Real-time sound effect feedback adjusting control method and system for karaoke system

Abstract

The application belongs to the field of sound control, and discloses a karaoke system real-time sound effect feedback adjustment control method and system, acquires use information of various sound effect parameter combinations in multiple historical time periods of a user, contains corresponding singing total duration, use frequency, parameter adjustment amplitude, singing satisfaction score, calculates historical preference yield coefficients corresponding to each historical time period according to this, generates a historical preference yield coefficient sequence in sequence, simultaneously acquires current preference yield coefficients in multiple preset time periods in a current time period, generates a current preference yield coefficient sequence in sequence. Then, the yield mean value and the change slope corresponding to the two sequences are calculated respectively, the preference similarity corresponding to each time period is matched, the historical time period sound effect parameter combination corresponding to the maximum preference similarity is selected, and the karaoke system sound parameter is adjusted in this way, so that the home karaoke system sound effect is more in line with the personal preference of the user.

Description

Technical Field

[0001] This application relates to the field of audio control technology, and more specifically, to a real-time sound effect feedback adjustment and control method and system for karaoke systems. Background Technology

[0002] Most existing home karaoke systems are equipped with customizable sound effect settings, allowing users to adjust various sound parameters according to their needs to suit different karaoke scenarios, song styles, and the acoustic characteristics of the environment. These settings typically cover multiple sound effect adjustment options, such as reverberation intensity, vocal gain, bass effect, and treble effect. Some high-end home karaoke systems also have preset fixed sound effect modes such as pop, folk, and rock. Users can directly select preset modes or make secondary fine-tuning of parameters based on preset modes, aiming to provide users with more freedom in sound effect adjustment and meet diverse sound effect usage needs.

[0003] Existing home karaoke systems only provide standardized sound effect adjustment paths, requiring users to manually adjust parameters each time they use the system. They cannot automatically identify and match users' personal sound effect preferences based on their historical usage data, nor can they achieve differentiated sound effect adaptation for different users' usage habits. When the same home karaoke system is used by multiple users, the sound effect parameters need to be readjusted each time a user changes. This operation is cumbersome and difficult to accurately match the actual preferences of the current user. Some users who are not familiar with parameter adjustment operations will find it even more difficult to adjust to the sound effect that suits their preferences. The overall user experience is poor and cannot meet users' needs for personalized karaoke sound effects. Summary of the Invention

[0004] The purpose of this application is to provide a real-time sound effect feedback adjustment and control method and system for a karaoke system, which solves the technical problem that the sound effect settings of existing home karaoke systems cannot meet the user's personal preferences, and achieves the technical effect of making the sound effect settings of the home karaoke system more in line with the user's personal preferences.

[0005] In a first aspect, embodiments of this application provide a real-time sound effect feedback adjustment and control method for a karaoke system. The method includes: acquiring sound effect usage information of a user for each sound effect parameter combination over multiple historical time periods; determining the user's historical preference benefit coefficient for each sound effect parameter combination over multiple historical time periods based on the user's sound effect usage information over multiple historical time periods; arranging the user's historical preference benefit coefficients for each sound effect parameter combination over multiple historical time periods in chronological order as a sequence of historical preference benefit coefficients corresponding to the user over multiple historical time periods; acquiring the user's current preference benefit coefficients for multiple preset time periods within the current time period; arranging the user's current preference benefit coefficients for multiple preset time periods within the current time period in chronological order as a sequence of current preference benefit coefficients corresponding to the user over multiple preset time periods within the current time period; wherein, the sound effect usage information includes each sound effect parameter... The system considers the total performance duration, frequency of use, parameter adjustment range, and performance satisfaction score of the combined number of data points. The parameter adjustment range represents the extent to which the user manually fine-tunes the sound effect parameters. It also determines the mean historical preference return and the slope of historical preference change corresponding to the historical preference return coefficient sequences for multiple historical time periods. Furthermore, it determines the mean current preference return and the slope of current preference change for multiple preset time periods within the current preference return coefficient sequence. Based on the mean historical preference return and slope of historical preference change for multiple historical time periods, and the mean current preference return and slope of current preference change for multiple preset time periods within the current preference return coefficient sequence, it determines the preference similarity between multiple historical time periods and multiple preset time periods within the current preference return coefficient sequence. Finally, it determines the maximum preference similarity among the preference similarities among multiple preset time periods and adjusts the sound parameters of the karaoke system according to the sound effect parameter combination of the historical time period corresponding to the maximum preference similarity.

[0006] In one possible implementation, based on the user's audio effect usage information for each audio effect parameter combination over multiple historical time periods, the historical preference benefit coefficient for each audio effect parameter combination over multiple historical time periods is determined. This includes: obtaining the weights of total performance duration, frequency of use, parameter adjustment range, and performance satisfaction rating; and determining the sum of the products of total performance duration and total performance duration weight, frequency of use and frequency of use weight, parameter adjustment range and parameter adjustment range weight, and performance satisfaction rating and performance satisfaction rating weight corresponding to the audio effect usage information for each audio effect parameter combination, as the historical preference benefit coefficient for each audio effect parameter combination over multiple historical time periods.

[0007] In another possible implementation, the preference similarity between multiple historical time periods and multiple preset time periods within the current preference coefficient sequence is determined based on the historical preference return mean and historical preference change slope corresponding to the historical preference return coefficient sequences of multiple historical time periods, and the current preference return mean and current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence. This includes: determining the difference between the historical preference return mean and the current preference return mean corresponding to multiple preset time periods within the current preference return coefficient sequence, as the preference return difference between the historical preference return coefficient sequences of multiple historical time periods; determining the difference between the historical preference change slope and the current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence, as the preference return difference coefficient between the historical preference return coefficient sequences of multiple historical time periods; and determining the product of the preference return difference and the preference return difference coefficient between the historical preference return coefficient sequences of multiple historical time periods, as the preference similarity between the multiple historical time periods and the multiple preset time periods within the current preference return coefficient sequence.

[0008] In another possible implementation, the method further includes: arranging the preference similarities corresponding to multiple preset time periods from largest to smallest, and determining a preset number of target preference similarities with the highest ranking; determining the target sound effect parameter combinations corresponding to the preset number of target preference similarities; determining the target sound effect parameter vectors corresponding to the preset number of target sound effect parameter combinations; obtaining recommended sound effect parameter combinations for the song selected by the user; determining the recommended sound effect parameter vectors corresponding to the recommended sound effect parameter combinations; determining the similarity between the preset number of target sound effect parameter vectors and the recommended sound effect parameter vectors, as the recommendation coefficients corresponding to the preset number of target sound effect parameter combinations; recommending the preset number of target sound effect parameter combinations to the user in descending order of recommendation coefficients; obtaining the selected sound effect parameter combinations selected by the user; and adjusting the audio parameters of the karaoke system according to the selected sound effect parameter combinations.

[0009] In another possible implementation, obtaining the recommended sound effect parameter combination for the song selected by the user includes: obtaining the number of times the song selected by the user has been sung; when the number of times sung is greater than or equal to the preset number of times sung, obtaining multiple historical performance records of the song to be sung and the historical performance time and historical performance sound effect parameter combination corresponding to the multiple historical performance records; obtaining the performance time weight corresponding to the historical performance time of the multiple historical performance records; and weighting and normalizing the historical performance sound effect parameter combination corresponding to the multiple historical performance records according to the performance time weight, as the recommended sound effect parameter combination for the song selected by the user.

[0010] In another possible implementation, the singing time weights corresponding to the historical singing times of multiple historical singing records are obtained, including: determining the time difference between the historical singing time and the current time corresponding to multiple historical singing records as the singing time distances corresponding to multiple historical singing records; determining and normalizing the ratio of the singing time distances corresponding to multiple historical singing records as the singing time weights corresponding to multiple historical singing records.

[0011] In another possible implementation, the method further includes: performing multi-dimensional style semantic annotation on multiple songs in the music library, as the multi-dimensional style semantics of the multiple songs; wherein, the multi-dimensional style semantics includes song rhythm semantics, song emotional semantics, instrument arrangement semantics, and singing difficulty semantics; obtaining the target multi-dimensional style semantics corresponding to the new song selected by the user to sing; determining the similarity between the multi-dimensional style semantics of the multiple songs and the target multi-dimensional style semantics, as the multi-dimensional similarity of the multiple songs; determining the style preference benefit coefficient corresponding to the multi-dimensional similarity of the multiple songs; determining the maximum style preference benefit coefficient among the style preference benefit coefficients of the multiple songs, obtaining the target sound effect parameter combination corresponding to the maximum style preference benefit coefficient; and adjusting the audio parameters of the karaoke system according to the target sound effect parameter combination.

[0012] In another possible implementation, the similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics is determined as the multi-dimensional similarity of multiple songs. This includes: obtaining the multi-dimensional weights corresponding to the multi-dimensional style semantics set by the user; determining the dimensional similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics respectively; and determining the sum of the products of the multi-dimensional similarity and the dimensional weights of multiple songs as the similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics.

[0013] In another possible implementation, determining the style preference benefit coefficient corresponding to the multi-dimensional similarity of multiple songs includes: obtaining the user's historical performance data and sound effect parameter combinations for multiple songs; determining the total performance duration, usage frequency, parameter adjustment range, and performance satisfaction score corresponding to different sound effect parameter combinations under the multi-dimensional style semantics based on the multi-dimensional style semantics of multiple songs and the user's historical performance data for multiple songs; determining the historical preference benefit coefficient corresponding to different sound effect parameter combinations under the multi-dimensional style semantics based on the total performance duration, usage frequency, parameter adjustment range, and performance satisfaction score corresponding to different sound effect parameter combinations under the multi-dimensional style semantics of multiple songs; and determining the product of the multi-dimensional similarity of multiple songs and the historical preference benefit coefficient as the style preference benefit coefficient corresponding to the multi-dimensional similarity of multiple songs.

[0014] Secondly, embodiments of this application provide a real-time sound effect feedback adjustment and control system for a karaoke system, including units for implementing the above-described method.

