Music pushing method and system based on travel and tourism big data

Abstract

The application discloses a music pushing method and system based on big data of tourism and travel, and belongs to the field of digital music and big data application, which comprises the following steps: acquiring an environmental dynamic pressure fluctuation signal and a building material pressure wave reflection characteristic and generating an environmental pressure fingerprint; extracting a rhythmic pressure paradigm to generate a rhythm synchronization parameter; combining an auditory preference feature to generate a sound pressure compensation model containing a frequency spectrum compensation curve; calculating a reverse spatial impulse response and fusing to generate a customized synthesis instruction; acquiring a plasticity audio track material for real-time synthesis, and outputting a customized audio stream. The application adopts a technical means of coupling spatial acoustic characteristics and individual hearing depth, can realize physical adaptation and rhythm synchronization of audio signals and tourism and travel environment, effectively offsets spatial reverberation and compensates environmental noise masking effect, and improves immersive listening experience and interactive satisfaction of tourists in complex tourism and travel scenes.

Description

Technical Field

[0001] This invention relates to the field of digital music and big data applications, and in particular to a music recommendation method and system based on cultural tourism big data. Background Technology

[0002] Fluid pressure measurement technology, especially the precise measurement of broad-spectrum dynamic pressure signals in the environment that include acoustic components—including subsonic pressure pulsations with frequencies below 20Hz and acoustic pressure components within the audible frequency range—is a fundamental technology in environmental sensing. These pressure fluctuation signals are widely present in nature in the form of sound waves and wind pressure changes. Applying this measurement technology to service systems that enhance user experience, such as information push systems in cultural and tourism scenarios, to provide more contextualized services through the quantitative perception of environmental parameters, is a current direction of technological development.

[0003] In existing technologies, some advanced audio guide or music recommendation systems for cultural and tourism scenarios have begun to utilize environmental information. For example, some systems use location information to recommend music related to the culture of the attraction, or use simple acoustic sensors to measure the total sound pressure level of ambient noise, i.e., volume, to perform automatic gain control on the played audio volume, ensuring that users can hear the content clearly in noisy environments. These solutions improve the user's listening experience to some extent.

[0004] However, existing technical solutions have significant limitations. First, their utilization of environmental pressure signals is very rudimentary and simplistic, merely measuring the total sound pressure level to adjust volume, failing to delve into the rich information contained in pressure fluctuation signals, such as rhythm and spectrum. Therefore, they cannot enable the pushed audio content to interact with the natural rhythm of the environment. Second, existing technologies typically directly push pre-made audio files, lacking effective means to handle pressure wave reflections caused by factors such as building structures—i.e., acoustic reverberation—leading to significant differences in listening experience across different acoustic spaces. Finally, existing systems lack a mechanism to integrate environmental pressure measurement, user behavior feedback, and audio signal generation, failing to form a collaborative closed-loop system. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a music recommendation method and system based on cultural and tourism big data. The method involves measuring and analyzing environmental dynamic pressure fluctuations and pressure wave reflection characteristics to generate customized synthesis instructions, which are then applied to malleable audio track materials. This enables the real-time synthesis of customized audio streams that are synchronized and coordinated with the physical environmental pressure characteristics.

[0006] The above objectives can be achieved through the following approach: Music recommendation methods based on cultural and tourism big data include: The system acquires the environmental dynamic pressure fluctuation signal of the tourist's current location and the pressure wave reflection characteristics determined by the building materials of the scenic spot, and integrates the environmental dynamic pressure fluctuation signal and the pressure wave reflection characteristics to generate an environmental pressure fingerprint. The environmental dynamic pressure fluctuation signal is frequency-decoupled, low-frequency pressure features outside the hearing frequency range are extracted, a beat paradigm is constructed, and acoustic spectrum features within the hearing frequency range are extracted to construct a spectrum compensation basis. Extract a periodic pressure paradigm from the environmental pressure fingerprint, and generate beat synchronization parameters based on the periodic pressure paradigm. Based on the beat-like pressure paradigm and the auditory preference features extracted from user historical behavior data, a sound pressure compensation model is generated, which includes a spectral compensation curve for adjusting the spectral energy distribution of the audio material. The inverse spatial impulse response is calculated based on the pressure wave reflection characteristics, and the inverse spatial impulse response is fused with the sound pressure compensation model to generate a customized synthesis instruction; Acquire malleable audio track materials that match the cultural attributes of the scenic spot, and apply the customized synthesis instructions to synthesize the malleable audio track materials in real time, outputting a customized audio stream that is synchronized with the environmental pressure beat.

[0007] Optionally, extracting a periodic pressure paradigm from the environmental pressure fingerprint and generating beat synchronization parameters based on the periodic pressure paradigm includes: Time-frequency analysis was performed on the environmental dynamic pressure fluctuation signal in the environmental pressure fingerprint to identify the dominant frequency component whose amplitude changes periodically with time. The center frequency and energy envelope of the dominant frequency component are extracted and together constitute the rhythmic pressure paradigm. The center frequency of the beat-pressure paradigm is matched with the internal rhythm of the malleable audio track material to generate a time mapping relationship for synchronizing the beats of the two, and the time mapping relationship is encapsulated as beat synchronization parameters.

[0008] Optionally, generating a sound pressure compensation model based on the beat-like pressure paradigm and auditory preference features extracted from user historical behavior data includes: By analyzing long-term listening records from users' historical behavior data, spectral centroid parameters representing preferences for the brightness or warmth of audio are extracted, and auditory preference features are constructed based on these spectral centroid parameters. The auditory preference features are compared with the spectral distribution in the environmental pressure fingerprint to generate an initial compensation curve for adjusting the gain of each frequency band in the malleable audio track material. Based on the energy intensity of the rhythmic pressure paradigm, the gain weights of the corresponding frequency bands in the initial compensation curve are dynamically adjusted to form the final spectral compensation curve, and the spectral compensation curve is encapsulated as a sound pressure compensation model.

[0009] Optionally, calculating the inverse spatial impulse response based on the pressure wave reflection characteristics and fusing the inverse spatial impulse response with the sound pressure compensation model to generate customized synthesis instructions includes: Based on the spatial geometric topology parameters of the current location and combined with the pressure wave reflection characteristics, a linear time-invariant system model describing the real spatial reverberation effect is constructed, and the spatial impulse response function is calculated. Perform a complex reciprocal operation on the spatial impulse response function in the frequency domain to generate an inverse spatial impulse response that is the opposite of the actual spatial reverberation effect. The frequency domain expression of the inverse spatial impulse response is multiplied point by point with the spectral compensation curve in the sound pressure compensation model to generate a customized synthesis instruction that simultaneously has reverberation cancellation and spectral shaping functions.

[0010] Optionally, the method further includes: By combining tourist movement path prediction data, areas of degraded communication link quality ahead of the path can be identified. Before tourists enter the area where the communication link quality degrades, a pre-caching operation is triggered for the beat synchronization parameters and the sound pressure compensation model within a future time window. When a tourist is in an area where the communication link quality degrades, the local clock, the cached beat synchronization parameters, and the sound pressure compensation model are used to perform local audio synthesis on the terminal, enabling uninterrupted playback of music.

