Bluetooth sound box intelligent regulation and control method and system and Bluetooth sound box

Abstract

The invention relates to the technical field of sound box equipment, in particular to an intelligent regulation and control method and system for a Bluetooth sound box and the Bluetooth sound box, and the method comprises the steps: judging whether audio delay meets the requirement or not; in response to the situation that the playing state does not meet the requirement, storing the duration and the transmission parameter into a fault data set; training a probability model; predicting the probability that the playing state in the next period does not meet the requirement and the duration; in response to the probability greater than a preset probability, adjusting the cache depth according to the duration, and in response to the probability less than or equal to the preset probability, obtaining a human voice ratio in the audio file; and adjusting the gain or cache depth of the medium-high frequency part in the audio according to the human voice ratio. A multi-task probability model is trained by using the strong relevance of historical transmission parameters and delay faults, and the delay exceeding risk and duration of the next period are predicted, so that graded accurate regulation and control based on prediction results are realized, and the operation stability of the Bluetooth sound box is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of speaker equipment, and in particular to a method and system for intelligently controlling a Bluetooth speaker, and a Bluetooth speaker. Background Art

[0002] With the rapid development of voice recognition technology, Bluetooth speakers are becoming increasingly popular due to their small size, low power consumption, and portability. However, since Bluetooth speakers cannot directly connect to the internet, they must first establish a Bluetooth connection with a mobile phone or other terminal to transmit the audio data to the speaker via Bluetooth for playback.

[0003] Patent publication number CN110418245A discloses a method, apparatus, terminal device, and computer-readable storage medium for reducing Bluetooth speaker response delay. The method includes: upon receiving a wake-up message, executing a wake-up interrupt operation to wake up the Bluetooth speaker; after the Bluetooth speaker wakes up, buffering received voice data; detecting whether invalid data exists in the buffered voice data; and, if invalid data exists in the buffered voice data, determining that the data other than the invalid data in the buffered voice data is valid data. This technical solution is affected by fluctuations in the Bluetooth signal, which in turn affects the audio latency of the audio file, hindering the audio experience and resulting in low stability. Summary of the Invention

[0004] The purpose of the present invention is to provide a Bluetooth speaker intelligent control method, system and Bluetooth speaker to solve the problem in the prior art that when the Bluetooth signal fluctuates, the data cache will be affected, thereby affecting the audio delay of the audio file, affecting the experience and low stability.

[0005] The present invention provides a Bluetooth speaker intelligent control method, comprising: Get the audio delay of Bluetooth speakers in real time; Determining whether the playback status meets the requirements based on the audio delay; In response to the audio delay being greater than the delay threshold, determining that the playback state does not meet the requirements, and storing the duration of the playback state not meeting the requirements and the transmission parameters when the playback state does not meet the requirements in a fault data set; training a probability model based on the fault data set; Based on the transmission parameters in the current cycle and the probability model, the probability that the playback status does not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements are predicted. In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration. In response to the probability being less than or equal to the preset probability, the proportion of human voice in the audio file is obtained, and the speaker operating parameters are adjusted according to the human voice proportion.

[0006] As a preferred technical solution of the method for intelligently controlling a Bluetooth speaker, in response to the audio delay being greater than a delay threshold, determining that the playback state does not meet the requirements, and storing the duration of the playback state not meeting the requirements and the transmission parameters when the playback state does not meet the requirements in a fault data set include: Classify the fault level according to the magnitude of the audio delay exceeding the threshold; Different storage weights are assigned to different levels of faults. The higher the audio delay exceeds the threshold, the higher the storage weight. Data is stored in descending order of weight value, and high-weight data is retained first when storage space is insufficient.

[0007] As an optimal technical solution for a method for intelligently controlling a Bluetooth speaker, the probability model is established based on the duration of time the playback state does not meet the requirements and the corresponding transmission parameters as a training set. The probability model is trained using a multi-task learning method. If the transmission parameters are input, the probability that the playback state does not meet the requirements and the duration of time the playback state does not meet the requirements are output.

[0008] As a preferred technical solution of a method for intelligently controlling a Bluetooth speaker, the method predicts the probability of a playback state not meeting the requirements in the next cycle and the duration of the non-compliance with the requirements based on the transmission parameters in the current cycle and the probability model, and in response to the probability being zero, increases the audio sampling rate to a higher level; In response to the probability being less than or equal to the preset probability and not being zero, separating the human voice and background noise in the audio file, calculating the proportion of the human voice in the total audio and recording the proportion as the human voice ratio, and determining the adjusted operating parameters according to the human voice ratio, including: a gain or a buffer depth of a mid-high frequency portion of the audio; In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration.

