Intelligent music creation auxiliary method based on music theory

Abstract

The invention discloses an intelligent music creation auxiliary method based on a music theory, and belongs to the technical field of music creation. According to the method, intelligent assistance of a music creation process is realized by fusing multi-modal intention input and a structured music theory rule base, a pre-constructed structured music theory rule base is utilized to dynamically map a user intention into a computable constraint parameter, and on the basis, a problem modeling technology is satisfied through constraint, so that the problem modeling efficiency is improved. A music creation problem is converted into a feasible solution space search problem in mathematics, a music fragment conforming to the intention of a user is efficiently searched in a solution space by means of a constraint propagation and backtracking algorithm, a result is output in a multi-mode form, and meanwhile interpretable feedback of the creation process is provided. By means of the method, music theoretical knowledge and user personalized requirements can be effectively integrated, the accuracy and usability of the generated result are improved, meanwhile, dependence on professional literacy of the user is reduced, and efficient and intelligent music creation assistance is achieved.

Description

Technical Field

[0001] This invention relates to the field of music composition technology, and more specifically, to an intelligent music composition assistance method based on music theory. Background Technology

[0002] With the development of artificial intelligence technology, data-driven music generation models, such as deep learning models (including RNN, LSTM, Transformer, Diffusion Model, etc.), have become a research hotspot. These models learn the statistical patterns of music by being trained on massive amounts of music data, thereby generating new music segments.

[0003] However, such methods have inherent technical limitations, including the uncontrollability and randomness of the generation process. Users cannot accurately specify the high-level attributes of the generated music, and the generation results heavily depend on the initial random seed, making it difficult to achieve targeted creation. At the same time, the correctness of music theory cannot be guaranteed, and the data-driven model lacks explicit music knowledge. Its generation results may contain phenomena that violate basic music theory, such as unresolved dissonant intervals, disordered chord progressions, and parallel fifths and octaves. These errors require users to have a high level of musical literacy to identify and correct, which reduces the efficiency of the auxiliary method. Summary of the Invention

[0004] The purpose of this invention is to provide an antibacterial water-based ink and its preparation method to solve the problems of uncontrollability and randomness in the generation process, the inability of users to accurately specify the high-level attributes of the generated music, the heavy reliance on the initial random seed, the difficulty in achieving targeted creation, the inability to guarantee the correctness of music theory, the lack of explicit music knowledge in the data-driven model, and the potential inclusion of phenomena that violate basic music theory, such as unresolved dissonant intervals, chord progression dysfunction, and parallel fifths and octaves. These errors require users to have a high level of musical literacy to identify and correct, which reduces the efficiency of the auxiliary process.

[0005] The steps of an intelligent music creation assistance method based on music theory include: S1. Receive user input describing their musical creative intent and specific musical constraints; S2. Based on the description of the creative intent, call and activate the corresponding music theory rule set from the pre-built music theory rule base; S3. The description of creative intent, specific musical constraints, and the set of activated music theory rules are jointly modeled into a multi-dimensional constraint satisfaction problem for the target music segment, which includes both hard and soft constraints; S4. The constraint satisfaction problem is solved using a hybrid constraint solving engine. First, a feasible solution that satisfies all hard constraints is generated. Then, the feasible solution is optimized based on the soft constraints to generate the final music data. S5. Output the generated music data.

[0006] This method effectively integrates music theory knowledge with users' personalized needs, improves the accuracy and usability of the generated results, reduces reliance on users' professional skills, and achieves efficient and intelligent music creation assistance.

[0007] Preferably, the music theory rule base is modular and configurable, and includes at least a harmony rule module, a melody rule module, a rhythm rule module, and a counterpoint rule module. The rule base supports dynamically loading rule subsets with different parameter configurations according to different music style tags.

[0008] Preferably, in the modeling step S3, the target music segment is formally represented as a network structure consisting of a time dimension. Each cell in the grid represents a music event variable to be solved. The variable includes pitch, duration, and dynamics attributes. The constraints act on these variables and their interrelationships.

[0009] Preferably, the hybrid constraint solving engine adopts a two-stage strategy of first searching and then optimizing. In the first stage, an algorithm based on backtracking search and constraint propagation is used to ensure that hard constraints are satisfied. In the second stage, a metaheuristic optimization algorithm is used to optimize the feasible solution obtained in the first stage with the weighted satisfaction of the soft constraints as the optimization objective.

