoustic model and manufacturing process parameters for iterative optimization. This step forms a closed-loop feedback system, which is crucial for achieving high-precision replication. Its purpose is to quantitatively evaluate the differences between the replica and the original, and automatically diagnose the sources of these differences, thereby driving improvements in the model or manufacturing process until the preset accuracy requirements are met.
[0013] Comparative analysis is based on objective quantitative comparison using signal processing techniques, covering differences in the time domain (e.g., decay time), frequency domain (e.g., spectrum, formants), time-frequency domain (e.g., spectrograms), and perceptual dimensions (e.g., loudness, sharpness). Calibration parameters are generated as quantitative adjustment suggestions based on the difference analysis results. For example, if the high-frequency decay of the replica is too rapid, instructions to increase the acoustic model damping parameters are generated; if a systematic shrinkage in overall size is found, instructions to adjust the printing scaling factor are generated. Iterative optimization means that this process may be repeated multiple times, with each iteration bringing the replica's characteristics closer to the original.
[0014] Preferably, in the multidimensional data collaborative acquisition step, the geometric structure data acquisition and acoustic characteristic data acquisition are time-synchronized through a synchronization trigger, and the output is a set of data pairs calibrated with spatiotemporal and excitation contexts.
[0015] By introducing dedicated synchronization trigger hardware, a unified precision clock is ensured that all acquisition devices share this clock. This allows each acoustic recording to be precisely correlated with a specific geometric state at a given moment. The resulting data sets contain not only the data itself but also rich metadata (such as timestamps, excitation types, and application locations). This significantly enhances data traceability and usability, laying a solid foundation for subsequent high-precision correlation modeling.
[0016] Preferably, the acoustic characteristic data acquisition uses a mechanical vibrator or a specially made soft hammer to apply a preset excitation signal at different positions on the string. The excitation signal includes one or more of the following: a broadband pulse signal, a linear sweep signal, and string-plucking analog signals with different forces and rates.
[0017] Using a mechanical exciter enables highly controllable and repeatable precise excitation, making it particularly suitable for scenarios requiring a high signal-to-noise ratio, such as measuring frequency response functions. A specially designed soft hammer simulates excitation closer to that of human hand playing. Wideband pulse signals are used to excite all modes of the instrument, quickly acquiring the overall frequency response; linear sweep signals provide high signal-to-noise ratio frequency response data; and simulated plucking signals at different forces and rates are used to capture the nonlinear response characteristics of the instrument under varying playing intensities. This diversified excitation strategy ensures that the acquired acoustic data comprehensively characterizes the dynamic behavior of the instrument.
[0018] Preferably, in the acoustic modeling and digital timbre synthesis steps, the physical acoustic model adopts a three-dimensional waveguide mesh model or a hybrid model of finite element method and digital waveguide.
[0019] Three-dimensional waveguide mesh models are particularly suitable for simulating air vibrations and the propagation of sound waves in complex cavities (such as piano cases), offering high computational efficiency and ease of real-time synthesis. Hybrid models combining the finite element method (FEM) and digital waveguides offer a more comprehensive solution. The FEM accurately simulates the vibrations of solid structures like the piano body, while the digital waveguide efficiently simulates string vibrations and sound radiation; the two interact through coupling conditions. This hybrid model can more precisely characterize the coupled vibrations of solids and air, thus maintaining both physical accuracy and computational efficiency.
[0020] Preferably, the optimization algorithm is based on gradient descent, and its objective function is to minimize the error between the model output and the measured acoustic response, expressed as: ; in, Represents the model parameter vector. The parameter is The output of the physical model under the i-th excitation signal, This represents the measured i-th acoustic response. represents the weighting coefficient, and N represents the number of activations.
[0021] Gradient descent-based algorithms efficiently find the parameter directions that minimize the objective function. The objective function is quantified as the weighted sum of squares of the differences between the model output and the measured data. The introduction of weighting coefficients allows for assigning higher weights to data points with high signal-to-noise ratios or greater importance, thus guiding the optimization process to focus more on key characteristics. This mathematically rigorous optimization framework ensures the objectivity and effectiveness of the model calibration process.
[0022] Preferably, in the acoustic-visual fusion verification and iterative calibration steps, the comparative analysis includes quantitatively calculating the acoustic response differences from the time domain, frequency domain, and perception dimensions.
[0023] Time-domain analysis can compare waveform envelopes, decay times, etc.; frequency-domain analysis can compare spectral shapes, formant frequencies and amplitudes, harmonic structures, etc.; the perceptual dimension introduces psychoacoustic models (such as loudness, pitch, sharpness, and roughness) to quantify differences in human hearing. This multi-dimensional, multi-scale method of difference quantification is far more comprehensive and reliable than comparing single indicators, and can more accurately pinpoint the problem, thereby generating more effective calibration instructions.