[0009] As a preferred technical solution for a method for intelligently controlling a Bluetooth speaker, the operating parameters for adjusting the parameters according to the proportion of human voice include: In response to the human voice proportion being greater than a proportion threshold, using a frequency band equalizer to perform gain processing on a mid-high frequency portion of the audio signal; In response to the human voice proportion being less than or equal to the proportion threshold, the cache depth is reduced.

[0010] As a preferred technical solution of a Bluetooth speaker intelligent control method, adjusting the cache depth according to the duration includes: Get the predicted duration; In the historical playback records, several time windows with a duration are selected; Calculate the data volume of the audio file played in each time window and determine the maximum data volume; Adjusting the current cache depth to the target cache depth based on the maximum value of the data volume in combination with a smooth transition algorithm, and loading subsequent audio data in advance; When caching the maximum value of the data volume, the target cache depth is 1.2 times or more of the required cache depth.

[0011] As an optimal technical solution for a Bluetooth speaker intelligent control method, the transmission parameters are transmission parameters within a preset time period before the audio delay exceeds the delay threshold, including: signal strength, retransmission rate, and throughput fluctuation rate.

[0012] As a preferred technical solution for a method for intelligently controlling a Bluetooth speaker, the method of obtaining the audio delay of the Bluetooth speaker in real time includes: Embed timestamps into audio data packets at the sending end of the audio file; Obtaining the local system time when decoding of the audio data packet starts, calculating the time difference between the time corresponding to the timestamp and the local system time when decoding of the audio data packet starts, to obtain an initial delay; Kalman filtering is performed on the initial delay to eliminate the influence of measurement noise and obtain a final audio delay.

[0013] The present invention also provides a Bluetooth speaker intelligent control system, comprising: Detection module to obtain the audio delay of the Bluetooth speaker; An analysis module, configured to determine whether a playback state meets the requirements based on the audio delay; a data collection module, in response to the audio delay being greater than a delay threshold, determining that the playback state does not meet the requirements, and storing a duration of the playback state not meeting the requirements and transmission parameters when the playback state does not meet the requirements in a fault data set; A training module, configured to train a probability model based on the fault data set; An adjustment module is used to predict the probability that the playback status does not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements based on the transmission parameters in the current cycle and the probability model. In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration. In response to the probability being less than or equal to the preset probability, the proportion of human voice in the audio file is obtained, and the speaker operating parameters are adjusted according to the human voice proportion.

[0014] The present invention also provides a Bluetooth speaker, comprising a digital signal processor, a main control processor, a memory, and a computer program stored in the memory and executable on the main control processor. When the computer program is executed by the processor, the steps of the above-mentioned Bluetooth speaker intelligent control method are implemented.

[0015] Compared with the existing technology, the beneficial effect of the present invention lies in that the probability prediction model is trained by the operating parameters stored in the fault data set when the playback status does not meet the requirements, so as to predict the probability of whether there will be a playback status that does not meet the requirements in the future and the duration of the duration based on the current operating parameters. The dual analysis decision combined with the proportion of human voice achieves a balance between audio delay and sound quality optimization, thereby increasing the stability of the Bluetooth speaker during operation.

[0016] Furthermore, when the probability is greater than the preset probability, the maximum amount of data of the audio file played within the duration is determined through historical playback records, and the cache depth is adjusted according to the maximum amount of data, the cache depth is dynamically expanded and data is preloaded, so that before the audio delay does not meet the requirements, the audio delay is avoided by playing the preloaded data, and an "anti-fluctuation buffer" is actively constructed. When the probability is less than or equal to the preset probability and is not zero, the audio voice ratio is separated, and the mid- and high-frequency gain is targetedly enhanced (for voice scenarios) or the cache depth is reduced (for music scenarios) to achieve the optimal balance between sound quality and delay. In application, it can also be combined with the collaborative optimization of the underlying hardware of the speaker: use timestamp filtering to eliminate monitoring errors, and use dedicated DSP to accelerate voice separation, so that when the Bluetooth data transmission fluctuates, the audio playback will not be interrupted due to network fluctuations or content switching, thereby further increasing the stability of the Bluetooth speaker during operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flowchart of the steps of the Bluetooth speaker intelligent control method according to an embodiment of the present invention; Figure 2 This is a structural block diagram of the Bluetooth speaker intelligent control system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present application by illustrating the examples of the present application.