[0011] Optionally, the operation of triggering the pre-caching of the beat synchronization parameters and the sound pressure compensation model within a future time window includes: Simultaneously with triggering the pre-buffering operation, a parameterized audio encoding mode is initiated; The parametric audio coding mode decomposes the malleable audio track material into a fundamental frequency component containing the main pitch and energy, and a set of residual parameters describing harmonic and transient characteristics; In the pre-caching operation, only the fundamental frequency component and the residual parameter set are cached, thereby reducing the amount of cached data.

[0012] Optionally, applying the customized synthesis instructions to perform real-time synthesis of the malleable audio track material includes: The gain of several frequency bands of the plastic audio track material is independently adjusted according to the spectral compensation curve in the sound pressure compensation model to generate a set of adjusted frequency band signals. The adjusted frequency band signals are superimposed and inverse Fourier transform is performed to generate a time-domain intermediate audio signal; The intermediate audio signal in the time domain is convolved with the inverse spatial impulse response in real time to generate and output a customized audio stream.

[0013] Optionally, the method further includes: Biomechanical response data of tourists is acquired through the inertial measurement unit of a smart terminal, and the biomechanical response data includes the periodic characteristics of head movements. Calculate the synchronization correlation between the periodic characteristics of the head movements and the beat-like pressure paradigm, and generate a synchronization satisfaction index; Based on the synchronization satisfaction index, the spectral centroid parameter in the auditory preference features is dynamically corrected to optimize the generated sound pressure compensation model.

[0014] Optionally, dynamically correcting the spectral centroid parameter in the auditory preference features based on the synchronization satisfaction index includes: When the synchronization satisfaction index is in the predetermined high satisfaction range, the current spectrum centroid parameter remains unchanged. When the synchronization satisfaction index does not reach the predetermined low satisfaction threshold, an exploratory adjustment is performed on the spectrum centroid parameter, which involves slightly shifting the frequency value within the target frequency threshold range. Using the corrected spectral centroid parameters, the sound pressure compensation model is regenerated and applied to subsequent audio synthesis.

[0015] Based on the same inventive concept, the present invention also provides a music recommendation system based on cultural and tourism big data, the system comprising: An environmental fingerprint generation module is used to acquire the dynamic environmental pressure fluctuation signal of the tourist's current location and the pressure wave reflection characteristics determined by the building materials of the scenic spot, and to fuse the dynamic environmental pressure fluctuation signal and the pressure wave reflection characteristics to generate an environmental pressure fingerprint. The beat parameter extraction module is used to extract a beat pressure paradigm with periodic regularity from the environmental pressure fingerprint, and generate beat synchronization parameters based on the beat pressure paradigm. The sound pressure compensation modeling module is used to generate a sound pressure compensation model based on the beat pressure paradigm and the auditory preference features extracted from the user's historical behavior data. The sound pressure compensation model includes a spectral compensation curve for adjusting the spectral energy distribution of the audio material. The instruction fusion generation module is used to calculate the inverse spatial impulse response based on the pressure wave reflection characteristics, and to fuse the inverse spatial impulse response with the sound pressure compensation model to generate customized synthesis instructions; The real-time synthesis output module is used to acquire malleable audio track materials that match the cultural attributes of the scenic spot, and to apply the customized synthesis instructions to synthesize the malleable audio track materials in real time, outputting a customized audio stream that is synchronized with the environmental pressure beat.

[0016] Compared with the prior art, the present invention has the following advantages: This invention achieves synchronization between the beat of the music and the natural rhythm of the environment by measuring the dynamic pressure fluctuation signal of the environment and extracting the rhythmic pressure pattern. It transforms background noise such as waterfall sounds and ocean waves caused by pressure fluctuations from auditory interference into part of the musical resonance. What users hear is no longer a simple superposition of music and noise, but a holistic soundscape that is deeply integrated with the physical environment in terms of rhythm, enhancing the immersion and uniqueness of the experience in cultural and tourism scenarios.

[0017] This invention combines the measurement and analysis of pressure wave reflection characteristics determined by the building materials of scenic spots. By calculating the inverse spatial impulse response, it pre-cancels reverberation and echoes in the environment. Simultaneously, it integrates auditory preference features extracted from users' historical behavior to generate a sound pressure compensation model. This ensures that the final customized audio stream delivered to the user's ears maintains high clarity while precisely matching the user's individual auditory comfort zone. This approach solves the shortcomings of traditional solutions that result in inconsistent sound quality in different acoustic spaces such as caves and halls, achieving personalized acoustic consistency across various scenarios.

[0018] This invention establishes a dynamic closed-loop feedback mechanism through real-time measurement of tourists' biomechanical response data, particularly the analysis of periodic head movements. It interprets unconscious bodily movements as feedback on satisfaction with the current audio effect and adjusts the parameters in the sound pressure compensation model accordingly in real time, enabling the system to learn and continuously optimize. This real-time adjustment based on pressure and biomechanical feedback allows the music recommendation scheme to continuously approach the optimal auditory experience for each user in a specific environment, achieving true deep personalization and adaptability.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims, and drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the music push method based on cultural and tourism big data according to an embodiment of the present invention.

[0022] Figure 2 This is a diagram illustrating the synchronization analysis of environmental pressure and audio rhythm in an embodiment of the present invention.

[0023] Figure 3 This is a heatmap of the spectral gain distribution of the sound pressure compensation model according to an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the structure of a music push system based on cultural tourism big data according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0026] Reference Figure 1 One embodiment of the present invention proposes a music push method based on cultural and tourism big data. The method uses the measurement and analysis of environmental dynamic pressure fluctuations and pressure wave reflection characteristics to generate customized synthesis instructions and apply them to malleable audio track materials. This method can synthesize customized audio streams that are synchronized and coordinated with the physical environmental pressure characteristics in real time.

[0027] The method described in this embodiment specifically includes: The system acquires the environmental dynamic pressure fluctuation signal of the tourist's current location and the pressure wave reflection characteristics determined by the building materials of the scenic spot, and integrates the environmental dynamic pressure fluctuation signal and the pressure wave reflection characteristics to generate an environmental pressure fingerprint. The environmental dynamic pressure fluctuation signal is frequency-decoupled, low-frequency pressure features outside the hearing frequency range are extracted, a beat paradigm is constructed, and acoustic spectrum features within the hearing frequency range are extracted to construct a spectrum compensation basis. Extract a periodic pressure paradigm from the environmental pressure fingerprint, and generate beat synchronization parameters based on the periodic pressure paradigm. Based on the beat-like pressure paradigm and the auditory preference features extracted from user historical behavior data, a sound pressure compensation model is generated, which includes a spectral compensation curve for adjusting the spectral energy distribution of the audio material. The inverse spatial impulse response is calculated based on the pressure wave reflection characteristics, and the inverse spatial impulse response is fused with the sound pressure compensation model to generate a customized synthesis instruction; Acquire malleable audio track materials that match the cultural attributes of the scenic spot, and apply the customized synthesis instructions to synthesize the malleable audio track materials in real time, outputting a customized audio stream that is synchronized with the environmental pressure beat.

