Sound evaluation method, device and system, electric equipment and readable storage medium
By screening and combining the mechanical information, product information and participant characteristics of the audio files, establishing a sound evaluation model, the problem that the sound quality evaluation system in the existing technology is difficult to accurately reflect the user's feelings, and a more accurate subjective evaluation is achieved.
Patent Information
- Application Number
- CN202411082857.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-08-08
AI Technical Summary
The existing sound quality evaluation system is difficult to ensure the accuracy of subjective evaluation, because a unified model cannot meet the personalized requirements of different types of sounds, resulting in the evaluation results deviating from the user's subjective feelings.
By entering the audio file to be evaluated into the preset sound database, filter out the target sample audio file corresponding to the audio file, establish a sound evaluation model, combine the mechanical information, product information and participant characteristics of the audio file, and filter the target sample audio file to Generate more precise subjective evaluations.
It improves the accuracy of sound evaluation, makes the evaluation results more in line with the user's subjective feelings, and meets the personalized needs of different types of sounds.
Smart Images

Figure CN118609604B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of sound quality evaluation, and in particular to a sound evaluation method, a sound evaluation device, a sound evaluation system, an electric device and a non-volatile computer-readable storage medium. Background Art
[0002] At present, in the sound quality evaluation system, the common practice is to establish an objective quantitative model for subjective evaluation of sound quality by linking the subjective evaluation of sound with objective parameters, so that engineers can directly obtain the subjective evaluation of the sound by inputting the objective parameters of the sound. However, for different types of sounds, people have different requirements and expectations for sound quality. A unified model will only get the average answer of the entire group, and it is difficult to guarantee the accuracy of the subjective evaluation. Summary of the invention
[0003] The embodiments of the present invention provide a sound evaluation method, a sound evaluation device, a sound evaluation system, an electric device and a non-volatile computer-readable storage medium.
[0004] The sound evaluation method of the embodiment of the present invention includes inputting an audio file of the audio to be evaluated into a preset sound database to obtain multiple target sample audio files; establishing a sound evaluation model based on the multiple target sample audio files; and outputting a subjective evaluation of the audio to be evaluated based on the sound evaluation model.
[0005] In some embodiments, the audio file includes mechanical information and product information of the audio to be evaluated, and the audio file to be evaluated is input into a preset sound database to obtain multiple target sample audio files, including determining, according to the mechanical information, multiple sample audio files in the sound database corresponding to the mechanical information as a first sample audio file; according to the product information, determining multiple sample audio files in the first sample audio file corresponding to the product information as a second sample audio file; and determining the target sample audio file according to the second sample audio file and participant characteristics.
[0006] In some embodiments, the mechanical information includes the type of electrical equipment corresponding to the audio to be evaluated, the power type and power parameters of the electrical equipment, and determining, based on the mechanical information, multiple sample audio files corresponding to the mechanical information in the sound database as the first sample audio file includes screening the multiple sample audio files in the sound database corresponding to the type, power type and power parameters of the electrical equipment to determine the first sample audio file.
[0007] In some embodiments, the product information includes the product positioning and style positioning of the electrical equipment corresponding to the audio to be evaluated, and determining, based on the product information, multiple sample audio files corresponding to the product information in the first sample audio file as the second sample audio file includes screening the multiple sample audio files in the first sample audio file corresponding to the product positioning and style positioning of the electrical equipment to determine the second sample audio file.
[0008] In certain embodiments, the sound evaluation method includes determining the participant characteristics of the audio to be evaluated based on user input, the participant characteristics including gender, age, and region; a first weight for the age, a second weight for the gender, and a third weight for the region; determining the target sample audio file based on the second sample audio file and the participant characteristics, including screening a plurality of the sample audio files in the second sample audio file corresponding to the participant characteristics to determine the target sample audio file; or screening the second sample file based on at least one of the first weight, the second weight, and the third weight to determine the target sample audio file.
[0009] In some embodiments, the sound database includes mechanical information of multiple sample audio files and product information of multiple sample audio files, and the sound evaluation method also includes establishing a mechanical classification model based on the mechanical information of the multiple sample audio files; establishing a product classification model based on the product information of the multiple sample audio files; when the mechanical information of the audio to be evaluated is empty, determining the mechanical information of the audio to be evaluated based on the mechanical classification model; when the product information of the audio to be evaluated is empty, determining the product information of the audio to be evaluated based on the product classification model.
[0010] In certain embodiments, the sound evaluation method further includes recording sound samples in batches and storing mechanical information of multiple sound samples; obtaining product information, subjective evaluation and language description of the sound samples from each participant, and recording the participant characteristics of the participants; and establishing the sound database based on the sound samples, the mechanical information of the sound samples, the product information, subjective evaluation, language description of the sound samples and the participant characteristics of the participants.