[0024] Preferably, in the sound-image fusion verification and iterative calibration steps, iterative optimization is performed until the difference between the acoustic response of the replica and the original instrument is lower than a preset acceptable threshold.
[0025] The preset acceptable thresholds can be set according to actual application requirements. For example, it can be stipulated that the error of all resonant frequencies is less than 1%, the spectral difference is less than 3dB in the main frequency band, or the difference in perceived sharpness is less than 0.1acum. This transforms the entire replication process from an artistic endeavor into a measurable and controllable engineering technique, ensuring that the final result meets the predetermined quality specifications.
[0026] Preferably, a digital replication system for ancient musical instruments includes: a multi-dimensional data acquisition module for synchronously or quasi-synchronously acquiring geometric structure data and acoustic characteristic data of the ancient musical instrument; an acoustic modeling and processing module for constructing a parameterized physical acoustic model based on the acoustic characteristic data; a digital manufacturing control module for controlling additive manufacturing equipment to prepare a physical copy of the ancient musical instrument; and a data fusion and verification calibration module for acquiring acoustic characteristic data of the physical copy, comparing and analyzing its acoustic response with the acoustic response of the original musical instrument and the output of the physical acoustic model, and generating and feeding back calibration parameters; wherein, each module works collaboratively through a system bus and a data management unit.
[0027] The acoustic modeling and processing module receives data from the acquisition module. Its internal algorithm first performs data preprocessing and feature extraction, then initializes the model structure according to the selected model type, and finally calls an optimization algorithm library to iteratively adjust the model parameters until the error between its output and the measured data is minimized. Ultimately, it outputs a calibrated, workable parametric acoustic model.
[0028] The digital manufacturing control module, including additive manufacturing equipment, a printing material management unit, and possibly a post-processing equipment control unit, receives geometric model files from the data processing module, slices them, and generates machine instructions to control the printer to manufacture solid copies layer by layer. It also manages printing parameters (such as layer thickness, temperature, and material) to ensure manufacturing accuracy and may control subsequent automated post-processing steps such as cleaning and curing.
[0029] Preferably, the multidimensional data acquisition module includes a high-precision structured light 3D scanner, a high-resolution digital camera array, an acoustic data acquisition unit composed of multiple condenser microphones and accelerometers, and a synchronization trigger; this module coordinates the operation of all sensors under the unified scheduling of the synchronization trigger. Optical devices capture the spatial information of the instrument, and acoustic devices record the response to excitation. All data is tagged with a unified timestamp and metadata, and after preliminary processing, outputs a standardized data packet with complete context, providing input for subsequent modules.
[0030] Preferably, the data fusion and verification calibration module includes a difference analysis algorithm unit, used to quantify acoustic response differences from the time domain, frequency domain, and perception dimensions.
[0031] The data fusion and verification calibration module drives the testing system to perform acoustic measurements on the manufactured replica, obtaining its response data. Then, its internal algorithm unit calculates the difference between this response and the original target from multiple dimensions. Based on preset rules and models, it diagnoses the root causes of the differences and generates specific, quantified calibration parameters (such as reducing the damping coefficient ξ of the acoustic model by 0.002 or increasing the manufacturing scaling factor by 0.1%). These parameters are fed back to the acoustic modeling module or the manufacturing control module to initiate the next iteration.
[0032] This invention provides a method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology. It has the following beneficial effects: 1. This invention ensures the spatiotemporal correlation between morphological and acoustic data from the source of data acquisition, and establishes their inherent mapping relationship using a parametric physical model. The resulting digital model and physical copy can faithfully reflect the organic unity between the original instrument's shape, acoustic characteristics and performance, thus solving the core problem of data fragmentation in existing technologies.
[0033] 2. This invention automatically calibrates model parameters through data-driven optimization algorithms, evaluates replication quality through quantitative indicators, and guides iterative optimization. It transforms the traditional replication process, which relies on craftsmanship experience and subjective perception, into a measurable, controllable, and repeatable modern engineering process, greatly reducing subjective uncertainty and improving the accuracy and consistency of replication.