[0019] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, the elements defined by the phrase "comprising..." do not exclude the presence of other identical elements in the process, method, article, or device comprising the elements.

[0020] See also Figure 1 As shown, it is a flowchart of the steps of the Bluetooth speaker intelligent control method according to an embodiment of the present invention, including: Step S1, obtaining the audio delay of the Bluetooth speaker in real time; Step S2, judging whether the playing status of the Bluetooth speaker meets the requirements based on the audio delay; Step S3: In response to the audio delay being greater than the delay threshold, determining that the playback state does not meet the requirements, and storing the duration of the playback state not meeting the requirements and the transmission parameters when the playback state does not meet the requirements in a fault data set; Step S4, training a probability model based on the fault data set; Step S5, based on the transmission parameters in the current cycle and the probability model, predict the probability that the playback status does not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements. In response to the probability being greater than the preset probability, adjust the cache depth according to the duration. In response to the probability being less than or equal to the preset probability, obtain the proportion of human voice in the audio file, and adjust the speaker operating parameters according to the proportion of human voice.

[0021] In detail, the probability prediction model is trained by using the operating parameters stored in the fault data set when the playback status does not meet the requirements, so as to predict the probability of whether the playback status will not meet the requirements in the future and the duration of the problem based on the current operating parameters. A dual analysis decision combined with the proportion of human voice is made to achieve a balance between audio delay and sound quality optimization, thereby increasing the stability of the Bluetooth speaker during operation.

[0022] Furthermore, determining whether the audio delay meets the requirements includes: In response to the audio delay being less than or equal to the delay threshold, determining that the requirement is met; In response to the audio delay being greater than the delay threshold, it is determined that the requirements are not met, the duration of the non-compliant playback state and the transmission parameters when the non-compliant playback state is obtained, and the obtained data are stored in a fault data set.

[0023] In implementation, the selection of the delay threshold is determined according to actual usage requirements. The maximum delay range acceptable to the human ear in different scenarios is different. For example, for daily music playback, the human ear is less sensitive to delay. If there is no visual synchronization requirement (such as no video screen), a delay of 200-300ms generally does not significantly affect the auditory experience. For video viewing and calls, the human ear is more sensitive to the asynchrony between audio and video, and the maximum acceptable delay is usually 150-200ms. If the delay exceeds 200ms, the asynchrony between audio and video will be clearly perceived, resulting in a decreased experience. For gaming scenarios, such as action games or real-time interactive scenarios, players have extremely high requirements for the delay of operation feedback. The delay usually needs to be controlled below 100ms, and ideally below 50ms, to avoid disconnection between operation and sound feedback. Preferably, the delay threshold of an embodiment of the present invention is 150ms.

[0024] Furthermore, in step S3, if the playback state does not meet the requirements, the duration of the playback state not meeting the requirements and the transmission parameters when the playback state does not meet the requirements are stored in the fault data set, including: Step S31, classifying the fault level according to the magnitude of the audio delay exceeding the threshold; Step S32: assigning differentiated storage weights to faults of different levels. The higher the magnitude of the audio delay exceeding the threshold, the higher the storage weight. Step S33: Store data in descending order of weight value, and prioritize retaining high-weight data when storage space is insufficient.

[0025] In implementation, the fault levels are divided into three levels: Level I: the amplitude of the audio delay exceeding the threshold is less than or equal to 50ms; Level II: the amplitude of the audio delay exceeding the threshold is greater than 50ms and less than or equal to 100ms; Level III: the amplitude of the audio delay exceeding the threshold is greater than 100ms; differentiated storage weights are assigned to different levels of faults, among which: Level I: weight 1; Level II: weight 3; Level III: weight 5.

[0026] In detail, by adopting fault level classification and differentiated weight storage strategy for the data in the fault data set, the system can give priority to retaining high-risk fault data, effectively utilize limited storage space, and ensure that key fault features are not discarded. This improves the accuracy of model training, increases the accuracy of prediction for the probability that the playback status will not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements, adjusts the operating parameters of the speaker according to the prediction results, and increases the ability of the speaker to resist the playback status not meeting the requirements, thereby increasing the stability of the Bluetooth speaker during operation.