[0010] Preferably, it also includes the following steps: S6. In response to the user's partial modification operation on the generated music data, the modification operation is regarded as a new hard constraint condition, and the hybrid constraint solving engine is triggered to re-solve the music data of the affected part in order to maintain the coherence of the music context.

[0011] This step allows users to make flexible adjustments during the creation process. When a generated music segment is found to have parts that do not meet expectations, there is no need to regenerate the entire music data. Only the specific part needs to be modified, and the system will automatically re-optimize the related music elements based on the modifications, ensuring the consistency and logic of the overall music style.

[0012] Preferably, it also includes the following steps: S7. Generate and display interpretability information corresponding to the generated music data, including but not limited to: visually highlighting the applied music theory rules and generating a structured music analysis report.

[0013] This step provides users with transparency in the creative process, enabling them not only to obtain the final musical work but also to understand the musical theory rules underlying the creation process. This helps users improve their knowledge of music composition, and is especially valuable for music learners and educators. At the same time, the structured music analysis report can show in detail the characteristics of musical fragments in terms of harmony, melody, rhythm, etc., making it easier for users to conduct in-depth analysis and research.

[0014] Preferably, the system includes a user interaction module for receiving user input, a constraint modeling and compilation module for constructing the multi-dimensional constraint satisfaction problem, a hybrid constraint solving engine module for executing the two-stage solution strategy, a music rendering and output module for outputting final music data, and an interpretability feedback generation module for generating and displaying interpretability information.

[0015] Preferably, the music theory rule base management module provides a graphical interface that allows users or experts to browse, edit, expand, and configure style templates for the rule base.

[0016] Compared with the prior art, the advantages of this invention are: 1. This method can effectively integrate music theory knowledge with users' personalized needs, improve the accuracy and usability of the generated results, reduce reliance on users' professional skills, and achieve efficient and intelligent music creation assistance.

[0017] 2. This step allows users to make flexible adjustments during the creation process. When a generated music segment is found to have a part that does not meet expectations, there is no need to regenerate the entire music data. Only the specific part needs to be modified, and the system will automatically re-optimize the related music elements based on the modifications to ensure the consistency and logic of the overall music style.

[0018] 3. This step provides users with transparency in the creative process, enabling them not only to obtain the final musical work but also to understand the musical theory rules underlying the creation process. This helps users improve their knowledge of music creation, and is especially valuable for music learners and educators. At the same time, the structured music analysis report can show in detail the characteristics of musical fragments in terms of harmony, melody, rhythm, etc., making it convenient for users to conduct in-depth analysis and research. Detailed Implementation

[0019] A smart music creation assistance method based on music theory rules, the specific steps of which are as follows: Step 1: Multimodal Intent and Constraint Input Multimodal intent and constraint input are used to receive high-level musical intents described by the user through a graphical interface or natural language. This intent is a structured data object. Simultaneously, it can receive specific musical constraints input by the user, providing finer-grained control. These constraints include: Basic metadata: key (e.g., C major, A minor), time signature (e.g., 4 / 4, 3 / 4, 6 / 8), tempo range; Style and emotional description: Defined by tags (such as "Baroque", "Bossa Nova", "Melancholy", "Exhilarating") or by selecting a preset style template (such as "Pop Piano Ballad"); Macro structure instructions: Define the musical form, such as "intro-verse-chorus-interlude-outro", and specify the starting measure and number of repetitions for each part; Instrument and voice configuration: Specify the instruments involved (such as piano, bass, drums) and define the role of each voice (such as melody and accompaniment, bass line). Harmonic framework constraints: Directly input chord sequences (such as C - G7 / B - Am - F), or specify start and stop chords, and prohibited chords; Melody-guided constraints: Specify the starting / ending note of the melody, the melody outline ("ascending", "descending", "arched"), the highest / lowest note limit, and the avoidance of specific intervals; Rhythm pattern constraints: Select a rhythm pattern (such as "dotted rhythm" or "Shuffle"), or specify the rhythm density of a specific voice.