[0015] The beneficial effects of the embodiments in this application compared with the prior art are:

[0016] This application provides a real-time sound effect feedback adjustment and control method for a karaoke system. The method includes: acquiring sound effect usage information of a user for each sound effect parameter combination over multiple historical time periods; determining the historical preference benefit coefficient of the user for each sound effect parameter combination over multiple historical time periods based on the user's sound effect usage information over multiple historical time periods; arranging the historical preference benefit coefficients of the user for each sound effect parameter combination over multiple historical time periods in chronological order as the historical preference benefit coefficient sequence corresponding to the user over multiple historical time periods; acquiring the current preference benefit coefficient of the user in multiple preset time periods within the current time period; arranging the current preference benefit coefficients of the user in multiple preset time periods within the current time period in chronological order as the current preference benefit coefficient sequence corresponding to the user in multiple preset time periods within the current time period; wherein, the sound effect usage information includes the total singing duration, usage frequency, parameter adjustment range, and singing satisfaction score for each sound effect parameter combination, and the parameter adjustment range characterizes the extent to which the user manually fine-tunes the sound effect parameters; This paper describes a method for determining the historical preference return mean and historical preference change slope corresponding to historical preference return coefficient sequences across multiple historical time periods. It also involves determining the current preference return mean and current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence. Based on these historical preference return mean and historical preference change slopes, and considering the current preference return mean and current preference change slope within the current preference return coefficient sequence, the method determines the preference similarity between multiple historical time periods and multiple preset time periods within the current preference return coefficient sequence. Finally, it determines the maximum preference similarity among the multiple preset time periods and adjusts the sound parameters of the karaoke system according to the sound effect parameter combination of the historical time period corresponding to the maximum preference similarity. In this embodiment, the method calculates the historical preference return mean and historical preference change slope of the historical preference return coefficient sequence, as well as the current preference return mean and current preference change slope of the current preference return coefficient sequence. By calculating preference similarity based on these two types of values, the method can accurately capture the changing trend of user sound effect preferences, avoid adjustment deviations caused by referring only to single-use data, improve the accuracy of sound effect matching, and make the home karaoke system more in line with the user's personal preferences. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the first real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0019] Figure 2 A schematic diagram illustrating the workflow of the first real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0020] Figure 3 A flowchart illustrating the second real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0021] Figure 4 A schematic diagram illustrating the workflow of the second real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0022] Figure 5 A flowchart illustrating the third real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0023] Figure 6 A schematic diagram illustrating the workflow of the third real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0024] Figure 7 A flowchart illustrating the fourth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0025] Figure 8 A flowchart illustrating the fifth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0026] Figure 9 A schematic diagram illustrating the workflow of the fifth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0027] Figure 10 A flowchart illustrating the sixth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0028] Figure 11 A schematic diagram illustrating the workflow of the sixth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0029] Figure 12 A flowchart illustrating the seventh real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0030] Figure 13 A schematic diagram illustrating the workflow of the seventh real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment;

[0031] Figure 14 This is a schematic diagram of the logic structure of a real-time sound effect feedback adjustment and control system for a karaoke system, provided in an embodiment of this application. Detailed Implementation

[0032] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0033] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0034] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0035] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0036] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0037] Existing home karaoke systems often fail to match users' preferred sound effects, resulting in a poor overall user experience and failing to meet users' needs for personalized karaoke sound effects.

[0038] Based on the above reasons, this application provides a real-time sound effect feedback adjustment and control method for a karaoke system. The method includes: acquiring sound effect usage information of a user for each sound effect parameter combination over multiple historical time periods; determining the historical preference benefit coefficient of the user for each sound effect parameter combination over multiple historical time periods based on the user's sound effect usage information over multiple historical time periods; arranging the historical preference benefit coefficients of the user for each sound effect parameter combination over multiple historical time periods in chronological order as the historical preference benefit coefficient sequence corresponding to the user over multiple historical time periods; acquiring the current preference benefit coefficient of the user in multiple preset time periods within the current time period; arranging the current preference benefit coefficients of the user in multiple preset time periods within the current time period in chronological order as the current preference benefit coefficient sequence corresponding to the user in multiple preset time periods within the current time period; wherein, the sound effect usage information includes the total singing duration, usage frequency, parameter adjustment range, and singing satisfaction score for each sound effect parameter combination, and the parameter adjustment range characterizes the extent to which the user manually fine-tunes the sound effect parameters; This paper describes a method for determining the historical preference return mean and historical preference change slope corresponding to historical preference return coefficient sequences across multiple historical time periods. It also involves determining the current preference return mean and current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence. Based on these historical preference return mean and historical preference change slopes, and considering the current preference return mean and current preference change slope within the current preference return coefficient sequence, the method determines the preference similarity between multiple historical time periods and multiple preset time periods within the current preference return coefficient sequence. Finally, it determines the maximum preference similarity among the multiple preset time periods and adjusts the sound parameters of the karaoke system according to the sound effect parameter combination of the historical time period corresponding to the maximum preference similarity. In this embodiment, the method calculates the historical preference return mean and historical preference change slope of the historical preference return coefficient sequence, as well as the current preference return mean and current preference change slope of the current preference return coefficient sequence. By calculating preference similarity based on these two types of values, the method can accurately capture the changing trend of user sound effect preferences, avoid adjustment deviations caused by referring only to single-use data, improve the accuracy of sound effect matching, and make the home karaoke system more in line with the user's personal preferences.

[0039] In some scenarios, the real-time sound effect feedback adjustment and control method for a karaoke system according to an embodiment of this application can be applied to home karaoke rooms. Users do not need to manually adjust the sound effects when singing karaoke. The system automatically matches the parameter combinations that users have previously preferred, and quickly adapts to the singing sound effect needs of different users.

[0040] In other scenarios, the real-time sound effect feedback adjustment and control method for a karaoke system according to the embodiments of this application can also be applied to online live karaoke scenarios. When the host switches singing styles, the system can match the corresponding preferred sound effect parameters in real time to ensure the sound effect of the real-time singing and the listening experience of the audience.

[0041] The following describes in detail, with specific examples, a real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment.

[0042] Figure 1 A flowchart illustrating the first real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 1 As shown in the embodiment of this application, a real-time sound effect feedback adjustment and control method for a karaoke system is also provided. The method further includes steps S110 to S120, which will be described in detail below.

[0043] S110. Obtain sound effect usage information for each sound effect parameter combination from multiple historical time periods. Based on the user's sound effect usage information for each sound effect parameter combination from multiple historical time periods, determine the user's historical preference benefit coefficient for each sound effect parameter combination from multiple historical time periods. Arrange the user's historical preference benefit coefficients for each sound effect parameter combination from multiple historical time periods in chronological order, as the user's corresponding historical preference benefit coefficient sequence from multiple historical time periods. Obtain the user's current preference benefit coefficients for multiple preset time periods within the current time period. Arrange the user's current preference benefit coefficients for multiple preset time periods within the current time period in chronological order, as the user's corresponding current preference benefit coefficient sequence from multiple preset time periods within the current time period. The sound effect usage information includes the total performance duration, usage frequency, parameter adjustment range, and performance satisfaction score for each sound effect parameter combination. The parameter adjustment range represents the extent to which the user manually fine-tunes the sound effect parameters.

[0044] In existing technologies, preferences are usually determined based solely on feedback from a single use of sound effects. This approach does not combine historical usage data over multiple time periods with dynamic adjustment of preference coefficients based on current usage. Consequently, it fails to accurately match users' real-time personalized needs for sound effect parameter combinations and cannot capture the changing trends of user preferences over time.

[0045] Figure 2 A schematic diagram of the workflow of the first real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 2 As shown in this implementation, the usage data of each sound effect parameter combination in different time periods can be extracted from the user's history of using the audio karaoke system by dividing it into time intervals. This data fully covers the user's actual usage behavior of different sound effects.

[0046] For example, when a user sings a song using a speaker karaoke system, the system can automatically record all sound effects usage and store the corresponding data according to pre-defined time intervals.

[0047] In this implementation, multiple historical time periods can be divided according to the user's usage cycle to adapt to the statistical needs of different scenarios.

[0048] For example, the historical period can be divided by calendar month, with each calendar month of the past 12 months as an independent historical period. Alternatively, it can be divided by user usage frequency, with each 10 hours of accumulated use of the audio karaoke system as a historical period. Or it can be divided by quarter, with each of the past 4 quarters as a historical period.

[0049] In this implementation, the historical preference benefit coefficient of each sound effect parameter combination in the corresponding historical time period can be calculated based on the usage information of each sound effect parameter combination extracted in each historical time period. The benefit coefficient can directly reflect the user's preference for the combination in the corresponding time period.

[0050] For example, after obtaining all usage information for three different combinations of sound effect parameters within a certain historical period, the corresponding historical preference benefit coefficient can be calculated for each combination of sound effect parameters.

[0051] In this implementation, the sound effect parameter combination consists of multiple sound effect parameters such as reverberation intensity, equalization level, reverberation type, and gain. Common types are set around different singing scenarios.

[0052] For example, the sound effect parameter combinations can be divided into KTV scene sound effect parameter combinations, recording studio scene sound effect parameter combinations, concert hall scene sound effect parameter combinations, open valley scene sound effect parameter combinations, and blues jazz scene sound effect parameter combinations. Each type corresponds to a fixed set of parameters such as reverberation intensity and equalization level.

[0053] In this implementation, the historical preference return coefficients corresponding to all historical time periods can be arranged sequentially in order of time from far to near or from near to far, generating a sequence that can be used for subsequent preference analysis.

[0054] For example, the historical preference benefit coefficients for each sound effect parameter combination in the past 12 calendar months are arranged sequentially from January to December to obtain the historical preference benefit coefficient sequence for that sound effect parameter combination.

[0055] In this implementation, the sound effect usage information includes statistical data in four dimensions, each with a corresponding unified statistical format, which can fully quantify the user's usage behavior.

[0056] For example, the total singing time is calculated as the total number of minutes the user sings using the sound effect parameter combination within the corresponding time period, rounded to one decimal place.

[0057] For example, the frequency of use is statistically represented as the percentage of the number of times a user uses the combination of sound effect parameters within the corresponding time period relative to the total number of times a user sings, presented as a percentage.