[0027] Furthermore, in step S4, a probability model is established based on the duration of the playback state not meeting the requirements and the corresponding transmission parameters as a training set. The probability model is trained using a multi-task learning method. If the transmission parameters are input, the probability of the playback state not meeting the requirements and the duration of the playback state not meeting the requirements are output. In actual applications, there is a direct influence relationship between the transmission parameters and the audio delay. Therefore, it is feasible to obtain a mapping relationship between the probability of the playback state not meeting the requirements and the duration of the playback state not meeting the requirements and the transmission parameters by collecting transmission parameter data and audio delay data for training. The process of training and establishing the model using known data is all existing technology and will not be elaborated here.

[0028] Specifically, through multi-task joint learning of the feature expression of shared transmission parameters and synchronous optimization of probability classification and duration regression accuracy, the model can simultaneously output the fault risk level and duration of the next cycle, providing a quantitative basis for cache depth adjustment, avoiding excessive or insufficient cache adjustment that leads to audio and video asynchrony or resource waste, maintaining a stable and efficient playback process, and further increasing the stability of the Bluetooth speaker during operation.

[0029] Furthermore, in step S5, the probability of the playback status not meeting the requirements in the next cycle and the duration of the playback status not meeting the requirements are predicted based on the transmission parameters in the current cycle and the probability model, including: In response to the probability being equal to zero, increasing the audio sampling rate to a higher level; Specifically, in practical applications, the commonly used audio sampling rates are: 8kHz, commonly used for low-quality voice communications.

[0030] 16kHz is commonly used in speech coding and speech recognition to improve speech clarity.

[0031] 32kHz, commonly used for broadcast, low-bandwidth audio of medium quality, suitable for voice.

[0032] 44.1kHz, commonly used in high-fidelity sound quality such as CD, MP3, WAV, suitable for music.

[0033] 48kHz is a standard commonly used for video and audio synchronization in film and video production.

[0034] 96kHz is often used in high-resolution audio applications where high-quality audio is sought.

[0035] 192kHz is commonly used in professional recording and high-fidelity audio for ultra-high resolution, pursuing the ultimate sound quality.

[0036] Preferably, the audio sampling rate levels from low to high are: 8kHz, 16kHz, 32kHz, 44.1kHz, 48kHz, 96kHz, and 192kHz.

[0037] In response to the probability being less than or equal to a preset probability and not being zero, separating the human voice and background noise in the audio file, calculating the proportion of the human voice in the total audio and recording the proportion as the human voice ratio, and determining the adjusted operating parameters according to the human voice ratio, including: a gain or a buffer depth of a mid-high frequency portion of the audio; In response to the probability being greater than a preset probability, the cache depth is adjusted according to the duration.

[0038] Furthermore, in step S5, the operating parameters to be adjusted are determined according to the proportion of human voice, including: In response to the human voice proportion being greater than a proportion threshold, using a frequency band equalizer to perform gain processing on a mid-high frequency portion of the audio signal; In implementation, the calculation of the proportion of human voice is achieved through audio signal processing technology, and the calculation of the proportion of human voice is achieved through a deep neural network sound source separation model. This model runs locally on the playback end and separates the voice and background sound components in the audio in real time. The proportion threshold is a preset value based on factors such as user experience, playback status environment and device performance. The proportion threshold can be adjusted according to real-time audio data. Preferably, in the embodiment of the present invention, the proportion threshold is 40%. The gain of the mid- and high-frequency parts of the audio is adjusted by an audio equalizer: by adjusting the audio equalizer, the gain is adjusted in the frequency band of 1kHz to 5kHz. This frequency band is the main frequency area for most languages ​​and human voices. Therefore, by increasing the gain in this frequency band, the performance of the human voice can be improved, making it clearer and easier to identify. The specific range of the gain (3-6dB) is dynamically adjusted based on the actual proportion of the human voice. The frequency band equalizer is adjusted based on the characteristics of the audio file, the acoustic performance of the current device, and the real-time network status. If the proportion of human voice reaches a threshold, the mid- and high-frequency gain is gradually increased to optimize the sound effect. The adjustment range of the frequency gain is dynamically adjusted based on the current network bandwidth, Bluetooth signal quality, and noise interference in the playback environment. The audio equalizer and the audio equalizer adjustment process are existing technologies and will not be described in detail here.

[0039] In response to the human voice ratio being less than or equal to the ratio threshold, the cache depth is reduced.