[0020] Step 2: Construction and Dynamic Mapping of a Structured Music Theory Rule Base The construction and dynamic mapping of a structured music theory rule base are used to build a hierarchical, modular music theory rule base. This rule base is stored using a declarative language (such as a domain-specific language DSL) or a database table structure. The core modules include: The Harmony Rules module covers chord progression rules (such as avoiding parallel fifths and correcting dissonant intervals), chord function classification (dominant, subdominant, and key modulation rules), and supports loading specific harmonic vocabulary through style tags (such as "classical" and "jazz"). Melody Rules Module: Includes pitch movement rules (such as stepwise as the main method and leap resolution), melody outline templates ("wave-shaped" and "step-shaped"), and ornament usage guidelines. It can configure exclusive melody generation logic for music from different cultural backgrounds (such as Chinese pentatonic scale and Indian ragga). The rhythm rules module defines beat subdivision rules (such as triplet alignment and syncopation), rhythm density gradient (transition patterns from dense to sparse), and multi-voice rhythmic counterpoint relationships. It also supports quick calling of specific style rhythm patterns through the sample rhythm pattern library. The counterpoint rules module includes rules for voice spacing (such as the need to prepare and resolve intervals of a fourth or greater), voice independence detection (avoiding too many common tones), and a template for generating imitation polyphony, which can generate complex counterpoint structures such as canons and fugues. Stylization rules module: Configure the parameters and weights of the above rules module for different music styles (such as "jazz", "classical", "pop"). For example, jazz allows more complex chord extensions (9, 11, 13) and dissonant intervals, and weakens the restriction of parallel fifths.

[0021] The dynamic mapping mechanism automatically activates the corresponding subset of parameters in the rule base by parsing the style tags and constraints input by the user (such as activating the blues scale rule and swing rhythm template under "jazz style"). At the same time, it allows users to manually adjust the rule weights through a graphical interface, achieving a balance between theoretical rigor and creative freedom.

[0022] Step 3: Modeling the constraint satisfaction problem of music fragments Constraint satisfaction problem modeling of musical segments is used to formally model the musical segment to be generated (e.g., a musical phrase or an entire work) as a constraint satisfaction problem on a spatiotemporal grid.

[0023] Variable definition: Each cell (Cell(i, j)) in the grid represents the note attribute to be determined for the j-th voice at the i-th time point (divided into the smallest time unit, such as a sixteenth note). Each variable is a composite data structure containing three sub-variables: pitch, duration, and dynamics. Constraint definition: Constraints act on these variables and fall into two main categories: Hard constraints: These are mandatory rules derived from music theory that must be 100% satisfied. Examples include: "The pitch of all notes must be within the scale of the current key," "Parallel octaves are prohibited," and "Suspensions must be resolved stepwise." Soft constraints, derived from musical aesthetics, style preferences, and user intent, are optimization goals that are expected to be maximized. Examples include: "Melody smoothness (minimum sum of squares of adjacent pitch differences)," "Moderate rhythmic complexity," and "Highest degree of match between harmonic rhythm and user-specified pattern." Each soft constraint is assigned a configurable weight, indicating its importance.

[0024] Step 4: Hybrid Constraint Solving and Music Generation Engine Hybrid constraint solving and music generation engine are used to design a two-stage hybrid solver engine for music generation: Phase 1: Searching the feasible solution space (based on constraint propagation and backtracking) The solver first processes all hard constraints. It starts from the beginning of the music and uses a backtracking search algorithm to try possible values ​​for each variable (note). During the search process, constraint propagation techniques such as inconsistency or forward checking are used to eliminate values ​​that would lead to no solution for subsequent variables in advance, which greatly compresses the search space and improves the solution efficiency. The output of this stage is one or more "theoretically correct" musical fragment drafts that satisfy all hard constraints.

[0025] Phase 2: The feasible solutions obtained in Phase 1 are used as the initial population and input into an optimization algorithm (such as genetic algorithm or simulated annealing algorithm); The fitness function of this optimization algorithm is defined as the sum of the weighted satisfaction of all soft constraints. The algorithm iteratively evolves the population through operations such as selection, crossover, and mutation to find the individual with the highest fitness. The final output is the optimal music segment that best meets the user's style preferences and aesthetic requirements while ensuring the correctness of music theory.

[0026] Real-time interaction and local re-solution mechanism: When a user manually modifies any note generated by the system, the system treats that modification as a new hard constraint. The system will lock the content before the user modified the area, and take the modification point as the new starting point. Based on the current music context, it will re-execute the modeling and solving process of S3 and S4 to locally regenerate the subsequent music content, thereby ensuring the overall coherence and consistency of the music.