[0058] For example, the statistical form of the parameter adjustment range is the sum of the absolute values ​​of the changes in each parameter manually fine-tuned by the user, with the unit being gear levels, and each change in a gear level is recorded as a value of 1.

[0059] For example, the statistical form of the singing satisfaction rating is the average of the 1-star to 5-star ratings given by users after singing using the combination of sound effect parameters.

[0060] In this implementation, the sound effect usage information of the four dimensions can be normalized separately and then weighted and summed to obtain the final historical preference benefit coefficient. The weights of the four dimensions can be set in advance based on experience.

[0061] For example, within a certain historical time period, the total performance duration of a certain KTV scene's sound effect parameter combination is 60 minutes, with a normalized value of 0.8. The usage frequency is 20%, with a normalized value of 0.6. The parameter adjustment range is 1 level, with a normalized value of 0.9. The average performance satisfaction rating is 4.5 stars, with a normalized value of 0.9. Pre-setting the weights of the four dimensions as 0.2, 0.3, 0.2, and 0.3, the historical preference benefit coefficient can be calculated as: 0.8 × 0.2 + 0.6 × 0.3 + 0.9 × 0.2 + 0.9 × 0.3 = 0.79.

[0062] In this implementation, the historical preference return coefficient sequence is arranged in chronological order, with each position corresponding to the return coefficient of a historical time period. The arrangement clearly reflects the process of preference changes.

[0063] For example, for the sound effect parameter combination of a home KTV, four historical time periods were divided from January 2024 to April 2024. The historical preference return coefficients for the four time periods were calculated to be 0.62, 0.71, 0.75, and 0.83, respectively. After arranging them in chronological order from January to April, the historical preference return coefficient sequence is [0.62, 0.71, 0.75, 0.83].

[0064] In this implementation, the current statistical period can be further divided into multiple consecutive short time periods. The current preference benefit coefficient for each combination of sound effect parameters in each short time period can be extracted to capture recent changes in user preferences.

[0065] For example, if the current statistical period is the current month, multiple consecutive preset time periods can be divided within the current month, and the preference benefit coefficient for each time period can be extracted.

[0066] In this implementation, multiple preset time periods within the current time period can be divided into equal durations to meet the needs of real-time preference statistics.

[0067] For example, if the current time period is a calendar month, the calendar month can be divided into 4 preset time periods by week, with each week being an independent preset time period. Alternatively, the current calendar month can be divided into preset time periods of 5 days each, resulting in a total of 6 consecutive preset time periods.

[0068] In this implementation, the current preference benefit coefficients of multiple preset time periods extracted in the current time period can be arranged in chronological order to generate a current preference benefit coefficient sequence that can be used to compare historical trends.

[0069] For example, the current preference benefit coefficients for a certain sound effect parameter combination in each of the four weeks of the month can be arranged in the order of the first to the fourth week to obtain the current preference benefit coefficient sequence for that combination.

[0070] S120. Determine the mean historical preference return and the slope of historical preference change corresponding to the historical preference return coefficient sequences of multiple historical time periods. Determine the mean current preference return and the slope of current preference change corresponding to multiple preset time periods within the current preference return coefficient sequence. Based on the mean historical preference return and slope of historical preference change corresponding to the historical preference return coefficient sequences of multiple historical time periods, and the mean current preference return and slope of current preference change corresponding to multiple preset time periods within the current preference return coefficient sequence, determine the preference similarity corresponding to the multiple historical time periods and the multiple preset time periods within the current preference return coefficient sequence. Determine the maximum preference similarity among the preference similarities corresponding to the multiple preset time periods, and adjust the audio parameters of the karaoke system according to the audio effect parameter combination of the historical time period corresponding to the maximum preference similarity.

[0071] In this implementation, for a sequence of historical preference return coefficients that has been arranged in chronological order for multiple historical time periods, the average value of all historical preference return coefficients in the sequence can be calculated, and the overall slope of the sequence can be obtained through linear fitting.

[0072] For example, for a 12-month historical preference return coefficient sequence corresponding to a certain sound effect parameter combination, the 12 historical preference return coefficients can be added together and divided by 12 to obtain the average historical preference return, and then the slope of historical preference change can be obtained by linear fitting using the least squares method.

[0073] In this implementation, for the current preference return coefficient sequence, that is, the return coefficient sequence obtained by arranging multiple preset time periods within the current time period, the average value of all current preference return coefficients in the sequence can be calculated, and the overall change slope of the sequence can be obtained by linear fitting.

[0074] For example, for the current preference return coefficient sequence of four preset time periods obtained by dividing the current month into weeks, the four current preference return coefficients can be added together and divided by 4 to obtain the mean of the current preference return. Then, the slope of the current preference change can be obtained by linear fitting using the least squares method.

[0075] In this implementation, the similarity between the historical time period and the current preset time period can be calculated by combining the calculated historical average preference return and historical preference change slope, as well as the current average preference return and current preference change slope. The higher the similarity, the closer the user's current preference is to the preference of the corresponding historical time period.

[0076] For example, the weight of the historical preference return mean is set to 0.6, and the weight of the historical preference change slope is set to 0.4. The historical feature value is calculated as 0.6 × historical preference return mean + 0.4 × historical preference change slope. Similarly, the current feature value is calculated as 0.6 × current preference return mean + 0.4 × current preference change slope. The preference similarity is 1 - |historical feature value - current feature value|. Given that the current historical preference return mean is 0.75 and the historical preference change slope is 0.05, the current preference return mean is 0.72 and the current preference change slope is 0.04, the preference similarity is calculated as 1 - |(0.6 × 0.75 + 0.4 × 0.05) - (0.6 × 0.72 + 0.4 × 0.04)| = 0.99.

[0077] In this implementation, it is possible to iterate through all preference similarities corresponding to the current preset time period and filter out the preference similarity with the largest value. This preference similarity corresponds to the historical time period in which the user's preference is most matched.

[0078] For example, the current time period is divided into four preset time periods. Each preset time period corresponds to four different historical time periods with four preference similarities. All sixteen preference similarities can be iterated through to select the preference similarity with the largest value.

[0079] In this implementation, the combination of sound effect parameters used in the historical time period corresponding to the maximum preference similarity can be extracted, and all parameters in the combination can be directly written into the karaoke system to complete the adjustment of the sound parameters of the karaoke system.

[0080] For example, the system filters out the concert hall sound effect parameter combination used by the user in the historical time period of March 2024, corresponding to the user with the highest preference similarity. The parameters of this combination are: reverberation intensity level 3, equalization level pop level 2, and gain +1dB. After receiving this parameter combination, the karaoke system directly replaces the original audio parameters with the corresponding parameters of this combination to complete the adjustment. The user can then directly use this parameter combination to output the sound effect when singing.

[0081] This implementation method collects the total singing time, frequency of use, parameter adjustment range, and singing satisfaction rating of each sound effect parameter combination for users in multiple historical time periods. It calculates the corresponding historical preference benefit coefficient and generates a sequence. By matching the similarity of the current preference benefit coefficient sequence, it finds the sound effect parameter combination that best matches the user's past preferences. This eliminates the need for users to manually adjust repeatedly, reducing the cost of sound effect adjustment during karaoke.

[0082] This implementation calculates the historical preference return mean and historical preference change slope of the historical preference return coefficient sequence, as well as the current preference return mean and current preference change slope of the current preference return coefficient sequence. Based on these two types of values, preference similarity is calculated, which can accurately capture the changing trend of user sound effect preferences, avoid adjustment deviations caused by referring only to single-use data, and improve the accuracy of sound effect matching. By selecting the sound effect parameter combination corresponding to the historical time period with the maximum preference similarity to adjust the audio parameters of the karaoke system, it can adapt to the user's current sound effect usage preferences in real time, quickly complete sound effect adjustments during the user's singing process, and not interrupt the singing process, thereby improving the smoothness and auditory experience of the karaoke process.

[0083] In some implementations, in S110 above, the historical preference benefit coefficient of the user for each combination of sound effect parameters in multiple historical time periods is determined based on the user's sound effect usage information for each combination of sound effect parameters in multiple historical time periods, including S111 to S112. S111 to S112 are explained in detail below.

[0084] S111: Obtain the weights of total performance duration, frequency of use, parameter adjustment range, and performance satisfaction rating.

[0085] In this implementation, the weights of the four dimensions corresponding to the historical preference benefit coefficient can be obtained. The four dimensions correspond to the four statistical items in the sound effect usage information. After obtaining them, they are stored as the parameter values ​​required for calculation.

[0086] In this implementation, the four weights are obtained from various sources, and the appropriate source can be selected according to the actual application scenario.

[0087] For example, the four weights can come from the system's default configuration. For newly registered users who have not accumulated usage data, the default weight reorganization pre-set by the karaoke app can be directly called to obtain the corresponding four weight values.

[0088] For example, the four weights can come from user-defined configurations. Users can manually adjust the weight values ​​of the four dimensions in the sound effect preference settings interface of the karaoke app, and then obtain them directly from the user configuration data after saving.

[0089] S112. Determine the sum of the following for each sound effect parameter combination: the product of the total singing time and the weight of the total singing time, the product of the usage frequency and the weight of the usage frequency, the product of the parameter adjustment range and the weight of the parameter adjustment range, and the product of the singing satisfaction score and the weight of the singing satisfaction score. This sum is used as the historical preference benefit coefficient for each sound effect parameter combination in multiple historical time periods.

[0090] In this implementation, after the normalized sound effect usage information extraction and the acquisition of the four corresponding weights are completed, the statistical value of each dimension and the product of the corresponding weights are calculated respectively, providing basic data for subsequent summation calculations.

[0091] For example, for the four dimensions of usage information of a certain sound effect parameter combination within a certain historical period, the corresponding normalized values ​​and the four weights can be extracted, and the four products can be calculated in sequence.

[0092] In this implementation, the corresponding product calculation can be completed according to the four dimensions, and each dimension adopts the calculation method of multiplying the statistical value by the corresponding weight.

[0093] For example, the existing normalized total performance duration value is 0.8, and the corresponding total performance duration weight is 0.2. The product of the two is 0.8 × 0.2 = 0.16.