[0034] 3. This invention reduces direct intervention and potential damage to precious cultural relics through micro-contact digital acquisition. The resulting high-fidelity digital model can not only be used for accurate physical replication, but also directly applied to a wide range of fields such as virtual reality, digital museums, music composition, and academic research, greatly expanding the ways in which cultural heritage is preserved, researched, and utilized. Attached Figure Description
[0035] Figure 1 This is a block diagram of the digital replication system of the present invention. Detailed Implementation
[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0037] Example: like Figure 1 As shown, this embodiment of the invention provides a digital replica system for ancient musical instruments based on acoustic modeling and 3D printing technology, comprising: a multi-dimensional data acquisition module for synchronously or quasi-synchronously acquiring geometric structure data and acoustic characteristic data of the ancient musical instrument; an acoustic modeling and processing module for constructing a parameterized physical acoustic model based on the acoustic characteristic data; a digital manufacturing control module for controlling additive manufacturing equipment to prepare a physical copy of the ancient musical instrument; and a data fusion and verification calibration module for acquiring acoustic characteristic data of the physical copy, comparing and analyzing its acoustic response with the acoustic response of the original musical instrument and the output of the physical acoustic model, generating and feeding back calibration parameters; wherein, each module works collaboratively through a system bus and a data management unit.
[0038] The steps for multidimensional data collaborative acquisition are as follows: The multi-dimensional data acquisition module is used to simultaneously acquire geometric structure data and acoustic characteristic data of the guqin to be replicated. The acoustic characteristic data acquisition is carried out under multiple preset excitation signals and performance states.
[0039] In one specific implementation, the data acquisition step is completed through a multidimensional data acquisition module. This module includes a high-precision structured light 3D scanner, a high-resolution digital camera array, an acoustic data acquisition unit consisting of multiple condenser microphones and accelerometers, and a synchronization trigger. The synchronization trigger can be a hardware clock synchronization device based on the IEEE 1588 Precision Time Protocol (PTP), capable of providing microsecond-level time synchronization signals to all acquisition devices. A data parsing unit, connected to each acquisition device, performs timestamp alignment, format unification, and metadata association on the received raw data, forming a set of data pairs that define the spatiotemporal and excitation context.
[0040] Taking the digitization of a Ming Dynasty Zhongni-style guqin as an example, after fixing the guqin on a soundproof and vibration-damping platform, the equipment was first calibrated: a standard calibration board was used to calibrate the intrinsic and extrinsic parameters of the 3D scanner and camera array to ensure a measurement accuracy of 0.05mm; an acoustic calibrator was used to calibrate the sensitivity of the condenser microphone to 94dB / 1kHz. During geometric acquisition, the scanner scanned from 12 angles, including the head, tail, and sides of the guqin, generating a dense point cloud of approximately 8 million points, while the camera array simultaneously acquired a texture image with a resolution of 2048×1536. During acoustic acquisition, a specially made soft hammer was used to apply three excitations at the 1 / 4, 1 / 2, and 3 / 4 positions of the 16th string: a 5ms wideband pulse (simulating striking), a 0.5-5kHz linear sweep signal (lasting 2s), and three different levels of string plucking simulation (light, medium, and heavy). All excitations were precisely applied through a mechanical vibrator and simultaneously recorded by a synchronous trigger. The final output format is: {PointCloud_001,(Chirp_Excitation,MicrophoneArray_Response)_002}, where the metadata includes information such as the stimulus type, location coordinates, and timestamp.
[0041] The steps of acoustic modeling and digital timbre synthesis are as follows: The acquired acoustic response was resampled at 48kHz, processed using a window function, and subjected to FFT transformation. The following acoustic features were extracted: Formant parameters: The first 6 formant frequencies Fn (n=1-6), bandwidth Bn, and amplitude An were identified using a peak detection algorithm, with an accuracy of 0.1Hz; Harmonic distortion: The total harmonic distortion of the 1st to 5th harmonics was calculated, requiring less than 2%; Attenuation characteristics: The distribution of the 60dB attenuation time (RT60) in the 250Hz-4kHz frequency band was calculated; Modal parameters: The frequencies and damping ratios of the first-order vibration modes of the instrument body were calculated using accelerometer data.
[0042] A hybrid model combining a 3D waveguide mesh and the finite element method is used as the basic framework. The waveguide mesh simulates air vibration, with the mesh node spacing Δx determined to be 3.4 mm (Δx = c / (2f_max)) based on the highest frequency of 5 kHz and a sampling rate of 48 kHz. The mesh topology is determined by the geometry of the instrument body. The finite element model simulates the solid vibration of the instrument body, using C3D8R eight-node reduced integral elements. The initial material properties are set as Young's modulus E0 = 10 GPa, Poisson's ratio ν = 0.3, and density ρ = 380 kg / m³. 3 The model has a total of parameter vectors to be optimized, P→=[E,ν,ρ,ξ1,ξ2, ...]^T, where ξ is the damping ratio of each mode.