[0027] Step 5: Multimodal Results Output and Interpretability Feedback Standard format output: The final generated music data is output in standard formats such as MIDI, MusicXML, or ABC Notation, which facilitates further editing and production in various music software; Structured analysis report generation: Automatically generates an analysis report that details the chord progressions, tonal changes, and formal structure used in the generated segment; Explainable Visual Interface: On the score display interface, areas where specific rules are applied are highlighted with different colors (e.g., "II-VI" is marked with a red box, and "suspension resolution" is indicated with an arrow). At the same time, a "rule traceability" function is provided. When the user clicks on any note, the system will pop up a prompt box to explain the main rules and constraints on which the note was generated.

[0028] An intelligent music creation assistance system that implements the above method includes the following core modules in its architecture: User interaction interface module: Provides graphical UI, natural language processing interface or API to capture user creative intent; Music Theory Knowledge Base Management Module: Responsible for the storage, retrieval, updating and version control of the rule base, and provides a rule editing interface for experts to expand their knowledge; Constraint Modeling and Compiler Module: Transforms user input and rules into formal CSP models. This module serves as a bridge connecting high-level intents and low-level solvers. Hybrid constraint solving engine module: The computational core of the system, integrating a backtracking searcher and an optimizer; Music Data Structure and Context Management Module: Maintains complete music information for the current creation project in memory, providing data sharing and context services for various modules; Music rendering and output module: responsible for serializing the solution results into standard music files; Visualization and Interpretation Engine Module: Generates visual scores and interpretable feedback information.

[0029] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and modifications can be made to the present invention without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for assisting intelligent music creation based on music theory rules, comprising the following steps: S1. Receive the user's description of the music creation intention and specific music constraints; S2. Based on the description of the creative intent, call and activate the corresponding music theory rule set from the pre-built music theory rule base; S3. The description of creative intent, specific musical constraints, and the set of activated music theory rules are jointly modeled into a multi-dimensional constraint satisfaction problem for the target music segment, which includes both hard and soft constraints; S4. The constraint satisfaction problem is solved using a hybrid constraint solving engine. First, a feasible solution that satisfies all hard constraints is generated. Then, the feasible solution is optimized based on the soft constraints to generate the final music data. S5. Output the generated music data.

2. The method according to claim 1, characterized in that: The music theory rule base is modular and configurable, and includes at least a harmony rule module, a melody rule module, a rhythm rule module, and a counterpoint rule module. The rule base supports dynamically loading rule subsets with different parameter configurations based on different music style tags.

3. The method according to claim 2, characterized in that: In the modeling step S3, the target music segment is formally represented as a network structure consisting of a time dimension. Each cell in the grid represents a music event variable to be solved. The variable includes pitch, duration, and dynamics attributes. The constraints apply to these variables and their interrelationships.

4. The method according to claim 1, characterized in that: The hybrid constraint solving engine adopts a two-stage strategy of first searching and then optimizing. In the first stage, an algorithm based on backtracking search and constraint propagation is used to ensure that hard constraints are satisfied. In the second stage, a metaheuristic optimization algorithm is used to optimize the feasible solution obtained in the first stage with the weighted satisfaction of the soft constraints as the optimization objective.

5. The method according to claim 1, characterized in that, It also includes the following steps: S6. In response to the user's partial modification operation on the generated music data, the modification operation is regarded as a new hard constraint condition, and the hybrid constraint solving engine is triggered to re-solve the music data of the affected part in order to maintain the coherence of the music context.

6. The method according to claim 1, characterized in that, It also includes the following steps: S7. Generate and display interpretability information corresponding to the generated music data, including but not limited to: visually highlighting the applied music theory rules and generating a structured music analysis report.

7. An intelligent music creation assistance system that implements the method of any one of claims 1-6, characterized in that: The system includes a user interaction module for receiving user input, a constraint modeling and compilation module for constructing the multi-dimensional constraint satisfaction problem, a hybrid constraint solving engine module for executing the two-stage solution strategy, a music rendering and output module for outputting the final music data, and an interpretability feedback generation module for generating and displaying interpretable information.

8. The system according to claim 7, characterized in that: The music theory rule base management module provides a graphical interface that allows users or experts to browse, edit, expand, and configure style templates for the rule base.