[0094] For example, the existing normalized usage frequency value is 0.6, the corresponding usage frequency weight is 0.3, and the product of the two is 0.6 × 0.3 = 0.18.

[0095] For example, the existing normalized parameter adjustment range value is 0.9, the corresponding parameter adjustment range weight is 0.2, and the product of the two is 0.9 × 0.2 = 0.18.

[0096] For example, the existing normalized performance satisfaction score is 0.9, and the corresponding performance satisfaction score weight is 0.3. The product of the two is 0.9 × 0.3 = 0.27.

[0097] In this implementation, the products of the four dimensions that have already been calculated can be added together, and the sum can be directly used as the historical preference benefit coefficient of the sound effect parameter combination in the corresponding historical time period.

[0098] For example, taking the results of the above four products, 0.16, 0.18, 0.18, and 0.27 are added together in sequence to get 0.16+0.18+0.18+0.27=0.79. Finally, the historical preference benefit coefficient corresponding to this sound effect parameter combination is 0.79.

[0099] This implementation method obtains the weights of total performance duration, frequency of use, parameter adjustment range, and performance satisfaction rating. The sum of the products of the four types of usage information and their corresponding weights yields the historical preference benefit coefficient, which can cover users' multi-dimensional feedback on sound effects, avoid the bias in preference judgment caused by a single evaluation dimension, and improve the comprehensiveness of preference judgment.

[0100] This implementation assigns corresponding weights to total performance duration, usage frequency, parameter adjustment range, and performance satisfaction score. The influence ratio of each dimension can be flexibly adjusted according to the emphasis of different scenarios, adapting to the preference judgment needs of different user groups and improving the scenario adaptability of preference benefit coefficient calculation. The historical preference benefit coefficient is obtained by multiplying and summing the quantified usage information and weights, transforming the user's vague sound effect preferences into calculable numerical indicators, reducing the computational complexity of subsequent preference similarity matching, and improving the computational efficiency of the entire sound effect adjustment process.

[0101] Figure 3 This is a flowchart illustrating the second real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment, as shown below. Figure 3 As shown, in some implementations, in S120 above, the preference similarity corresponding to multiple historical time periods and multiple preset time periods in the current preference coefficient sequence is determined based on the historical preference return mean and historical preference change slope corresponding to the historical preference return coefficient sequence of multiple historical time periods and the current preference return mean and current preference change slope corresponding to multiple preset time periods in the current preference return coefficient sequence, including S121 to S122. S121 to S122 are explained in detail below.

[0102] S121. Determine the difference between the historical preference return mean corresponding to the historical preference return coefficient sequence for multiple historical time periods and the current preference return mean corresponding to multiple preset time periods within the current preference return coefficient sequence, and use this difference as the preference return difference coefficient corresponding to the historical preference return coefficient sequence for multiple historical time periods. Determine the difference between the historical preference change slope corresponding to the historical preference return coefficient sequence for multiple historical time periods and the current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence, and use this difference as the preference return difference coefficient corresponding to the historical preference return coefficient sequence for multiple historical time periods.

[0103] Figure 4 This is a schematic diagram of the workflow of the second real-time sound effect feedback adjustment and control method for a karaoke system provided in the embodiments of this application, as shown below. Figure 4 As shown, in this implementation, for each set of historical preference return coefficient sequences and the current preference return coefficient sequences of the current preset time period, the calculated historical preference return mean and the current preference return mean can be extracted respectively, the difference between the two can be calculated, and the difference can be used as the preference return difference of the corresponding combination.

[0104] For example, given a set of historical preference return coefficient sequences and a current preference return coefficient sequence, the difference between the two mean values ​​can be obtained by directly extracting the two mean values.

[0105] For example, the average historical preference return corresponding to the historical preference return coefficient sequence of a certain historical period is 0.75, and the average current preference return corresponding to the current preference return coefficient sequence of a certain current preset period is 0.72. The difference is calculated as 0.75-0.72=0.03, and the final preference return difference is 0.03.

[0106] In this implementation, for each set of historical preference return coefficient sequences and the current preference return coefficient sequences of the current preset time period, the calculated historical preference change slope and the current preference change slope can be extracted respectively, the difference between the two can be calculated, and the difference can be used as the preference return difference coefficient of the corresponding combination.

[0107] For example, given a set of historical preference return coefficient sequences and a current preference return coefficient sequence, the two slope values ​​can be extracted directly and the difference can be calculated to obtain the corresponding preference return difference coefficient.

[0108] In this implementation, when calculating the difference between the historical preference change slope and the current preference change slope, the result is obtained by directly subtracting the current preference change slope from the historical preference change slope.

[0109] For example, the slope of the historical preference change corresponding to the historical preference return coefficient sequence of a certain historical period is 0.05, and the slope of the current preference change corresponding to the current preference return coefficient sequence of a certain current preset period is 0.04. The difference is calculated to be 0.05-0.04=0.01, and the final preference return difference coefficient is 0.01.

[0110] S122. Determine the product of the preference return difference and the preference return difference coefficient corresponding to the historical preference return coefficient sequence of multiple historical time periods, and use it as the preference similarity corresponding to multiple preset time periods within the multiple historical time periods and the current preference return coefficient sequence.

[0111] In this implementation, the product of the already calculated preference benefit difference and preference benefit difference coefficient can be calculated, and the calculated product result can be used as the preference similarity between the corresponding historical time period and the current preset time period combination.

[0112] For example, for a certain historical time period and a certain current preset time period, the obtained preference payoff difference and preference payoff difference coefficient are extracted, and the product of the two is directly calculated to obtain the corresponding preference similarity.

[0113] In this implementation, the preference similarity can be obtained by directly multiplying the two obtained values, and the calculation logic is clear.

[0114] For example, following the previous calculation results, the preference payoff difference for the corresponding combination is 0.03, and the preference payoff difference coefficient is 0.01. The product of these two is 0.03 × 0.01 = 0.0003, resulting in a preference similarity of 0.0003 for this combination. The closer the preference payoff difference and the preference payoff difference coefficient are to 0, and the smaller the product, the higher the corresponding preference similarity.

[0115] This implementation calculates the difference between the historical average preference return and the current average preference return to obtain the preference return difference value, and calculates the difference between the historical slope of preference change and the current slope of preference change to obtain the preference return difference coefficient. Combining the two types of differences, preference similarity is calculated, which covers both the overall level and trend of preference changes, thereby improving the accuracy of preference matching.

[0116] This implementation uses the product of the preference gain difference and the preference gain difference coefficient as the preference similarity, transforming preference differences into a quantifiable single value. This simplifies the similarity comparison process, reduces the computational complexity of subsequent selection of the maximum preference similarity, and improves computational efficiency. At the same time, it considers the differences in gain levels and the differences in change trends when calculating similarity, avoiding matching bias caused by referring to only a single dimension. Even if there are small fluctuations in short-term preferences, it can still match historical sound effect parameter combinations that meet the user's actual needs, improving the reliability of sound effect adjustment.

[0117] Figure 5 A flowchart illustrating the third real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 5 As shown, in some implementations, the above method also includes S130 to S140, which will be described in detail below.

[0118] S130. Sort the preference similarity corresponding to multiple preset time periods from largest to smallest, and determine the top preset number of target preference similarities. Then, determine the target sound effect parameter combinations corresponding to the preset number of target preference similarities. Finally, determine the target sound effect parameter vector corresponding to the preset number of target sound effect parameter combinations.

[0119] Figure 6 A schematic diagram illustrating the workflow of the third real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 6 As shown, in this implementation, the preference similarities corresponding to all current preset time periods can be sorted in descending order of value, the target preference similarities that are ranked first can be selected according to a preset number, and then the target sound effect parameter combination corresponding to each target preference similarity can be extracted.

[0120] For example, there are currently 8 preference similarities corresponding to 8 preset time periods. The top 3 preference similarities can be selected as the target preference similarities according to the sorting results, and then the corresponding 3 target sound effect parameter combinations can be extracted.

[0121] In this implementation, the preset quantity can be determined according to different rules, and the corresponding quantity of preference similarity can be selected directly according to the sorting results during the filtering process.

[0122] For example, the preset number can be fixed at 3, or it can be determined by rounding down to 1 / 3 of the total number of preference similarities. If there are a total of 10 preference similarities, the preset number is 3. After sorting, the top N preference similarities (N is the preset number) are directly selected as the target preference similarities.

[0123] In this implementation, the target sound effect parameter combination corresponding to each target preference similarity can be directly extracted by binding the preference similarity and the sound effect parameter combination.

[0124] For example, after calculating each preference similarity, the preference similarity is directly bound and stored with the corresponding historical time period's stored sound effect parameter combination. After filtering to obtain three target preference similarities, the sound effect parameter combination bound to each target preference similarity is read sequentially to obtain the three corresponding target sound effect parameter combinations.

[0125] In this implementation, the extracted multiple target sound effect parameter combinations can be converted into corresponding target sound effect parameter vectors, which facilitates subsequent calculation and processing.

[0126] For example, all parameters in each target sound effect parameter combination can be arranged in a fixed order to generate a one-dimensional target sound effect parameter vector.

[0127] In this implementation, the target sound effect parameter vector arranges the values ​​of all sound effect parameters in a preset fixed order to form a one-dimensional vector.

[0128] For example, with the fixed sorting order being reverberation intensity, equalization level, gain, and reverberation type encoding, the parameters of a certain target sound effect parameter combination are reverberation intensity 3, equalization level 2, gain 1, and reverberation type encoding 2, and the corresponding target sound effect parameter vector is [3, 2, 1, 2].

[0129] S140. Obtain the recommended sound effect parameter combination for the song selected by the user. Determine the recommended sound effect parameter vector corresponding to the recommended sound effect parameter combination. Determine the similarity between a preset number of target sound effect parameter vectors and recommended sound effect parameter vectors, using this as the recommendation coefficient for the preset number of target sound effect parameter combinations. Recommend the preset number of target sound effect parameter combinations to the user in descending order of recommendation coefficient. Obtain the selected sound effect parameter combination chosen by the user. Adjust the audio parameters of the karaoke system according to the selected sound effect parameter combination.