[0028] Specifically, the core of this invention lies in establishing an end-to-end adaptive processing link from multi-dimensional environmental perception to real-time generation of personalized audio signals. First, by integrating architectural acoustics of scenic spots, real-time environmental noise, and passenger flow data, an "acoustic fingerprint vector" characterizing the dynamic properties of the physical space is constructed. Second, environmental noise is innovatively treated as a resource, from which regular environmental beat signals are extracted, and beat synchronization parameters are generated for synchronization, achieving alignment between musical rhythm and natural cadence. Based on this, the system does not simply push finished music, but rather pulls highly malleable resonant audio tracks and combines them with the user's historical listening preferences and real-time acoustic fingerprints to generate a personalized resonance mapping. This mapping can fine-tune the timbre of the music according to user preferences and allow the music to complement the environmental noise in the frequency spectrum. Finally, by calculating a "virtual inverse impulse response" to pre-cancel environmental reverberation, this response is superimposed with the personalized resonance mapping to generate customized decoding parameters. The audio tracks are then synthesized in real-time at the terminal, ultimately outputting an audio stream that is deeply coupled with the specific environment and individual user in terms of rhythm, frequency spectrum, and spatial sense.

[0029] Optionally, extracting a periodic pressure paradigm from the environmental pressure fingerprint and generating beat synchronization parameters based on the periodic pressure paradigm includes: Time-frequency analysis was performed on the environmental dynamic pressure fluctuation signal in the environmental pressure fingerprint to identify the dominant frequency component whose amplitude changes periodically with time. The center frequency and energy envelope of the dominant frequency component are extracted and together constitute the rhythmic pressure paradigm. The center frequency of the beat-pressure paradigm is matched with the internal rhythm of the malleable audio track material to generate a time mapping relationship for synchronizing the beats of the two, and the time mapping relationship is encapsulated as beat synchronization parameters.

[0030] Specifically, the original acquired signal is a full-band, broad-spectrum dynamic pressure signal. First, it undergoes frequency division preprocessing. In the dynamic branch, a low-pass filter is used to extract the signal components from 0.1Hz to 20Hz. This component represents the environmental physical rhythms, such as wind pressure or fluid pressure, and is used to subsequently extract the environmental fundamental beat frequency in the beat-like pressure paradigm. Acoustic compensation branch: A high-pass filter is used to extract signal components above 20Hz. This component represents the ambient background noise and is used to calculate the ambient noise energy in the subsequent sound pressure compensation model. This frequency division and decoupling method ensures the physical accuracy of beat synchronization while guaranteeing the scientific validity of spectrum compensation within the auditory range. Environmental pressure fingerprints, as a comprehensive vector integrating spatial reflection characteristics and dynamic fluctuation signals, contain both time-domain and frequency-domain features of the environmental acoustic field. To accurately extract background energy fluctuations with natural rhythms, a real-time sound pressure sequence with a sampling rate of 44.1 kHz is first acquired using the microphone array of a smart terminal. A Fast Fourier Transform (FFT) is then performed using a Hanning window with a length of 2048 points and an overlap rate of 50%, transforming it to the frequency domain to obtain the environmental pressure spectrum matrix. By performing a sliding window analysis on the energy evolution sequence of each frequency point in this matrix, the dominant frequency component with a stable periodic change in amplitude is identified. The core algorithm for identifying this component is based on the autocorrelation coefficient of the energy peak on the time axis: when the energy sequence at a certain frequency point is delayed... Autocorrelation coefficient at When the peak value exceeds a preset stability threshold of 0.7 and exhibits continuity over at least three delay periods, the frequency point is determined to be an environmental beat component. The center frequency of this dominant frequency component is extracted as the environmental fundamental beat frequency, and a peak detection algorithm is used to extract local maxima points in its energy envelope whose amplitude exceeds twice the standard deviation of the sequence average. These constitute the environmental pulse time point sequence, and together they form a beat-stress paradigm. To achieve alignment between the pushed music and the environment, beat synchronization parameters need to be generated based on malleable audio track materials. The malleable audio track material is a structured audio file containing XML metadata descriptors, whose metadata tags pre-set rhythm parameters conforming to industry standards or specific specifications. The environmental fundamental beat frequency is defined as... It originates from the strongest periodic fluctuation frequency detected by environmental pressure fingerprints; the internal reference rhythm frequency corresponding to the tag in the audio track metadata is defined as... Calculate the frequency scaling factor. : ; This is used to adjust the playback speed of audio tracks using a time-domain companding algorithm to achieve frequency alignment. Simultaneously, the first significant peak moment in the environmental pulse timeline is defined as the environmental reference start moment. ; Define the internal reference start time corresponding to the tag in the audio track metadata as Calculate the time offset. : ; By frequency scaling factor With time offset The common encapsulation constitutes the beat synchronization parameters, which are used for subsequent resampling and phase shifting of the audio signal on the time axis to ensure that the accent phase of the output audio completely coincides with the physical pulse of the environmental pressure. For example... Figure 2 As shown, the distribution relationship between environmental pressure pulses and the original rhythm of the audio track on the time axis is illustrated. By placing the diagram below, the solid lines representing the environmental beat-like pressure paradigm and the dashed lines representing the internal rhythm of the audio track are clearly marked, visually presenting the phase and frequency differences between the two before synchronization processing.

[0031] For example, if a tourist is near a large waterfall, the environmental pressure spectrum matrix collected by the microphone array shows energy concentration in the 1.0Hz to 2.0Hz frequency band. The algorithm detects the autocorrelation coefficient of the energy sequence at the 1.2Hz frequency point. Reaching 0.85 satisfies the stability criterion of greater than 0.7, thus determining the environmental fundamental beat frequency. The frequency was 1.2 Hz. Simultaneously, the energy envelope identified the first significant pulse moment exceeding two standard deviations. The time is 0.5s. At this point, the acquired metadata for the malleable audio track is calibrated to 1.0Hz (60BPM), marking the start time. The value is 0.2s. Calculate the frequency scaling factor according to the formula. Calculate the time offset The generated beat synchronization parameters include a 1.2x rate scaling instruction and a 0.3s start delay instruction. By applying these parameters, the audio track is accelerated in real time and triggered to play after a 0.3s delay, ensuring that the drumbeats and the pressure fluctuation pulses generated by the waterfall's impact are precisely synchronized in perceived phase.

[0032] Optionally, generating a sound pressure compensation model based on the beat-like pressure paradigm and auditory preference features extracted from user historical behavior data includes: By analyzing long-term listening records from users' historical behavior data, spectral centroid parameters representing preferences for the brightness or warmth of audio are extracted, and auditory preference features are constructed based on these spectral centroid parameters. The auditory preference features are compared with the spectral distribution in the environmental pressure fingerprint to generate an initial compensation curve for adjusting the gain of each frequency band in the malleable audio track material. Based on the energy intensity of the rhythmic pressure paradigm, the gain weights of the corresponding frequency bands in the initial compensation curve are dynamically adjusted to form the final spectral compensation curve, and the spectral compensation curve is encapsulated as a sound pressure compensation model.