[0040] Specifically, when the proportion of human voices is low, the system will reduce the buffer depth to optimize audio playback latency. This operation can reduce playback latency and improve the real-time performance of the audio stream, avoiding audio stuttering or delays. The buffer depth can be reduced to the minimum value that does not affect the playback status. Dynamically adjusting the buffer depth is an existing technology in the audio adjustment process and will not be detailed here.

[0041] Furthermore, the cache depth is increased according to the duration, including: Get the predicted duration; In the historical playback records, several time windows with a duration are selected; Calculate the data volume of the audio file played in each time window and determine the maximum data volume; Adjust the current cache depth to the target cache depth based on the maximum data volume combined with a smooth transition algorithm, and load subsequent audio data in advance; When caching the maximum value of the data volume, the target cache depth is 1.2 times or more of the required cache depth. The historical playback record is the playback record of the audio file after the Bluetooth connection.

[0042] Specifically, the target cache depth is set at 1.2 times the maximum historical playback data volume, and a smooth transition algorithm is used to ensure that the cache adjustment process is imperceptible to the user. Actual tests have shown that the algorithm can complete cache depth adjustment within 200ms, and the audio quality fluctuation during the adjustment process is less than 0.5dB. This dynamic caching mechanism effectively addresses sudden data fluctuations in Bluetooth transmission, ensuring smooth playback while avoiding the resource waste or excessive latency caused by traditional fixed caching strategies, thereby further improving the stability of the Bluetooth speaker during operation.

[0043] Specifically, the transmission parameters are transmission parameters within a preset time period before the audio delay is greater than the delay threshold, including: signal strength, retransmission rate, and throughput fluctuation rate.

[0044] Furthermore, obtaining the audio delay of the Bluetooth speaker in real time includes: Synchronize the local clocks of the Bluetooth speaker and the data transmitter; Embed timestamps into audio data packets at the sending end of the audio file; Obtain the local system time when the audio data packet starts decoding, calculate the time difference between the time corresponding to the timestamp and the local system time when the audio data packet starts decoding, and obtain the initial delay; The initial delay is processed by Kalman filtering to eliminate the influence of measurement noise and obtain the final audio delay.

[0045] Specifically, this method uses Kalman filtering to process audio delay data, effectively eliminating measurement noise caused by Bluetooth signal fluctuations. Actual tests have shown that this method can reduce delay measurement error from ±20ms to within ±5ms, improving the system's sensitivity to small delay variations. Accurate delay measurement is the foundation for intelligent control. The application of Kalman filtering enables the system to detect delay change trends earlier, allowing for more timely adjustment decisions, further increasing the stability of the Bluetooth speaker during operation.

[0046] See also Figure 2 As shown in FIG, it is a structural block diagram of the Bluetooth speaker intelligent control system according to an embodiment of the present invention, including: Detection module to obtain the audio delay of the Bluetooth speaker; An analysis module, configured to determine whether a playback state meets the requirements based on the audio delay; a data collection module, in response to the audio delay being greater than a delay threshold, determining that the playback state does not meet the requirements, and storing a duration of the playback state not meeting the requirements and transmission parameters when the playback state does not meet the requirements in a fault data set; A training module, configured to train a probability model based on the fault data set; An adjustment module is used to predict the probability that the playback status does not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements based on the transmission parameters in the current cycle and the probability model. In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration. In response to the probability being less than or equal to the preset probability, the proportion of human voice in the audio file is obtained, and the speaker operating parameters are adjusted according to the human voice proportion.

[0047] Furthermore, an embodiment of the present invention also provides a Bluetooth speaker, including a digital signal processor, a main control processor, a memory, and a computer program stored in the memory and executable on the main control processor, wherein the computer program implements any step of the above method when executed by the processor.

[0048] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. A person skilled in the art would be able to make other variations or modifications based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for intelligently controlling a Bluetooth speaker, characterized in that: include: Get the audio delay of Bluetooth speakers in real time; Determining whether the playback status meets the requirements based on the audio delay; In response to the audio delay being greater than the delay threshold, determining that the playback state does not meet the requirements, and storing the duration of the playback state not meeting the requirements and the transmission parameters when the playback state does not meet the requirements in a fault data set; training a probability model based on the fault data set; Based on the transmission parameters in the current cycle and the probability model, the probability that the playback status does not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements are predicted. In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration. In response to the probability being less than or equal to the preset probability, the proportion of human voice in the audio file is obtained, and the speaker operating parameters are adjusted according to the proportion of human voice.