[0130] In this implementation, after the user selects a song to sing, a pre-matched recommended combination of sound effect parameters for that song can be obtained, and sound effect options that better suit the user's needs can be selected based on the user's personal preferences.

[0131] In this implementation, the sources for the recommended sound effect parameter combinations are divided into several different types, which can be selected according to the scene.

[0132] For example, it can be obtained from a matching library of song styles and sound effects, and the corresponding recommended sound effect parameter combination can be matched according to the genre of the song to be sung, and popular songs can be matched with popular sound effect combinations.

[0133] For example, the most frequently used sound effect parameter combinations for a song can be extracted from cloud-based trending data and used by all users as recommended combinations.

[0134] For example, the sound effect parameters used in the original version of the song can be extracted and compiled into a recommended combination.

[0135] In this implementation, the obtained recommended sound effect parameter combinations can be arranged in the same parameter order as the target sound effect parameter vector to generate the corresponding recommended sound effect parameter vector.

[0136] For example, by referencing the order of the target sound effect parameter vector, the parameters in the recommended sound effect parameter combination are arranged in the same order to obtain a recommended sound effect parameter vector with the same corresponding dimension.

[0137] In this implementation, the cosine similarity between each target sound effect parameter vector and the recommended sound effect parameter vector can be calculated, and the calculated similarity can be used as the recommendation coefficient for the corresponding target sound effect parameter combination.

[0138] For example, for three preset number of target sound effect parameter vectors, the similarity is calculated with the recommended sound effect parameter vectors in turn to obtain three corresponding recommendation coefficients.

[0139] In this implementation, cosine similarity is used to calculate the similarity between two vectors of the same dimension, and the range of the calculation result is adapted to the requirements of the recommendation coefficient.

[0140] For example, the existing target sound effect parameter vector is [3, 2, 1, 2], the recommended sound effect parameter vector is [3, 2, 2, 1], the cosine similarity between the two vectors is 0.93, and the recommendation coefficient of the target sound effect parameter combination is 0.93.

[0141] In this implementation, all preset combinations of target sound effect parameters can be sorted from high to low according to their corresponding recommendation coefficients, and the sorted combinations can be displayed to the user in order to complete the recommendation process.

[0142] For example, the three target sound effect parameters can be sorted by recommendation coefficient from largest to smallest and then displayed sequentially in the sound effect recommendation section of the app.

[0143] In this implementation, the recommendation results can be displayed in a list format, with each option labeled with the corresponding sound effect name and matching degree.

[0144] For example, in the sound effect selection interface of the karaoke system, a personalized recommendation area is set up, which displays the target sound effect parameter combinations in descending order of recommendation coefficient. The combinations are arranged from top to bottom as follows: concert hall sound effect (matching degree 93%), KTV sound effect (matching degree 85%), and recording studio sound effect (matching degree 72%). Users can intuitively view the sorted recommendation results.

[0145] In this implementation, the user's click operation in the recommendation list can be detected to obtain the final selected sound effect parameter combination.

[0146] For example, when a user clicks on a sound effect option in the recommended list, the karaoke system directly identifies the sound effect parameter combination corresponding to that click and marks it as the selected sound effect parameter combination.

[0147] In this implementation, the user completes the selection through click interaction, and the system directly reads the parameter combination bound to the corresponding option.

[0148] For example, when a user clicks on the concert hall sound effect option that is ranked first in the recommended list, the system reads the target sound effect parameter combination that is pre-bound to the option, marks the combination as the selected sound effect parameter combination, and completes the acquisition process.

[0149] In this implementation, all audio parameter values ​​in the selected audio effect parameter combination can be read, and the current audio parameters of the karaoke system can be replaced with the parameter values ​​corresponding to the selected audio effect parameter combination to complete the adjustment.

[0150] For example, it can read parameters such as reverberation intensity and equalization level of the selected sound effect parameter combination, directly replace the original parameters of the karaoke system, and allow users to use the adjusted parameters to output sound effects when singing.

[0151] This implementation method filters out the top-ranked preset number of target sound effect parameter combinations corresponding to target preference similarity, and calculates the recommendation coefficient by combining the recommended sound effect parameter combinations of the song to be sung by the user. It combines user preferences and song adaptation requirements to avoid the recommended sound effects only matching user habits but not matching the song style, thereby improving the adaptability of sound effect recommendations.

[0152] This implementation calculates the similarity between the target sound effect parameter vector and the recommended sound effect parameter vector to obtain the recommendation coefficient. It transforms the sound effect parameters into quantifiable and comparable vector indicators, enabling rapid ranking of the suitability of multiple sound effects, improving the computational efficiency of sound effect recommendations, and reducing user waiting time. Options are pushed to users from high to low according to the recommendation coefficient, allowing users to independently select the desired combination of sound effect parameters for adjustment. This reduces the tedious steps of manual parameter adjustment for users, respects users' autonomy in making choices, and enhances the user's karaoke experience.

[0153] Figure 7 A flowchart illustrating the fourth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 7 As shown, in some implementations, in the above-mentioned S140, the recommended sound effect parameter combination of the song to be sung by the user is obtained, including S141 to S142. S141 to S142 will be explained in detail below.

[0154] S141. Obtain the number of times the user-selected song has been performed. When the number of performances is greater than or equal to the preset number of performances, obtain multiple historical performance records of the song and the corresponding historical performance times and combinations of historical performance sound effect parameters. Obtain the performance time weights corresponding to the historical performance times of multiple historical performance records.

[0155] In this implementation, after the user selects a song to sing, the cumulative number of times the song has been sung can be extracted from the user's historical singing statistics.

[0156] In this implementation, the number of performances is directly counted as the number of times a user has performed the same song, and the statistical logic is clear.

[0157] For example, only the number of times a user has performed the song completely within the past year is counted; performance records older than one year are not included in the statistics.

[0158] In this implementation, the extracted number of performances can be compared with the preset number of performances. When the number of performances is greater than or equal to the preset number of performances, the process of reading the historical performance records of the song is triggered.

[0159] For example, if the preset number of performances is 3 and the extracted number of performances is 5, the trigger condition is met, and multiple historical performance records of the song are read.

[0160] In this implementation, the preset number of performances can be set according to different rules to adapt to different recommendation scenarios.

[0161] For example, the number of times can be fixed at 3, and the same preset value can be used regardless of the song type. The recommendation process based on history will only be triggered when the number of times the song is sung reaches 3 or more.

[0162] For example, the settings can be adjusted according to the average frequency of the song, with popular songs preset to 5 times and niche songs preset to 2 times.

[0163] In this implementation, all historical performance records of the song to be sung can be extracted from the user's historical stored data, and the historical performance time and historical performance sound effect parameter combination corresponding to each record can be extracted separately.

[0164] In this implementation, each historical performance record contains a clear time stamp and a corresponding combination of sound effect parameters, and the storage format is uniform.

[0165] For example, the historical performance time is stored in the form of a complete timestamp, with the content being "2024-05-01 20:30:15".

[0166] For example, the historical vocal effect parameter combinations are stored in a fixed parameter order, with the content being "Reverb Intensity 3, Equalizer Pop 2, Gain +1dB, Reverb Type Concert Hall", which is consistent with the previous parameter format.

[0167] In this implementation, the corresponding time weight can be matched and obtained based on the historical performance time of each historical performance record. The more recent the time, the higher the weight, and the better it reflects the user's current preference.

[0168] For example, historical performance records closer to the current time are assigned a higher performance time weight, while those further back are assigned a lower performance time weight.

[0169] In this implementation, intervals are divided according to their distance from the current time, and different intervals are assigned different weights. The assignment logic is in line with the changing patterns of user preferences.

[0170] For example, the historical performance time is divided into three intervals: historical performance records within 1 month are assigned a weight of 0.6, those between 1 and 3 months are assigned a weight of 0.3, and those more than 3 months are assigned a weight of 0.1.

[0171] For example, a linear decay assignment is used, with weight = 1 / (current date - number of days since the last performance + 1), and the fewer the number of days, the higher the weight.

[0172] In this implementation, the performance time weight can be obtained from a pre-set weight matching rule, or it can be calculated in real time according to the calculation rule.

[0173] For example, when assigning values ​​using interval division, the corresponding performance time weight is obtained by directly matching the interval where the historical performance time is located from the pre-configured interval-weight correspondence table.

[0174] For example, when using linear decay calculation, the difference in days is directly calculated based on the current time and the historical performance time, and then the corresponding performance time weight is obtained in real time through the calculation formula.

[0175] S141. Weight the combination of historical singing sound effect parameters corresponding to multiple historical singing records according to the weight of singing time, and normalize it to serve as the recommended combination of sound effect parameters for the song to be sung by the user.

[0176] In this implementation, for each dimension of sound effect parameters, the parameter value corresponding to each historical performance record is multiplied by the corresponding performance time weight, and then the weighted results of all historical records are added together. After completing the weighted summation, the result is normalized to finally obtain the recommended sound effect parameter combination corresponding to the song.

[0177] For example, for three historical performance records of a song to be performed, weighted summation and normalization are performed according to the corresponding parameter dimensions, and finally the recommended combination of sound effect parameters for the song is obtained.

[0178] In this implementation, each sound effect parameter dimension is weighted and calculated separately, and each parameter is calculated independently without confusion.

[0179] For example, given three historical performance records with reverberation intensity parameters of 3, 2, and 3, and corresponding performance time weights of 0.6, 0.3, and 0.1, the weighted calculation yields weighted results of 1.8, 0.6, and 0.3 for each record. Adding these three weighted results gives a weighted sum of 1.8 + 0.6 + 0.3 = 2.7, completing the weighted calculation for the reverberation intensity dimension. Similarly, the calculation is performed sequentially for all parameter dimensions.

[0180] In this implementation, the result of the weighted summation is divided by the sum of the weights of all performance times to complete the normalization process and eliminate the bias caused by the total weights not being 1.

[0181] For example, following the weighted calculation results of the reverberation intensity dimension above, the sum of the performance time weights for the three records is 0.6 + 0.3 + 0.1 = 1, and the normalized reverberation intensity value is 2.7 / 1 = 2.7. If the total weight is 1.2, the weighted sum is 3.6, and the normalized value is 3.6 / 1.2 = 3.