[0033] Specifically, auditory preference features, as a multi-dimensional vector characterizing individual sensory characteristics, are fundamentally based on quantifying users' long-term aesthetic preferences for sound pressure frequency distribution. By retrieving nearly a year's worth of digital music platform listening behavior data from a cloud server, including the top 100 most played audio tracks and their corresponding equalizer adjustments, feature extraction was performed using a spectral centroid algorithm. The spectral centroid parameter is defined as... The unit is Hertz (Hz). This parameter represents the physical center of the audio power spectrum on the frequency axis and is used to characterize the brightness or warmth of the timbre. It is calculated by the following formula: ; in, Indicates the first The center frequency corresponding to each frequency point is in Hz, which comes from the frequency sampling points after performing a short-time Fourier transform on the historical listening data. Indicates the first The average power spectral density value corresponding to each frequency point, in W / Hz, is derived from the statistical mean of the energy of the above sampling points; The total number of selected frequency points is fixed at 1024. The calculated... This constitutes an auditory preference feature. This feature is then compared with the spectral distribution in the environmental stress fingerprint. The environmental stress fingerprint... The environmental noise energy at a frequency point is defined as The first in the malleable audio track material The original energy of each frequency point is defined as By comparing the initial compensation curves generated for adjusting the gain of each frequency band of the audio track, its first... Initial gain adjustment value for each frequency point The calculation formula is: ; in, The gain adjustment coefficient is a dimensionless empirical constant with a value of 0.5. To make the audio more rhythmically penetrating in dynamic environments, the initial compensation curve needs to be weighted according to the energy intensity of the beat pressure paradigm. The average peak sound pressure level at the environmental pulse moment in the beat pressure paradigm is defined as... The unit is decibels (dB), and this parameter is derived from the time-domain energy envelope peak value in the environmental pressure fingerprint; the baseline environmental sound pressure level is defined as... The value is fixed at 60dB. Calculate the gain weighting factor. : ; The final spectral compensation curve in the first Final gain weight at each frequency point for This curve is encapsulated as a sound pressure compensation model to guide subsequent synthesis modules in reshaping the spectral energy distribution of the audio track. For example... Figure 3 As shown, a black and white ladder heatmap illustrates the evolution of the gain distribution of the sound pressure compensation model across 32 sub-bands, demonstrating how the system dynamically adjusts the weights of different frequency components based on environmental fingerprints.

[0034] For example, if a tourist's historical listening data is used to calculate their spectral centroid parameter... The frequency of 800Hz indicates a preference for warm, deep tones. When this tourist was sightseeing, environmental pressure fingerprint monitoring detected ambient noise energy at a frequency of 1000Hz. It is 70dB, while the original energy of the currently playing malleable audio track at this frequency is... The gain is 60dB. Based on the initial gain calculation formula, the preferred gain adjustment coefficient is substituted. and the center frequency of the frequency point The initial gain adjustment value for this frequency point is calculated. At this point, the average peak sound pressure level of the environmental beat pressure paradigm was detected. It is 80dB, combined with the reference sound pressure level. The gain weighting factor is calculated. Ultimately, the final gain weight at this 1000Hz frequency point... The sound pressure compensation model boosts the volume of the audio track by 11.68 dB in that frequency band, effectively offsetting the ambient noise masking at 1000 Hz, fine-tuning it according to the user's warm tone preference at 800 Hz, and further enhancing the compensation when the ambient beat is enhanced.

[0035] Optionally, calculating the inverse spatial impulse response based on the pressure wave reflection characteristics and fusing the inverse spatial impulse response with the sound pressure compensation model to generate customized synthesis instructions includes: Based on the spatial geometric topology parameters of the current location and combined with the pressure wave reflection characteristics, a linear time-invariant system model describing the real spatial reverberation effect is constructed, and the spatial impulse response function is calculated. Perform a complex reciprocal operation on the spatial impulse response function in the frequency domain to generate an inverse spatial impulse response that is the opposite of the actual spatial reverberation effect. The frequency domain expression of the inverse spatial impulse response is multiplied point by point with the spectral compensation curve in the sound pressure compensation model to generate a customized synthesis instruction that simultaneously has reverberation cancellation and spectral shaping functions.

[0036] Specifically, pressure wave reflection characteristics, as a core parameter characterizing the acoustic environment of scenic buildings, reflect the energy loss and phase evolution of sound waves at specific spatial boundaries. To quantify this physical effect, a linear time-invariant system model is constructed based on the absorption coefficient of the building material and spatial geometric parameters corresponding to the current location. This model describes the reverberation, delay, and attenuation characteristics of the sound signal during propagation. Furthermore, considering the computational redundancy of mobile terminals, the generation of the inverse spatial impulse response does not employ full-space finite element calculations but is based on a pre-defined simplified parameterized model. In practical implementation, a basic response operator library is pre-established based on typical geometric structures of scenic buildings, such as corridors, domes, and plazas. During real-time calculation, the processor only needs to extract the current geometric feature distance. It then calls the corresponding basic operator and, based on the sound absorption coefficient in the 'Building Material Database', performs the calculation. The energy decay curve of the impulse response is weighted and corrected. This 'basic operator + parameter correction' calculation method greatly reduces the order of the convolution kernel, for example, controlling it to within 512 orders, thus ensuring the engineering feasibility of real-time convolution processing on mobile terminals for tourists and effectively avoiding audio-visual desynchronization problems caused by computational latency. The spatial impulse response function serves as the time-domain representation of this linear time-invariant system model, and its calculation process follows the mirror source algorithm. The expression of the spatial impulse response function in the frequency domain is defined as... : ; in, It represents frequency, with the unit Hz, and is derived from the frequency variable when performing spectrum analysis on audio signals; This indicates the order of sound wave reflection, and its value ranges from 0 to... , The preset maximum number of reflections is fixed at 50. For the first The sound pressure amplitude attenuation coefficient after secondary reflection is dimensionless and is derived from the product of the frequency-related sound absorption coefficients corresponding to the pressure wave reflection characteristics of building materials. For the first The time delay of the reflected sound reaching the visitor's receiving point, measured in seconds (s), is calculated by dividing the geometric distance between the visitor's location and the building boundary by the speed of sound, 340 m / s. Let be the imaginary unit, defined as satisfying The constant is the foundation for constructing complex exponential signals. And the sign... and the index term it belongs to Together they constitute the rotation factor on the complex plane, whose physical meaning lies in describing the first... The secondary reflected sound wave produces a phase lag relative to the original signal. The amount of phase lag is determined by the frequency. With delay time The product of these terms determines the phase evolution of a sound wave as it propagates through space due to physical distance. This is achieved by including an imaginary unit. The computation of linear time-invariant system models allows for the precise characterization of the dual effects of spatial reverberation on the amplitude and phase of audio signals in the complex domain, providing a complete analytical foundation for subsequent calculations of the inverse spatial impulse response. In obtaining... Subsequently, in order to cancel the reverberation interference from the physical space at the playback end, it is necessary to calculate the inverse spatial impulse response. The inverse spatial impulse response is defined in the frequency domain as... Considering that the spatial impulse response function in real-world environments typically has a zero in the frequency domain, directly calculating the reciprocal would lead to gain explosion. Therefore, a regularization technique is used to perform the complex reciprocal operation: ; in, The complex conjugate of the spatial impulse response function; Indicates its amplitude; This is a regularization parameter used to prevent the denominator from being zero; its value is [value to be filled in]. Calculated This is the transfer function that has the opposite effect to the real spatial reverberation. Subsequently, the inverse spatial impulse response is fused with the sound pressure compensation model. The discrete gain sequence in the frequency domain of the spectral compensation curve in the sound pressure compensation model is then presented. ,in Indicates the frequency point index. At the corresponding frequency point Complex values ​​at the location and Perform pointwise multiplication to generate customized synthesis instructions. : ; Generated custom synthesis instructions As a complex vector, it contains both phase compensation information for reducing reverberation and spectral gain information for optimizing the listening experience.