[0043] The optimization uses the Levenberg-Marquardt algorithm, with the objective function being: ; in, Represents the model parameter vector. The parameter is The output of the physical model under the i-th excitation signal, This represents the measured i-th acoustic response. The weighting coefficient is set according to the signal-to-noise ratio. The impulse response is set to 1.2, the sweep response to 0.8, and the string-plucking response to 1.0. N represents the number of excitations.
[0044] The norm uses perceptually weighted spectral distance, and its calculation formula is: ; Where X(k) and Y(k) are spectral coefficients, and ε is a small constant to prevent division by zero, with a value of 10. -6 After 500 iterations of optimization, the algorithm converged, yielding the optimal parameters [E=9.7GPa, ν=0.31, ρ=375kg / m]. 3 ,ξ1=0.005, ξ2=0.007, ...].
[0045] The optimized model was compiled into a VST3 audio plugin, which receives MIDI input and calculates and outputs a 44.1kHz / 24bit audio stream in real time. For example, when receiving MIDI note A4 (440Hz) with a velocity value of 80, the plugin internally solves the wave equation and synthesizes a piano tone containing 7 significant formants (107Hz, 315Hz, 662Hz, 1205Hz, 2350Hz, 3580Hz, 4980Hz), with a spectral correlation coefficient of 0.97 with the measured sample.
[0046] The physical manufacturing steps are as follows: The processed 3D model is imported into the printer control system. Based on the material characteristics of the cultural relic, a brown photosensitive resin (elastic modulus 9.5 GPa, density 380 kg / m³) is selected. 3 The printing material was [material name - missing]. The printing layer thickness was set to 0.05mm, and a support structure was used to ensure the forming accuracy of the instrument's curved surface. After printing, the instrument was cleaned with isopropyl alcohol and UV cured. Then, steel strings (0.8mm-1.2mm in diameter) were installed and adjusted to the standard pitch (A4=440Hz). The final physical replica's geometric dimensions deviated from the original guqin by less than 0.15mm, and its weight deviation was less than 3%.
[0047] The data fusion and verification calibration module includes a difference analysis algorithm unit, used to quantify acoustic response differences from the time domain, frequency domain, and sensory dimensions. The acoustic-visual fusion verification and iterative calibration steps are as follows: The physical entity replica undergoes acoustic testing, and its response is compared with the original data. Based on the comparison results, calibration parameters are generated and iteratively optimized. This is achieved through a data fusion and verification calibration module, whose internal difference analysis algorithm unit performs the following operations: Data Acquisition for Copy A: The initial acquisition process is completely repeated for the 3D-printed copy to obtain the copy's acoustic response dataset under the same excitation settings. For example, a linear sweep signal is applied at the same location, and the response of the microphone array is recorded.
[0048] B. Multidimensional Difference Analysis: Differences were quantified from three dimensions: Time domain: The difference in the decay time of the impulse response ΔRT60 was calculated. At the 250Hz frequency point, the replica's decay time was 3.2s while the original was 3.5s, a difference of 0.3s; Frequency domain: The difference in the 1 / 3 octave band spectrum was calculated. At the 1205Hz resonant peak, the replica's amplitude was 2.8dB lower; Perceptual dimension: The psychoacoustic sharpness was calculated. The replica's value was 1.35acum while the original was 1.28acum.
[0049] C-parameter calibration and feedback: Analysis of the causes of differences: If the spectral differences show a systematic shift and are related to the dimensional shrinkage found in the geometric scan (e.g., the body length shrinkage of 0.12%), then manufacturing compensation parameters (dimensional scaling factor 1.0012) are generated and fed back to the printing system; if the differences are manifested as changes in damping characteristics (e.g., excessively rapid high-frequency attenuation), then the damping parameter ξ of the acoustic model is adjusted, changing ξ3500Hz from 0.012 to 0.009, and the model is re-optimized.
[0050] D-iteration termination condition: Set acceptable thresholds: spectral difference less than 3dB, perceived sharpness difference less than 0.1acum. After 2 iterations, the replica response meets all threshold requirements, and the iteration terminates.