[0182] In this implementation, the final recommended sound effect parameter combination adopts the same format as the historical singing sound effect parameter combination, with each parameter corresponding to a specific value to suit subsequent adjustment needs.

[0183] For example, the final recommended sound effect parameter combination is as follows: reverb intensity 2.7, equalization level 2, gain +0.9dB, reverb type code 1, where continuous numerical parameters are retained to one decimal place, and discrete parameters are taken as the nearest integer, directly corresponding to the parameter format that the karaoke system can recognize.

[0184] With this implementation, when the number of times a user selects a song to sing is greater than or equal to the preset number of times it can be sung, the combination of historical singing sound effect parameters and the corresponding singing time weights corresponding to multiple historical singing records of the song are obtained, and the recommended sound effect parameter combination is obtained by weighted summation and normalization. This combination can fit the user's historical singing habits of the song and avoid the problem of inconsistency between general recommendations and user habits.

[0185] This implementation assigns corresponding performance time weights to different historical performance times, with more recent performance records receiving higher weights. This allows for priority matching of the user's recent sound effect preferences for the song, capturing changes in the user's sound effect preferences for a single song, and improving the timeliness of recommended sound effects for a single song.

[0186] In some implementations, in S141 above, obtaining the performance time weights corresponding to the historical performance times of multiple historical performance records includes: determining the time difference between the historical performance time and the current time corresponding to multiple historical performance records, as the performance time distance corresponding to multiple historical performance records; and determining and normalizing the ratio of the performance time distances corresponding to multiple historical performance records, as the performance time weight corresponding to multiple historical performance records.

[0187] In this implementation, for each historical performance record, the time difference between the current time and the corresponding historical performance time can be calculated, and this time difference can be used as the performance time distance corresponding to this historical performance record.

[0188] For example, if the current time is October 1, 2024, and the historical performance time is September 1, 2024, the time difference is calculated to be 30 days. This 30 days is the performance time distance of the corresponding historical performance record.

[0189] In this implementation, the ratio of the maximum singing time distance to the singing time distance of each record is obtained by dividing the maximum singing time distance by the singing time distance of each record, ensuring that the smaller the time difference, the larger the ratio.

[0190] For example, there are three historical performance records with performance time intervals of 10 days, 30 days, and 60 days, respectively. The maximum performance time interval is 60 days. Calculate the ratio for each record as 60 / 10=6, 60 / 30=2, and 60 / 60=1, respectively, and complete the ratio calculation for all records.

[0191] In this implementation, the ratios of all calculated performance time distances can be summed to obtain a total. Then, the ratio of each record is divided by the total, and the result is the normalized performance time weight of the corresponding historical performance record. The sum of all weights is 1.

[0192] For example, normalize the ratios of all performance time distances so that the sum of the weights meets the requirement of 1, which is suitable for subsequent weighted calculations.

[0193] In this implementation, the ratio of a single record is directly divided by the sum of all ratios to complete the normalization process and obtain the required performance time weight.

[0194] For example, taking the ratio calculation results of the above three records, the three ratios are 6, 2 and 1 respectively, and the sum is 6+2+1=9. The normalized weights are 6 / 9≈0.67, 2 / 9≈0.22 and 1 / 9≈0.11 respectively. Finally, the performance time weights corresponding to the three historical performance records are obtained, and the sum of the three weights is 1.

[0195] This implementation calculates the time difference between the historical performance time and the current time corresponding to multiple historical performance records as the performance time distance. The ratio of the performance time distance is normalized to obtain the performance time weight. This method can quantify the reference value of different historical records based on their time proximity. The more recent the record, the more reasonable the weight allocation, thus improving the accuracy of the recommended sound effect parameter combination.

[0196] This implementation normalizes the ratio of performance time distances to obtain performance time weights, eliminating the need for manual configuration of weight parameters. It can automatically adapt to the differences in performance time intervals among different users, reducing the complexity of custom configuration and improving the universality of weight calculation. Calculating weights based on the proportion of time differences avoids weight bias caused by differences in the time span of historical records, accurately reflecting the influence of performance records from different periods on the current recommendation, and improving the reliability of single-song sound effect recommendations.

[0197] Figure 8 A flowchart illustrating the fifth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 8 As shown, in some implementations, the above method also includes S210 to S220, which will be described in detail below.

[0198] S210. Perform multi-dimensional style semantic annotation on multiple songs in the music library, which serves as the multi-dimensional style semantics for multiple songs. Among them, the multi-dimensional style semantics include song rhythm semantics, song emotional semantics, instrument arrangement semantics, and singing difficulty semantics.

[0199] Figure 9 A schematic diagram illustrating the workflow of the fifth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 9 As shown, in this implementation, each song stored in the music library can be semantically annotated from four pre-defined dimensions, and the annotation results can be integrated as the multi-dimensional style semantics of the song.

[0200] In this implementation, the four dimensions correspond to different song features, and each dimension has clear annotation options.

[0201] For example, the annotation dimension of song rhythm semantics is divided into three options: fast rhythm, medium rhythm, and slow rhythm, and the corresponding categories are divided according to the number of beats per minute of the song.

[0202] For example, the labeling dimensions of song emotional semantics are divided into four options: cheerful and lyrical, sad and lyrical, passionate and enthusiastic, and calm and soothing, and the corresponding categories are divided according to the lyrics and melody features.

[0203] For example, the semantic annotation dimensions of instrument arrangement are divided into four categories: guitar playing and singing, piano accompaniment, band arrangement, and electronic synthesis, and the corresponding categories are divided according to the main accompanying instruments of the song.

[0204] For example, the semantic labeling dimensions of singing difficulty are divided into three options: easy, medium, and difficult, and the corresponding categories are divided according to the vocal range and modulation frequency.

[0205] In this implementation, the audio features and metadata of the song can be extracted first, and then the automatic annotation can be completed through a pre-trained semantic classification model. Alternatively, manual review and correction of the annotation results can be used.

[0206] For example, the audio file of a song is input into a trained classification model, the model outputs annotation results in four dimensions, and then the reviewers confirm and correct them before storing them in the music library.

[0207] In this implementation, the entire music library is annotated through a combination of automatic annotation and manual review, resulting in a clear process and accurate results.

[0208] For example, for a song to be labeled, the BPM value of the audio is first extracted to determine the rhythm semantics of the song, then the melody and lyric features are extracted to determine the emotional semantics of the song, the main accompanying instruments are identified to determine the instrument arrangement semantics, the vocal range is analyzed to determine the singing difficulty semantics, after automatic labeling is completed, errors are corrected by manual review, and finally the multi-dimensional style semantics of the song are stored.

[0209] S220. Obtain the target multi-dimensional style semantics corresponding to the new song selected by the user. Determine the similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics, using this as the multi-dimensional similarity of the multiple songs. Determine the style preference benefit coefficients corresponding to the multi-dimensional similarity of the multiple songs. Determine the maximum style preference benefit coefficient among the style preference benefit coefficients of the multiple songs, and obtain the target sound effect parameter combination corresponding to the maximum style preference benefit coefficient. Adjust the audio parameters of the karaoke system according to the target sound effect parameter combination.

[0210] In this implementation, after the user selects a new song to be sung that has no historical singing records, the pre-annotated multi-dimensional style semantics of the song can be extracted from the annotation data of the music library and used as the target multi-dimensional style semantics for matching.

[0211] In this implementation, the target multi-dimensional style semantics can be directly obtained from the pre-annotated music library data, without the need for temporary annotation.

[0212] For example, when a new song is added to the music library, it has already been annotated with multi-dimensional style semantics. When the user selects the song, the system directly reads the stored annotation results to obtain the target multi-dimensional style semantics, which includes the song rhythm semantics: fast-paced, the song emotion semantics: passionate and enthusiastic, the instrument arrangement semantics: band arrangement, and the singing difficulty semantics: difficult.

[0213] In this implementation, for each song in the music library that a user has historically sung, the similarity between the multi-dimensional style semantics of the song and the target multi-dimensional style semantics can be calculated, and the calculation result can be used as the multi-dimensional similarity of the song.

[0214] For example, if there are ten songs in the music library that the user has historical singing records for, the similarity between the multi-dimensional style semantics of each song and the target multi-dimensional style semantics is calculated in turn to obtain ten multi-dimensional similarity scores.

[0215] In this implementation, matching is performed on each of the four dimensions. If the same dimension is labeled, 1 point is awarded; otherwise, 0 points are awarded. The total score is divided by the number of dimensions to obtain the multi-dimensional similarity.

[0216] For example, the target multi-dimensional style semantics are fast-paced, exciting, band arrangement, and difficult, while the multi-dimensional style semantics of an existing song are fast-paced, exciting, band arrangement, and medium. There are three matches in the four dimensions, and the multi-dimensional similarity is 3 / 4 = 0.75.

[0217] In this implementation, the corresponding style preference benefit coefficient can be calculated by combining the multi-dimensional similarity of each song and the historical preference benefit coefficient of the song. The higher the benefit coefficient, the higher the degree of matching.

[0218] For example, multiplying the multidimensional similarity with the historical preference gain coefficient of the song yields the style preference gain coefficient of the corresponding song.

[0219] In this implementation, the style preference benefit coefficients of all existing songs can be traversed, the style preference benefit coefficient with the largest value can be selected, and the corresponding sound effect parameter combination can be extracted as the final target sound effect parameter combination for matching.

[0220] In this implementation, each song sung by a user is bound to the combination of sound effect parameters used in that song. After filtering out the maximum style preference benefit coefficient, the bound parameter combination can be read directly.

[0221] In this implementation, the current audio parameters of the karaoke system can be replaced with all parameter values ​​corresponding to the target sound effect parameter combination to complete the adjustment, allowing users to directly use the matched parameters to sing.

[0222] For example, the system reads the values ​​of each parameter in the target sound effect parameter combination, overwrites the original parameters of the karaoke system, and outputs the corresponding sound effect directly when the user starts singing.