[0037] For example, if a tourist is in a stone-paved corridor, the amplitude of the spatial impulse response function at a frequency of 1000Hz can be calculated based on the characteristics of pressure wave reflection. The complex expression corresponding to a phase delay of 0.5. for At this point, the sound pressure compensation model calculates the final gain weight for the tourist at 1000Hz. The value is 11.68 dB, which translates to a linear gain of approximately 3.84. First, the complex value of the inverse space impulse response at 1000 Hz is calculated, and then the regularization parameter is substituted... ,calculate Next, this complex value is multiplied point-by-point with the spectral compensation gain of 3.84 to obtain the customized synthesis instruction at the 1000Hz frequency point. The instruction was then sent, and through complex multiplication of the audio track material, both pre-compensation for the sound wave reflection at 1000Hz in the stone corridor was achieved, and a personalized spectral boost of 11.68dB was completed.

[0038] Optionally, the method further includes: By combining tourist movement path prediction data, areas of degraded communication link quality ahead of the path can be identified. Before tourists enter the area where the communication link quality degrades, a pre-caching operation is triggered for the beat synchronization parameters and the sound pressure compensation model within a future time window. When a tourist is in an area where the communication link quality degrades, the local clock, the cached beat synchronization parameters, and the sound pressure compensation model are used to perform local audio synthesis on the terminal, enabling uninterrupted playback of music.

[0039] Specifically, movement path prediction data, as a core antecedent variable determining the timing of pre-caching, relies on probabilistic modeling of tourists' historical movement trajectories and the distribution of attraction heatmaps. By collecting GPS coordinate sequences every second from smart terminals, a Hidden Markov Model is used to predict the potential location of tourists within the next minute. The coordinates of the tourist's current location are defined as follows: The predicted path point sequence is By querying the digital map of scenic area signal quality pre-installed on local terminals or in the cloud, predicted waypoints are extracted. Corresponding signal strength prediction value The unit is decibels in milliseconds (dBm). The criterion for identifying regions of communication link quality degradation is defined as: when... The connection remains below the preset threshold. , A value typically set at -105 dBm, and a duration exceeding 5 seconds, indicates a communication link quality degradation zone ahead. Once this zone is identified, a countdown timer for entering it is immediately calculated. : ; in, This represents the geographical distance between the current location and the edge of the weak signal area, in meters (m), derived from the Euclidean distance calculation in the path planning algorithm. This represents the average moving speed of tourists over the past 5 minutes, measured in meters per second (m / s), calculated using a fusion of inertial measurement unit (IMU) and positioning data. When Less than the system's preset pre-cached startup threshold hour, The value is 15 seconds, triggering a future time window. The beat synchronization parameters and sound pressure compensation model are downloaded and pre-cached to the terminal memory. The value is set to 60 seconds. At this time, the terminal obtains the environmental pressure fingerprint evolution trend corresponding to the predicted path within this time window from the server via high-priority instructions, and pre-calculates the corresponding frequency scaling factor and spectral gain weight. When a tourist actually enters a weak signal area causing a communication interruption, the terminal switches to local self-oscillation mode, using the high-precision local clock signal generated by the local quartz crystal oscillator as a reference to drive the cached synchronization parameters. The cached parameters are defined as follows: The start time of each synchronization cycle is The current local clock time is The terminal calculates the current real-time synchronization offset to be applied using the following formula: ; in, The pre-buffered frequency scaling factor is derived from the aforementioned frequency alignment calculation; The pre-buffered time offset is derived from the prediction of environmental pulse moments. In this way, even in a completely offline environment, the terminal can still use local computing power to maintain the continuity of audio synthesis and achieve uninterrupted music playback.

[0040] For example, if a tourist is moving at a speed of 1.2 m / s into the depths of a cave, path prediction indicates that they will enter a completely signal-blocked area after 30 meters. The time required to enter this area is calculated using the formula. At this point, recognizing that the time is approaching the pre-caching activation threshold, the pre-caching operation is immediately initiated, downloading the preset environmental pressure fingerprint data inside the cave to the mobile phone, such as the 1.5Hz dominant frequency of dripping water. When the tourist enters the cave after walking for 25 seconds, the mobile network signal disappears. At this time, the terminal detects a decline in communication link quality and immediately calls the local clock. If the local clock is currently running at the 10th second, and the first beat of the pre-cached sequence begins at the 9.8th second, with a frequency scaling factor of 1.1 and a time offset of 0.1 seconds, the terminal calculates the current synchronization offset as follows: The local synthesis engine adjusts the audio stream in real time based on this 0.32-second offset to ensure that the music beats remain precisely synchronized with the physical pressure fluctuations of the cave environment even when the network is down inside the cave.

[0041] Optionally, the operation of triggering the pre-caching of the beat synchronization parameters and the sound pressure compensation model within a future time window includes: Simultaneously with triggering the pre-buffering operation, a parameterized audio encoding mode is initiated; The parametric audio coding mode decomposes the malleable audio track material into a fundamental frequency component containing the main pitch and energy, and a set of residual parameters describing harmonic and transient characteristics; In the pre-caching operation, only the fundamental frequency component and the residual parameter set are cached, thereby reducing the amount of cached data.