2. The Bluetooth speaker intelligent control method according to claim 1, characterized in that: In response to the audio delay being greater than a delay threshold, determining that the playback state does not meet the requirements, and storing the duration of the playback state not meeting the requirements and the transmission parameters when the playback state does not meet the requirements in a fault data set, include: Classify the fault level according to the magnitude of the audio delay exceeding the threshold; Different storage weights are assigned to different levels of faults. The higher the audio delay exceeds the threshold, the higher the storage weight. Data is stored in descending order of weight value, and high-weight data is retained first when storage space is insufficient.

3. The Bluetooth speaker intelligent control method according to claim 1, characterized in that: The probability model is established based on the duration of the playback status not meeting the requirements and the corresponding transmission parameters as a training set. The probability model is trained using a multi-task learning method. If the transmission parameters are input, the probability of the playback status not meeting the requirements and the duration of the playback status not meeting the requirements are output.

4. The method for intelligently controlling a Bluetooth speaker according to claim 1, wherein: predicting the probability of the playback state not meeting the requirements in the next cycle and the duration of the non-compliance with the requirements based on the transmission parameters in the current cycle and the probability model, and increasing the audio sampling rate to a higher level in response to the probability being zero; In response to the probability being less than or equal to the preset probability and not being zero, separating the human voice and background noise in the audio file, calculating the proportion of the human voice in the total audio and recording the proportion as the human voice ratio, and determining the adjusted operating parameters according to the human voice ratio, including: a gain or a buffer depth of a mid-high frequency portion of the audio; In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration.

5. The method for intelligently controlling a Bluetooth speaker according to claim 4, wherein: The operating parameters for adjustment determined according to the proportion of human voice include: In response to the human voice proportion being greater than a proportion threshold, using a frequency band equalizer to perform gain processing on a mid-high frequency portion of the audio signal; In response to the human voice proportion being less than or equal to the proportion threshold, the cache depth is reduced.

6. The Bluetooth speaker intelligent control method according to claim 4, characterized in that: The adjusting the cache depth according to the duration includes: Get the predicted duration; In the historical playback records, several time windows with a duration are selected; Calculate the data volume of the audio file played in each time window and determine the maximum data volume; Adjusting the current cache depth to the target cache depth based on the maximum value of the data volume in combination with a smooth transition algorithm, and loading subsequent audio data in advance; When caching the maximum value of the data volume, the target cache depth is 1.2 times or more of the required cache depth.

7. The method for intelligently controlling a Bluetooth speaker according to claim 2, wherein: The transmission parameters are transmission parameters within a preset time period before the audio delay is greater than the delay threshold, including: signal strength, retransmission rate, and throughput fluctuation rate.

8. The method for intelligently controlling a Bluetooth speaker according to claim 1, wherein: The real-time acquisition of the audio delay of the Bluetooth speaker includes: Embed timestamps into audio data packets at the sending end of the audio file; Obtaining the local system time when decoding of the audio data packet starts, calculating the time difference between the time corresponding to the timestamp and the local system time when decoding of the audio data packet starts, to obtain an initial delay; Kalman filtering is performed on the initial delay to eliminate the influence of measurement noise and obtain a final audio delay.

9. A Bluetooth speaker intelligent control system, used to implement the Bluetooth speaker intelligent control method according to any one of claims 1 to 8, characterized in that: include: Detection module to obtain the audio delay of the Bluetooth speaker; An analysis module, configured to determine whether a playback state meets the requirements based on the audio delay; a data collection module, in response to the audio delay being greater than a delay threshold, determining that the playback state does not meet the requirements, and storing a duration of the playback state not meeting the requirements and transmission parameters when the playback state does not meet the requirements in a fault data set; A training module, configured to train a probability model based on the fault data set; An adjustment module is used to predict the probability that the playback status does not meet the requirements in the next cycle and the duration when the playback status does not meet the requirements based on the transmission parameters in the current cycle and the probability model. In response to the probability being greater than the preset probability, the cache depth is adjusted according to the duration. In response to the probability being less than or equal to the preset probability, the proportion of human voice in the audio file is obtained, and the speaker operating parameters are adjusted according to the human voice proportion.

10. A Bluetooth speaker, characterized in that: The method comprises a digital signal processor, a main control processor, a memory and a computer program stored in the memory and executable on the main control processor, wherein the computer program implements the steps of the method according to any one of claims 1 to 8 when executed by the processor.

Images