[0223] This implementation method performs multi-dimensional style semantic annotation on multiple songs in the music library, including song rhythm semantics, song emotion semantics, instrument arrangement semantics, and singing difficulty semantics. It matches the target multi-dimensional style semantics of the new song to be sung, which can cover the multi-dimensional style features of the song and provide a comprehensive reference for the sound effect matching of the new song, avoiding the adaptation deviation caused by a single style matching dimension.

[0224] This implementation calculates the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics to obtain multi-dimensional similarity, corresponding to a style preference benefit coefficient. The target sound effect parameter combination corresponding to the maximum style preference benefit coefficient is selected to adjust the audio parameters, eliminating the need for users to manually set the sound effects for new songs and reducing the parameter adjustment cost when singing new songs. Based on style similarity matching historical sound effect preferences, when users sing new songs that have not been recorded, sound effects that match the style and user preferences can be quickly matched, filling the gap in sound effect adjustment when there is no historical singing data for new songs and improving the sound effect adaptation capability in all scenarios.

[0225] Figure 10 A flowchart illustrating the sixth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 10 As shown, in some implementations, in the above S210, the similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics is determined as the multi-dimensional similarity of multiple songs, including S211 to S212. S211 to S212 will be explained in detail below.

[0226] S211. Obtain the weights of multiple dimensions corresponding to the multi-dimensional style semantics set by the user.

[0227] Figure 11 A schematic diagram of the workflow of the sixth real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 11 As shown in this implementation, before calculating multi-dimensional similarity, the weight values ​​that the user has pre-set for each of the four dimensions of multi-dimensional style semantics can be extracted to provide basic parameters for subsequent weighted similarity calculation.

[0228] For example, users can set the weights of four dimensions in the preference settings interface of the karaoke system. Before calculating the similarity of new songs, the system directly reads the stored user-set weights.

[0229] In this implementation, the four dimensions of weights are obtained from different sources to suit the different user settings.

[0230] For example, the source of the data can be user-defined settings. Users can drag sliders in the settings interface to adjust the weights of the song rhythm semantic dimension, the song emotion semantic dimension, the instrument arrangement semantic dimension, and the singing difficulty semantic dimension. After saving, the four weights can be obtained directly from the user's personal settings data.

[0231] For example, the source of the data can be learned from the user's historical singing data. The system analyzes the degree of influence of different dimensions on the user's preferences in the user's historical choices, automatically learns and generates the weights of the four dimensions and stores them, and obtains them directly from the learning results during calculation.

[0232] For example, the source can be the system default configuration. For new users who have not manually set up or accumulated enough data, the system can directly call the pre-set default four-dimensional weights and obtain them from the default configuration table.

[0233] S212. Determine the dimensional similarity of the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics for each of the multiple dimensions. The sum of the products of the multi-dimensional similarity and dimensional weights of the multiple songs is used as the similarity between the multi-dimensional style semantics of the multiple songs and the target multi-dimensional style semantics.

[0234] In this implementation, for each song with existing historical performance records, the multi-dimensional style semantics of the song and the target multi-dimensional style semantics can be compared one by one according to the corresponding dimensions, and the dimensional similarity of each dimension can be calculated.

[0235] For example, for an existing song, we compare the four dimensions of song rhythm semantics, song emotional semantics, instrument arrangement semantics, and singing difficulty semantics in turn, and obtain the dimension similarity corresponding to each of the four dimensions.

[0236] In this implementation, the dimensional similarity is recorded as 1 if the dimensions are the same, and as 0 if the dimensions are different. The calculation logic for discrete classification dimensions is simple and clear.

[0237] For example, the target song's rhythmic semantics are both fast-paced, and the existing song's rhythmic semantics are both fast-paced; since they share the same dimension, their dimensional similarity is 1. The target song's emotional semantics are passionate and energetic, while the existing song's emotional semantics are calm and soothing; since they share different dimensions, their dimensional similarity is 0. The remaining two dimensions are calculated using the same rules, resulting in dimensional similarities of 1, 0, 1, and 1 for the four dimensions, respectively.

[0238] In this implementation, the dimensional similarity of each dimension can be multiplied by the corresponding dimensional weight, and then the product of all four dimensions can be added together. The sum obtained is the final similarity between the multidimensional style semantics of the song and the multidimensional style semantics of the target.

[0239] For example, for an existing song, the products of the four dimensions are calculated separately and then summed to obtain the final similarity result for that song.

[0240] In this implementation, the product of each dimension is calculated sequentially according to the dimension order, and then the sum is obtained by accumulating them one by one. The calculation process is straightforward.

[0241] For example, the dimensional similarity of the four dimensions are 1, 0, 1, and 1, respectively, and the corresponding dimensional weights are 0.3, 0.3, 0.2, and 0.2, respectively. The product of each dimension is calculated as 0.3, 0, 0.2, and 0.2, respectively. The sum of the four products is 0.3 + 0 + 0.2 + 0.2 = 0.7. Finally, the similarity between the song and the target multidimensional style semantics is 0.7.

[0242] This implementation method obtains the weights of multiple dimensions corresponding to the multi-dimensional style semantics set by the user. The multi-dimensional similarity is obtained by summing the product of the dimensional similarity of multiple dimensions and the corresponding dimensional weights. The weights can be flexibly adjusted according to the different emphases of the user on the semantics of song rhythm, song emotion, instrument arrangement, and singing difficulty, so as to adapt to the style preference needs of different users.

[0243] This implementation calculates the multi-dimensional style semantics of multiple songs and the multi-dimensional style semantics of the target song in each dimension, and then combines the weighted aggregation to obtain the final similarity. This can accurately quantify the style matching degree between different songs and the new song to be sung, avoid the bias caused by general matching, and improve the accuracy of style matching.

[0244] Figure 12 A flowchart illustrating the seventh real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 12 As shown, in some implementations, in the above S210, the style preference benefit coefficients corresponding to the multi-dimensional similarity of multiple songs are determined, including S213 to S214. S213 to S214 will be explained in detail below.

[0245] S213. Obtain the user's historical performance data and sound effect parameter combinations for multiple songs. Based on the multi-dimensional style semantics of multiple songs and the user's historical performance data for multiple songs, determine the total performance duration, frequency of use, parameter adjustment range, and performance satisfaction score corresponding to different sound effect parameter combinations under the multi-dimensional style semantics.

[0246] Figure 13 A schematic diagram of the workflow of the seventh real-time sound effect feedback adjustment and control method for a karaoke system provided in this application embodiment is shown below. Figure 13As shown, in this implementation, the user's personal storage data can be used to extract singing-related data for all songs the user has sung in the past, as well as the combination of sound effect parameters used for each performance, providing basic data for subsequent style-based statistical preferences.

[0247] For example, if a user has accumulated 50 songs in the karaoke system, the system will extract all the singing data corresponding to these 50 songs, as well as the combination of sound effect parameters used for each song.

[0248] In this implementation, the user's historical singing data contains information from multiple dimensions, covering all the content required for preference statistics.

[0249] For example, the user's historical performance data includes the complete performance duration of each song, the corresponding combination of sound effect parameters used for each song, the user's adjustment record of the sound effect parameters during each performance, the satisfaction rating given by the user after the performance, and the multi-dimensional style semantic annotation of each song. All information is categorized and stored for easy subsequent statistics.

[0250] In this implementation, all historical songs can be classified according to multi-dimensional style semantics. For different combinations of sound effect parameters under the same style classification, the preference data of four dimensions are statistically analyzed to obtain four statistical values ​​corresponding to each combination of sound effect parameters.

[0251] For example, all fast-paced, energetic songs are grouped into the same category. For the three different combinations of sound effect parameters used by users in this category, the total performance time, frequency of use, parameter adjustment range, and performance satisfaction score of each combination are statistically analyzed.

[0252] In this implementation, after classifying by style, data from four dimensions are statistically analyzed for each combination of sound effect parameters, resulting in a clear statistical logic.

[0253] For example, regarding the KTV sound effect parameter combination under the fast-paced, energetic style category, the total singing time for all users using this combination was calculated to be 120 minutes. The usage frequency of this combination was calculated as 40% of the total singing times under this style category. The total number of parameter adjustments made manually by users during use was calculated, resulting in a parameter adjustment range of 2 levels. The average user rating after all uses was calculated, resulting in a singing satisfaction rating of 4.2 stars. This completes the statistical analysis of this combination across four dimensions.

[0254] S214. Based on the user's total performance time, usage frequency, parameter adjustment range, and performance satisfaction score corresponding to different combinations of sound effect parameters under multi-dimensional style semantics, determine the historical preference benefit coefficient corresponding to different combinations of sound effect parameters under multi-dimensional style semantics. Determine the product of the multi-dimensional similarity of multiple songs and the historical preference benefit coefficient as the style preference benefit coefficient corresponding to the multi-dimensional similarity of multiple songs.

[0255] In this implementation, for each different combination of sound effect parameters under the multi-dimensional style semantic classification, the historical preference benefit coefficient corresponding to the combination can be calculated using the four dimensions of data that have been statistically completed and the corresponding weights. The benefit coefficient reflects the user's preference for the corresponding combination under this style.

[0256] For example, under the fast-paced, exciting, and passionate style category, there are three different combinations of sound effect parameters, and the corresponding historical preference gain coefficient is calculated for each combination.

[0257] In this implementation, the statistical data of the four dimensions are first normalized, and then multiplied by their respective weights and summed to obtain the final historical preference return coefficient.

[0258] For example, the normalized result of the KTV sound effect parameter combination under a certain style is: total singing time 0.8, usage frequency 0.6, parameter adjustment range 0.9, singing satisfaction score 0.9, with corresponding weights of 0.2, 0.3, 0.2, and 0.3 respectively. The calculated historical preference return coefficient is 0.79, thus completing the calculation of the historical preference return coefficient for this combination.

[0259] In this implementation, for each song with existing historical performance records, the multi-dimensional similarity of the song and the historical preference benefit coefficient of the sound effects used in the song under the style of the song can be extracted, and the product of the two can be calculated to obtain the style preference benefit coefficient of the song.