[0042] Specifically, while triggering the pre-buffering operation, a parameterized audio encoding mode is initiated. Its core function is to convert the originally complex waveform audio into a series of low-data-volume control parameters, enabling efficient data transmission before network conditions deteriorate. This mode utilizes an audio source separation algorithm to decompose the malleable audio track material, representing it as a superposition of the fundamental frequency component and the residual parameter set. The fundamental frequency component is defined as... It represents the pitch profile with the most concentrated energy in the audio track, derived from the dominant harmonic features extracted after performing cepstral analysis on the audio track; the residual parameter set is defined as... It includes Mel-frequency cepstral coefficients and high-frequency transient components used to describe timbre texture, derived from the residual distribution after subtracting the fundamental frequency reconstructed signal from the original signal. To minimize the amount of buffered data, a compression factor is calculated. : ; in, and These represent the number of bytes of data occupied in memory for the fundamental frequency component and the residual parameter set, respectively, derived from data statistics after parametric modeling; This represents the number of data bytes in the original waveform file over the same time period, calculated using standard PCM encoding. By caching only these two key parameters, the terminal can reconstruct audio using a local sine synthesizer and noise excitation source during network outages. The locally reconstructed audio signal is defined as... Its calculation logic is as follows: ; in, and These are respectively from the fundamental frequency components The extracted first The amplitude sequence and instantaneous frequency sequence of each harmonic, respectively, are in the form of dimensionless gain and Hertz (Hz). This corresponds to the initial phase value; The total number of harmonics participating in the synthesis is fixed at 12; Represented by residual parameter set As filter coefficients, environmental texture components are generated by exciting convolution of white noise. Through this parameterized reconstruction, the amount of cached data is only 10% to 15% of the original waveform, ensuring that pre-downloading is completed within a limited time before entering the weak signal region.

[0043] For example, if it is necessary to pre-cachate a 60-second mono flexible audio track with a sampling rate of 44100Hz and a bit depth of 16bit, the original waveform data size would be... Approximately 5.29 MB. The extracted fundamental frequency component after activating parametric coding mode. It only contains frequency and amplitude information for each frame, and the data volume is small. The calculated value is 0.4 MB; residual parameter set Recorded 13th-order Mel-frequency cepstral coefficients, data volume The calculated value is 0.3MB. The compression factor is then calculated using the formula. That is, the data volume is compressed to 13.2% of the original. When tourists walk into the cave and the network is disconnected, the mobile terminal calls the local synthesis engine, if the amplitude of the first harmonic... The instantaneous frequency is 0.8. For 440Hz, residual parameters The generated filter gain is -10dB at 3kHz. The terminal then calculates the core melody signal using cosine superposition and adds filtered noise texture. Finally, a highly accurate and small-sized audio stream is generated locally on the phone, ensuring that music can continue to play even without a network connection, and the listening experience is indistinguishable from online streaming.

[0044] Optionally, applying the customized synthesis instructions to perform real-time synthesis of the malleable audio track material includes: The gain of several frequency bands of the plastic audio track material is independently adjusted according to the spectral compensation curve in the sound pressure compensation model to generate a set of adjusted frequency band signals. The adjusted frequency band signals are superimposed and inverse Fourier transform is performed to generate a time-domain intermediate audio signal; The intermediate audio signal in the time domain is convolved with the inverse spatial impulse response in real time to generate and output a customized audio stream.

[0045] Specifically, the real-time synthesis module, as the end of the entire processing chain, is responsible for processing the customized synthesis instructions generated by the aforementioned calculations. The task of transforming this into the physical implementation of the final audio product. In this stage, the malleable audio track material first enters the frequency domain decomposition engine. A polyphase filter bank is then used to divide the material into... Sub-band The value of is fixed at 32. Define the first... The original complex spectrum signal of each sub-band is This originates from the frequency band segmentation after performing a short-time Fourier transform on the malleable audio track material. Based on the spectral compensation curve provided in the sound pressure compensation model, the discrete gain weights at corresponding frequency points are extracted. Each sub-band undergoes independent gain adjustment to generate the adjusted frequency band signal. Its calculation logic is as follows: ; in, The dimensionless spectral compensation curve is the first The final gain weight for each frequency point is derived from the logarithmic inverse calculation of the gain weight in decibel format. Then, this... The adjusted frequency band signals are superimposed, and an inverse Fourier transform (IFFT) is performed to convert the signals back from the frequency domain to the time domain, generating a time-domain intermediate audio signal. To achieve precise offsetting of ambient reverberation, the intermediate signal must be convolved with the inverse spatial impulse response. The final generated customized audio stream is defined as follows: The calculation formula is as follows: ; in, The time-domain function of the inverse spatial impulse response is derived from the aforementioned frequency-domain expression. The result of performing the inverse Fourier transform; symbol This represents a linear convolution operation; The integral variable represents the time delay offset of the signal in the time domain. This convolution process is implemented in the time domain using an overlapping addition method, which can utilize the characteristics of a pre-calculated inverse filter to cancel out multipath reflection interference that will occur in the physical space before the audio output.

[0046] For example, suppose the current malleable audio track material is divided into 32 frequency bands. Taking the 15th frequency band with a center frequency of 1000Hz as an example, its original complex spectrum signal... The amplitude is 1.0. If the sound pressure compensation model calculates the final gain weight at this frequency to be 11.68 dB, the corresponding linear gain weight... The calculated value is approximately 3.84. The adjusted frequency band signal is then calculated using the gain adjustment formula. The amplitude is After superimposing all 32 frequency bands and performing an inverse Fourier transform, a time-domain intermediate audio signal with a peak amplitude of 0.5 is obtained. At this moment, if the reverse spatial impulse response of the stone-paved corridor... exist If the tap coefficient at that point is 0.6, then according to the convolution calculation formula, the output audio stream at that moment... A portion of the components are from Contribution. By performing the aforementioned real-time convolution on a pulse sequence lasting hundreds of milliseconds, a sound wave component completely opposite to the building's reflection path is pre-superimposed in the visitor's headphones. This ensures that when music is reflected in the physical corridor, the reflected sound and the pre-superimposed component interfere and cancel each other out at the visitor's eardrum, ultimately presenting dry, clear, and personalized music that suits individual preferences.

[0047] Optionally, the method further includes: Biomechanical response data of tourists is acquired through the inertial measurement unit of a smart terminal, and the biomechanical response data includes the periodic characteristics of head movements. Calculate the synchronization correlation between the periodic characteristics of the head movements and the beat-like pressure paradigm, and generate a synchronization satisfaction index; Based on the synchronization satisfaction index, the spectral centroid parameter in the auditory preference features is dynamically corrected to optimize the generated sound pressure compensation model.

[0048] Specifically, biomechanical response data, as a crucial feedback indicator for measuring real-time user immersion, essentially transforms unconscious bodily movements into evaluation logic for audio synthesis effects. This is achieved by acquiring triaxial acceleration and triaxial angular velocity data at a sampling frequency of 100Hz using the inertial measurement unit built into the smart terminal. The periodic characteristics of head movements are defined as vectors. It includes the frequency of head movement along the vertical axis. and motion amplitude Using short-time energy analysis and zero-crossing rate detection algorithms, the periodic head movements in sync with music are extracted from the angular velocity time-series signal. A synchronization satisfaction index is defined as... It represents the degree of coupling between user behavior and the environment and music beat, and the calculation formula is: ; in, This indicates the frequency of head movements, measured in Hz, and is derived from the identification of the main peak of the Fast Fourier Transform of inertial data. The environmental fundamental beat frequency in the beat stress paradigm is expressed in Hz. This is the weighting coefficient, with a value of 0.7; The reference motion amplitude constant is fixed at 1.0; The measured motion amplitude is characterized by the variance of the acceleration vector magnitude. This formula quantifies frequency synchronization through an exponential decay term and motion intensity through a logarithmic term. The calculated synchronization satisfaction index... Used to trigger the spectral centroid parameter in auditory preference features Correction. Corrected spectral centroid parameters. The calculation formula is: ; in, The original spectral centroid parameter, in Hz, is derived from the spectral centroid statistics of users' historical listening records; This is the learning rate factor, with a value of 0.2; The exploratory offset step size is selected between -50Hz and 50Hz based on the current environmental energy distribution. Through this closed-loop correction mechanism, when the user's actions become out of sync with the environmental rhythm, the brightness and darkness of the timbre will be automatically fine-tuned to seek a more suitable auditory balance point that matches the user's current physiological state, and a new sound pressure compensation model will be generated accordingly.