[0051] The method and system for digitally replicating ancient musical instruments provided in this invention achieve high-precision digital reproduction integrating form, sound, and performance through a multi-dimensional data collaborative acquisition mechanism and a sound-form fusion verification and iterative calibration process. By acquiring multi-dimensional data pairs with associated context through a spatiotemporally synchronized acquisition system, the problem of the separation of morphological and acoustic data in the prior art is solved. A precise form-sound mapping relationship is established through parametric physical modeling and optimization algorithms. More importantly, a closed-loop calibration mechanism of manufacturing-measurement-comparison enables quantitative evaluation and iterative optimization of the replication process, transforming the experience-dependent nature of traditional processes into a data-driven, precise engineering process, ultimately achieving a high standard of acoustic fidelity that can be quantitatively verified.
[0052] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology, characterized by: Includes the following steps: S1. Multidimensional Data Collaborative Acquisition When replicating the ancient musical instrument, geometric structure data acquisition and acoustic characteristic data acquisition are performed simultaneously. The acoustic characteristic data acquisition is carried out under multiple preset excitation signals and performance states. S2. Acoustic Modeling and Digital Timbre Synthesis Based on the acoustic characteristic data, a parameterized physical acoustic model is constructed. The physical acoustic model uses the geometric structure data as initial constraints, and its model parameters are solved by fitting the measured acoustic data through an optimization algorithm. S3. Solid Manufacturing Using the aforementioned geometric data, a physical replica of the ancient musical instrument is prepared using additive manufacturing technology. S4. Sound and Image Fusion Verification and Iterative Calibration Acoustic characteristic data are collected from the physical entity replica to obtain its acoustic response; the acoustic response of the replica is compared and analyzed with the acoustic response of the original instrument body and the output of the physical acoustic model, and calibration parameters are generated based on the comparison results. The calibration parameters are fed back to the acoustic model and manufacturing process parameters for iterative optimization.
2. The method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology according to claim 1, characterized in that: In the multidimensional data collaborative acquisition step, the acquisition of geometric structure data and acoustic characteristic data is synchronized in time through a synchronization trigger, and the output is a set of data pairs that are calibrated with spatiotemporal and excitation context.
3. The method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology according to claim 2, characterized in that: The acoustic characteristic data acquisition uses a mechanical vibrator or a specially made soft hammer to apply a preset excitation signal at different positions on the strings. The excitation signal includes one or more of the following: a broadband pulse signal, a linear sweep signal, and string-plucking analog signals with different force and speed.
4. The method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology according to claim 1, characterized in that: In the acoustic modeling and digital timbre synthesis steps, the physical acoustic model adopts a three-dimensional waveguide mesh model or a hybrid model of finite element method and digital waveguide.
5. The method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology according to claim 1, characterized in that: The optimization algorithm employs a gradient descent-based approach, with its objective function being to minimize the error between the model output and the measured acoustic response, expressed as: ； in, Represents the model parameter vector. The parameter is The output of the physical model under the i-th excitation signal, This represents the measured i-th acoustic response. represents the weighting coefficient, and N represents the number of activations.
6. The method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology according to claim 1, characterized in that: In the aforementioned acoustic-visual fusion verification and iterative calibration steps, the comparative analysis includes quantitatively calculating the acoustic response differences from the time domain, frequency domain, and perception dimensions.
7. The method for digitally replicating ancient musical instruments based on acoustic modeling and 3D printing technology according to claim 1, characterized in that: In the sound-image fusion verification and iterative calibration steps, iterative optimization is performed until the difference between the acoustic response of the copy and the original instrument is lower than a preset acceptable threshold.
8. A digital replication system for ancient musical instruments, used to implement the method according to any one of claims 1-7, characterized in that: include: A multi-dimensional data acquisition module is used to synchronously or quasi-synchronously acquire geometric structure data and acoustic characteristic data of ancient musical instruments; an acoustic modeling and processing module is used to construct a parameterized physical acoustic model based on the acoustic characteristic data; A digital manufacturing control module is used to control additive manufacturing equipment to produce a physical copy of the ancient musical instrument; The data fusion and verification calibration module is used to collect acoustic characteristic data of the physical entity copy, compare and analyze its acoustic response with the acoustic response of the original instrument body and the output of the physical acoustic model, and generate and feed back calibration parameters; wherein, each module works in concert through the system bus and data management unit.
9. The ancient musical instrument digital replication system according to claim 8, characterized in that: The multidimensional data acquisition module includes an optical data acquisition unit consisting of a high-precision structured light 3D scanner and a high-resolution digital camera array, an acoustic data acquisition unit consisting of multiple capacitive microphones and accelerometers, and a synchronization trigger.
10. The ancient musical instrument digital replication system according to claim 8, characterized in that: The data fusion and verification calibration module includes a difference analysis algorithm unit, which is used to quantify acoustic response differences from the time domain, frequency domain, and perception dimensions.