[0260] For example, for an existing song, extract its multi-dimensional similarity to the new song to be sung, then extract the historical preference benefit coefficient corresponding to the sound effects used in the song, and calculate the product of the two to obtain the style preference benefit coefficient.

[0261] In this implementation, the final style preference gain coefficient is obtained by directly multiplying the two values, and the calculation logic is simple and straightforward.

[0262] For example, the multi-dimensional similarity between an existing song and a new song to be sung is 0.7. The historical preference gain coefficient of the combination of sound effect parameters used in the song under the corresponding style is 0.79. The product is calculated to be 0.55. Finally, the style preference gain coefficient corresponding to the song is 0.55.

[0263] This implementation combines users' historical performance data for multiple songs to calculate the historical preference benefit coefficients corresponding to different combinations of sound effect parameters under multi-dimensional style semantics. The coefficients are then multiplied by the multi-dimensional similarity to obtain the style preference benefit coefficients. This approach takes into account both the song style matching degree and the user's sound effect usage preferences, avoiding the situation where recommended sound effects only match the style but do not conform to the user's usage habits, thus improving the adaptability of sound effect matching for new songs.

[0264] This implementation calculates historical preference benefit coefficients by referencing users' total singing time, usage frequency, parameter adjustment range, and singing satisfaction ratings under multi-dimensional style semantics. This comprehensively covers users' sound effect preference characteristics for songs of the same style, making the calculation of style preference benefit coefficients more in line with users' actual usage habits and improving the accuracy of preference determination. The product of multi-dimensional similarity and historical preference benefit coefficients is used as the style preference benefit coefficient, integrating style matching degree and preference fit into a single quantitative indicator. This allows for the rapid selection of the optimal target sound effect parameter combination, improving the computational efficiency of sound effect adjustment for new songs.

[0265] This application also provides another real-time sound effect feedback adjustment and control system for a karaoke system, including a unit for implementing the method described above.

[0266] Figure 14 A schematic diagram of the logic structure of a real-time sound effect feedback adjustment and control system for a karaoke system provided in this application embodiment is shown below. Figure 14 As shown, the system 1 of this embodiment includes a processing unit 11, a storage unit 12, and a transceiver unit 13. The processing unit 11 is used to process data, the storage unit 12 is used to store data, and the transceiver unit 13 is used to send and receive data. The processing unit 11, the storage unit 12, and the transceiver unit 13 cooperate with each other to implement the above-described method. The beneficial effects of the embodiments of this application have been described in the above-described method and will not be repeated here.

[0267] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0268] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0269] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0270] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0271] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0272] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0273] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0274] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A real-time sound effect feedback adjustment and control method for a karaoke system, characterized in that, The method includes: The system acquires sound effect usage information for each sound effect parameter combination across multiple historical time periods. Based on this information, it determines the historical preference benefit coefficients for each sound effect parameter combination across multiple historical time periods. These historical preference benefit coefficients are then arranged chronologically to form the corresponding historical preference benefit coefficient sequence for each user across multiple historical time periods. The system also acquires the current preference benefit coefficients for the user across multiple preset time periods within the current time period. These current preference benefit coefficients are then arranged chronologically to form the corresponding current preference benefit coefficient sequence for each user across multiple preset time periods within the current time period. The sound effect usage information includes the total performance duration, usage frequency, parameter adjustment range, and performance satisfaction score for each sound effect parameter combination. The parameter adjustment range represents the extent to which the user manually fine-tunes the sound effect parameters. Determine the mean historical preference return and the slope of historical preference change corresponding to the historical preference return coefficient sequence for multiple historical time periods; determine the mean current preference return and the slope of current preference change corresponding to multiple preset time periods within the current preference return coefficient sequence; based on the mean historical preference return and the slope of historical preference change corresponding to the historical preference return coefficient sequence for multiple historical time periods and the mean current preference return and the slope of current preference change corresponding to multiple preset time periods within the current preference return coefficient sequence, determine the preference similarity corresponding to multiple historical time periods and multiple preset time periods within the current preference return coefficient sequence; determine the maximum preference similarity among the preference similarities corresponding to multiple preset time periods, and adjust the audio parameters of the karaoke system according to the audio effect parameter combination of the historical time period corresponding to the maximum preference similarity.

2. The method according to claim 1, characterized in that, Based on user usage information for each sound effect parameter combination across multiple historical time periods, determine the user's historical preference benefit coefficient for each sound effect parameter combination across multiple historical time periods, including: Obtain the weights for total performance duration, frequency of use, parameter adjustment range, and performance satisfaction rating. The sum of the following factors is used to determine the historical preference benefit coefficient of users for each sound effect parameter combination: the product of total singing time and total singing time weight, the product of usage frequency and usage frequency weight, the product of parameter adjustment range and parameter adjustment range weight, and the product of singing satisfaction score and singing satisfaction score weight.

3. The method according to claim 2, characterized in that, Based on the historical preference return mean and historical preference change slope corresponding to the historical preference return coefficient sequences of multiple historical time periods, and the current preference return mean and current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence, the preference similarity corresponding to multiple historical time periods and multiple preset time periods within the current preference return coefficient sequence is determined, including: The difference between the historical preference return mean corresponding to the historical preference return coefficient sequence of multiple historical time periods and the current preference return mean corresponding to multiple preset time periods within the current preference return coefficient sequence is determined as the preference return difference corresponding to the historical preference return coefficient sequence of multiple historical time periods; the difference between the historical preference change slope corresponding to the historical preference return coefficient sequence of multiple historical time periods and the current preference change slope corresponding to multiple preset time periods within the current preference return coefficient sequence is determined as the preference return difference coefficient corresponding to the historical preference return coefficient sequence of multiple historical time periods. The product of the preference return difference and the preference return difference coefficient corresponding to the historical preference return coefficient sequence of multiple historical time periods is determined as the preference similarity corresponding to multiple preset time periods within the multiple historical time periods and the current preference return coefficient sequence.

4. The method according to claim 3, characterized in that, The method further includes: Arrange the preference similarity corresponding to multiple preset time periods from largest to smallest, determine the top preset number of target preference similarities, and determine the target sound effect parameter combinations corresponding to the preset number of target preference similarities; determine the target sound effect parameter vectors corresponding to the preset number of target sound effect parameter combinations. Obtain recommended sound effect parameter combinations for the song selected by the user; determine the recommended sound effect parameter vector corresponding to the recommended sound effect parameter combination; determine the similarity between a preset number of target sound effect parameter vectors and recommended sound effect parameter vectors, and use this as the recommendation coefficient for the preset number of target sound effect parameter combinations; recommend the preset number of target sound effect parameter combinations to the user in descending order of recommendation coefficient; obtain the selected sound effect parameter combination chosen by the user; adjust the audio parameters of the karaoke system according to the selected sound effect parameter combination.

5. The method according to claim 4, characterized in that, Obtain recommended sound effect parameter combinations for the song selected by the user to sing, including: Get the number of times the user-selected song has been sung; when the number of times it has been sung is greater than or equal to the preset number of times it has been sung, get multiple historical performance records of the song to be sung and the historical performance time and historical performance sound effect parameter combination corresponding to the multiple historical performance records; get the performance time weight corresponding to the historical performance time of the multiple historical performance records; The combined audio parameters of historical performances corresponding to multiple historical performance records are weighted, summed, and normalized according to the weight of performance time, and used as the recommended audio parameter combination for the song to be performed by the user.

6. The method according to claim 5, characterized in that, Obtain the performance time weights corresponding to the historical performance times of multiple historical performance records, including: Determine the time difference between the historical performance time and the current time corresponding to multiple historical performance records, and use it as the performance time distance between multiple historical performance records; determine the ratio of the performance time distance between multiple historical performance records and normalize it, and use it as the performance time weight between multiple historical performance records.

7. The method according to claim 6, characterized in that, The method further includes: Multi-dimensional style semantic annotation is performed on multiple songs in the music library to serve as the multi-dimensional style semantics of multiple songs; among which, the multi-dimensional style semantics include the rhythm semantics, the emotional semantics, the instrument arrangement semantics, and the singing difficulty semantics. Obtain the target multi-dimensional style semantics corresponding to the new song selected by the user; determine the similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics, and use this as the multi-dimensional similarity of multiple songs; determine the style preference benefit coefficient corresponding to the multi-dimensional similarity of multiple songs; determine the maximum style preference benefit coefficient among the style preference benefit coefficients of multiple songs, and obtain the target sound effect parameter combination corresponding to the maximum style preference benefit coefficient; adjust the audio parameters of the karaoke system according to the target sound effect parameter combination.

8. The method according to claim 7, characterized in that, Determine the similarity between the multi-dimensional style semantics of multiple songs and the target multi-dimensional style semantics, and use this as the multi-dimensional similarity of multiple songs, including: Obtain the weights of multiple dimensions corresponding to the multi-dimensional style semantics set by the user; Determine the dimensional similarity of the multidimensional style semantics of multiple songs and the multidimensional style semantics of the target song, respectively; determine the sum of the products of the multidimensional similarity and the dimensional weights of multiple songs as the similarity between the multidimensional style semantics of multiple songs and the target multidimensional style semantics.

9. The method according to claim 8, characterized in that, Determine the style preference payoff coefficients corresponding to the multi-dimensional similarity of multiple songs, including: Obtain historical performance data and sound effect parameter combinations of multiple songs by the user; based on the multi-dimensional style semantics of multiple songs and the user's historical performance data of multiple songs, determine the total performance duration, frequency of use, parameter adjustment range and performance satisfaction score corresponding to different sound effect parameter combinations under the multi-dimensional style semantics. Based on the total performance time, frequency of use, range of parameter adjustment, and performance satisfaction score of users under different combinations of sound effect parameters under multi-dimensional style semantics, the historical preference benefit coefficients corresponding to different combinations of sound effect parameters under multi-dimensional style semantics are determined; the product of the multi-dimensional similarity of multiple songs and the historical preference benefit coefficients is determined as the style preference benefit coefficients corresponding to the multi-dimensional similarity of multiple songs.

10. A real-time sound effect feedback adjustment and control system for a karaoke system, characterized in that, Includes units for implementing the method of any one of claims 1 to 9.

Images