[0049] For example, if the environmental basic beat frequency The frequency of head movements in tourists influenced by the music is 1.2Hz. The Hz frequency was measured by the inertial measurement unit, and the motion amplitude was 1.15 Hz. The measured value is 1.5. Substitute this into the weighting coefficient. and reference amplitude Calculate the synchronous satisfaction index If the preset ideal satisfaction range starts at 1.2, then the current value of 1.06 is considered not to have reached the predetermined threshold. At this point, if the original spectral centroid parameter... 800Hz, exploratory offset step size If the frequency is set to 50Hz, then the corrected spectral centroid parameter... Although the magnitude of a single correction is small, with the continuous biomechanical feedback from tourists, the spectral centroid parameter will gradually drift towards the frequency point that makes users feel most comfortable and most rhythmic, thus realizing the dynamic evolution of the sound pressure compensation model.

[0050] Optionally, dynamically correcting the spectral centroid parameter in the auditory preference features based on the synchronization satisfaction index includes: When the synchronization satisfaction index is in the predetermined high satisfaction range, the current spectrum centroid parameter remains unchanged. When the synchronization satisfaction index does not reach the predetermined low satisfaction threshold, an exploratory adjustment is performed on the spectrum centroid parameter, which involves slightly shifting the frequency value within the target frequency threshold range. Using the corrected spectral centroid parameters, the sound pressure compensation model is regenerated and applied to subsequent audio synthesis.

[0051] Specifically, during the correction of the spectral centroid parameters, a clear gradient adjustment logic is followed to ensure a smooth transition in listening experience. The threshold for the high satisfaction range is defined as... The value is set to 1.2; a low satisfaction threshold is defined. The value is 0.8. When the synchronous satisfaction index... When the value is greater than or equal to 1.2, it is determined that the current audio synthesis effect is highly matched with the user's physiological feedback, and the current spectral centroid parameter is maintained. Unchanged. When When the value is less than 0.8, it is determined that the user is significantly uncomfortable with the current timbre or that excessive environmental interference is causing a break in beat perception. In this case, exploratory adjustments are initiated. The adjustment process utilizes a correction step size function. ; in, and These are the synchronization satisfaction indices for the current time and the previous sampling time, respectively. It is a symbolic function; This is the scaling factor, with a value of 0.01. This logic ensures that when satisfaction is trending upwards, the adjustment direction continues along the original path; otherwise, it explores in reverse. Using the corrected spectral centroid parameters, the sound pressure compensation modeling module will re-invoke the formula to update the initial compensation curve: ; This real-time parameter reconstruction based on biofeedback makes the sound pressure compensation model no longer a static preset, but a dynamic model that can be optimized in real time according to changes in biomechanical characteristics caused by tourists' fatigue and emotional fluctuations.

[0052] For example, if tourists have a simultaneous satisfaction index at the beginning of their visit... The threshold value is 1.3, greater than 1.2, while maintaining the spectral centroid parameter at 800Hz. After one hour of sightseeing, due to changes in the environmental noise spectrum, the frequency of tourists' head movements decreases, leading to... It dropped to 0.75, hitting the low threshold of 0.8. The previous timeframe was detected. It is 0.78, currently. The value is 0.75, and the trend is negative. Therefore, based on the sign function and a scaling factor of 0.01, the correction step size is calculated as follows: The new spectral centroid parameter is set to 792Hz, shifting the track's timbre towards a warmer direction. Subsequently, the initial gain for each frequency band is recalculated using 792Hz as the denominator. If a certain frequency point The original gain was 10.16dB, and the frequency is 1000Hz. The updated gain will be changed to... Through these subtle but continuous adjustments, the music recommendation system can constantly adapt to the real-time physiological state of tourists, optimizing their overall interactive experience in complex cultural and tourism scenarios.

[0053] Based on the same inventive concept, such as Figure 4As shown, the present invention also provides a music recommendation system based on cultural tourism big data, the system comprising: An environmental fingerprint generation module is used to acquire the dynamic environmental pressure fluctuation signal of the tourist's current location and the pressure wave reflection characteristics determined by the building materials of the scenic spot, and to fuse the dynamic environmental pressure fluctuation signal and the pressure wave reflection characteristics to generate an environmental pressure fingerprint. The beat parameter extraction module is used to extract a beat pressure paradigm with periodic regularity from the environmental pressure fingerprint, and generate beat synchronization parameters based on the beat pressure paradigm. The sound pressure compensation modeling module is used to generate a sound pressure compensation model based on the beat pressure paradigm and the auditory preference features extracted from the user's historical behavior data. The sound pressure compensation model includes a spectral compensation curve for adjusting the spectral energy distribution of the audio material. The instruction fusion generation module is used to calculate the inverse spatial impulse response based on the pressure wave reflection characteristics, and to fuse the inverse spatial impulse response with the sound pressure compensation model to generate customized synthesis instructions; The real-time synthesis output module is used to acquire malleable audio track materials that match the cultural attributes of the scenic spot, and to apply the customized synthesis instructions to synthesize the malleable audio track materials in real time, outputting a customized audio stream that is synchronized with the environmental pressure beat.

[0054] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0055] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A music pushing method based on travel and tourism big data, characterized in that, The method includes: The system acquires the environmental dynamic pressure fluctuation signal of the tourist's current location and the pressure wave reflection characteristics determined by the building materials of the scenic spot, and integrates the environmental dynamic pressure fluctuation signal and the pressure wave reflection characteristics to generate an environmental pressure fingerprint. The environmental dynamic pressure fluctuation signal is frequency-decoupled, low-frequency pressure features outside the hearing frequency range are extracted, a beat paradigm is constructed, and acoustic spectrum features within the hearing frequency range are extracted to construct a spectrum compensation basis. Extract a periodic pressure paradigm from the environmental pressure fingerprint, and generate beat synchronization parameters based on the periodic pressure paradigm. Based on the beat-like pressure paradigm and the auditory preference features extracted from user historical behavior data, a sound pressure compensation model is generated, which includes a spectral compensation curve for adjusting the spectral energy distribution of the audio material. The inverse spatial impulse response is calculated based on the pressure wave reflection characteristics, and the inverse spatial impulse response is fused with the sound pressure compensation model to generate a customized synthesis instruction; Acquire malleable audio track materials that match the cultural attributes of the scenic spot, and apply the customized synthesis instructions to synthesize the malleable audio track materials in real time, outputting a customized audio stream that is synchronized with the environmental pressure beat.

2. The music recommendation method based on cultural tourism big data according to claim 1, characterized in that, Extracting a periodic pressure paradigm from the environmental pressure fingerprint, and generating beat synchronization parameters based on the periodic pressure paradigm, includes: Time-frequency analysis was performed on the environmental dynamic pressure fluctuation signal in the environmental pressure fingerprint to identify the dominant frequency component whose amplitude changes periodically with time. The center frequency and energy envelope of the dominant frequency component are extracted and together constitute the rhythmic pressure paradigm. The center frequency of the beat-pressure paradigm is matched with the internal rhythm of the malleable audio track material to generate a time mapping relationship for synchronizing the beats of the two, and the time mapping relationship is encapsulated as beat synchronization parameters.

3. The music recommendation method based on cultural tourism big data according to claim 1, characterized in that, Based on the aforementioned rhythmic pressure paradigm and auditory preference features extracted from user historical behavior data, a sound pressure compensation model is generated, including: By analyzing long-term listening records from users' historical behavior data, spectral centroid parameters representing preferences for the brightness or warmth of audio are extracted, and auditory preference features are constructed based on these spectral centroid parameters. The auditory preference features are compared with the spectral distribution in the environmental pressure fingerprint to generate an initial compensation curve for adjusting the gain of each frequency band in the malleable audio track material. Based on the energy intensity of the rhythmic pressure paradigm, the gain weights of the corresponding frequency bands in the initial compensation curve are dynamically adjusted to form the final spectral compensation curve, and the spectral compensation curve is encapsulated as a sound pressure compensation model.

4. The music recommendation method based on cultural tourism big data according to claim 1, characterized in that, The inverse spatial impulse response is calculated based on the pressure wave reflection characteristics, and the inverse spatial impulse response is fused with the sound pressure compensation model to generate a customized synthesis instruction, including: Based on the spatial geometric topology parameters of the current location and combined with the pressure wave reflection characteristics, a linear time-invariant system model describing the real spatial reverberation effect is constructed, and the spatial impulse response function is calculated. Perform a complex reciprocal operation on the spatial impulse response function in the frequency domain to generate an inverse spatial impulse response that is the opposite of the actual spatial reverberation effect. The frequency domain expression of the inverse spatial impulse response is multiplied point by point with the spectral compensation curve in the sound pressure compensation model to generate a customized synthesis instruction that simultaneously has reverberation cancellation and spectral shaping functions.

5. The music recommendation method based on cultural tourism big data according to claim 1, characterized in that, The method further includes: By combining tourist movement path prediction data, areas of degraded communication link quality ahead of the path can be identified. Before tourists enter the area where the communication link quality degrades, a pre-caching operation is triggered for the beat synchronization parameters and the sound pressure compensation model within a future time window. When a tourist is in an area where the communication link quality degrades, the local clock, the cached beat synchronization parameters, and the sound pressure compensation model are used to perform local audio synthesis on the terminal, enabling uninterrupted playback of music.

6. The music recommendation method based on cultural tourism big data according to claim 5, characterized in that, The operation of triggering the pre-caching of the beat synchronization parameters and the sound pressure compensation model within a future time window includes: Simultaneously with triggering the pre-buffering operation, a parameterized audio encoding mode is initiated; The parametric audio coding mode decomposes the malleable audio track material into a fundamental frequency component containing the main pitch and energy, and a set of residual parameters describing harmonic and transient characteristics; In the pre-caching operation, only the fundamental frequency component and the residual parameter set are cached, thereby reducing the amount of cached data.

7. The music recommendation method based on cultural tourism big data according to claim 4, characterized in that, The real-time synthesis of the malleable audio track material using the customized synthesis instructions includes: The gain of several frequency bands of the plastic audio track material is independently adjusted according to the spectral compensation curve in the sound pressure compensation model to generate a set of adjusted frequency band signals. The adjusted frequency band signals are superimposed and inverse Fourier transform is performed to generate a time-domain intermediate audio signal; The intermediate audio signal in the time domain is convolved with the inverse spatial impulse response in real time to generate and output a customized audio stream. The real-time convolution operation is based on a preset simplified order response kernel generated by the linear time-invariant system to reconstruct the frequency domain of the customized audio stream to compensate for acoustic distortion in the physical space.

8. The music recommendation method based on cultural tourism big data according to claim 1, characterized in that, The method further includes: Biomechanical response data of tourists is acquired through the inertial measurement unit of a smart terminal, and the biomechanical response data includes the periodic characteristics of head movements. Calculate the synchronization correlation between the periodic characteristics of the head movements and the beat-like pressure paradigm, and generate a synchronization satisfaction index; Based on the synchronization satisfaction index, the spectral centroid parameter in the auditory preference features is dynamically corrected to optimize the generated sound pressure compensation model.

9. The music recommendation method based on cultural tourism big data according to claim 8, characterized in that, Based on the synchronization satisfaction index, the dynamic correction of the spectral centroid parameter in the auditory preference features includes: When the synchronization satisfaction index is in the predetermined high satisfaction range, the current spectrum centroid parameter remains unchanged. When the synchronization satisfaction index does not reach the predetermined low satisfaction threshold, an exploratory adjustment is performed on the spectrum centroid parameter, which involves slightly shifting the frequency value within the target frequency threshold range. Using the corrected spectral centroid parameters, the sound pressure compensation model is regenerated and applied to subsequent audio synthesis.

10. A music recommendation system based on cultural tourism big data, characterized in that: The system is used for the music recommendation method based on cultural and tourism big data as described in any one of claims 1-9, the system comprising: An environmental fingerprint generation module is used to acquire the environmental dynamic pressure fluctuation signal of the tourist's current location and the pressure wave reflection characteristics determined by the building materials of the scenic spot, and to fuse the environmental dynamic pressure fluctuation signal and the pressure wave reflection characteristics to generate an environmental pressure fingerprint. In this module, the environmental dynamic pressure fluctuation signal is decoupled by frequency division, low-frequency pressure features outside the hearing frequency range are extracted to construct a beat paradigm, and acoustic spectrum features within the hearing frequency range are extracted to construct a spectrum compensation basis. The beat parameter extraction module is used to extract a beat pressure paradigm with periodic regularity from the environmental pressure fingerprint, and generate beat synchronization parameters based on the beat pressure paradigm. The sound pressure compensation modeling module is used to generate a sound pressure compensation model based on the beat pressure paradigm and the auditory preference features extracted from the user's historical behavior data. The sound pressure compensation model includes a spectral compensation curve for adjusting the spectral energy distribution of the audio material. The instruction fusion generation module is used to calculate the inverse spatial impulse response based on the pressure wave reflection characteristics, and to fuse the inverse spatial impulse response with the sound pressure compensation model to generate customized synthesis instructions; The real-time synthesis output module is used to acquire malleable audio track materials that match the cultural attributes of the scenic spot, and to apply the customized synthesis instructions to synthesize the malleable audio track materials in real time, outputting a customized audio stream that is synchronized with the environmental pressure beat.

Images