[0011] In some embodiments, establishing the sound database includes obtaining product information, subjective evaluations and language descriptions of the sound samples from the participants multiple times to verify whether each of the participants is reliable; if so, establishing the sound database.
[0012] In certain embodiments, after establishing the sound database, the sound evaluation method further includes re-obtaining the participant's product information, subjective evaluation, and language description of the sound sample to verify whether the participant is reliable; if so, maintaining the sound database; otherwise, deleting the product information, subjective evaluation, and language description of the sound sample corresponding to the participant in the sound database; obtaining the product information, subjective evaluation, and language description of the sound sample from a new participant, and verifying whether the new participant is reliable; if so, updating the sound database; otherwise, maintaining the sound database.
[0013] In some embodiments, establishing a sound evaluation model based on multiple target sample audio files includes establishing an initial sound evaluation model based on the target sample audio files; verifying whether the initial sound evaluation model meets preset requirements; if so, using the initial sound evaluation model as the sound evaluation model; if not, prompting the user.
[0014] In some embodiments, the target sample audio file includes an actual evaluation value, and the verification of whether the initial sound evaluation model meets the preset requirements includes obtaining the predicted value of the initial sound evaluation model for the audio to be evaluated, and the actual value corresponding to the predicted value in the target sample audio file; determining whether the difference between the predicted value and the actual value is less than or equal to a preset threshold; if so, determining that the initial sound evaluation model meets the preset requirements.
[0015] In certain embodiments, before inputting the audio file to be evaluated into the preset sound database, the sound evaluation method further includes acquiring the audio file to be evaluated and preprocessing the audio file to be evaluated.
[0016] The sound evaluation device of the embodiment of the present invention comprises an acquisition module, a creation module and an output module. The acquisition module is used to input the audio file of the audio to be evaluated into a preset sound database to obtain a plurality of target sample audio files. The creation module is used to establish a sound evaluation model according to the plurality of target sample audio files. The output module is used to output the subjective evaluation of the audio to be evaluated according to the sound evaluation model.
[0017] The sound evaluation system according to the embodiment of the present invention comprises a processor, a memory, and a computer program, wherein the computer program is stored in the memory and executed by the processor, and the computer program comprises instructions for executing the sound evaluation method according to any of the above embodiments.
[0018] The electric device of the embodiment of the present invention comprises a main body, a sound evaluation system and one or more processors. The sound evaluation system is arranged in the main body. One or more processors are arranged in the main body, and the one or more processors are used to execute the instructions of the sound evaluation method of any embodiment of the claim.
[0019] A non-volatile computer-readable storage medium according to an embodiment of the present invention includes a computer program. When the computer program is executed by a processor, the processor is caused to execute the sound evaluation method according to any one of the above embodiments.
[0020] In the sound evaluation method, sound evaluation device, sound evaluation system, electrical equipment and non-volatile computer-readable storage medium of the embodiments of the present invention, after the audio file of the audio to be evaluated is input into the preset sound database, the target sample audio file corresponding to the audio file of the audio to be evaluated is determined according to the preset sound database. In this way, after the sound evaluation model is established through the target sample audio file, the sound evaluation model is more suitable for evaluating the audio to be evaluated, ensuring that when the subjective evaluation of the audio to be evaluated is output, the subjective evaluation is more accurate and more in line with the user's subjective feelings.
[0021] Additional aspects and advantages of the embodiments of the present invention will be given in part in the following description and in part will become apparent from the following description or will be learned through the practice of the embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:
[0023] Figure 1 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0024] Figure 2 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0025] Figure 3 is a schematic diagram of a scenario of a sound evaluation method according to some embodiments of the present invention;
[0026] Figure 4 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0027] Figure 5 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0028] Figure 6 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0029] Figure 7 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0030] Figure 8 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0031] Fig. 9 is a flow chart of a sound evaluation method according to certain embodiments of the present invention;
[0032] Fig.10 is a schematic diagram of a scenario of a sound evaluation method according to some embodiments of the present invention;
[0033] Fig.11 is a schematic diagram of a scenario of a sound evaluation method according to some embodiments of the present invention;
[0034] Fig.12 is a schematic diagram of a module of a sound evaluation device according to some embodiments of the present invention;
[0035] Fig.13 is a schematic diagram of the structure of a sound evaluation system in some embodiments of the present invention;
[0036] Fig.14 It is a schematic diagram of the structure of an electrical device according to some embodiments of the present invention;
[0037] Fig.15 It is a schematic diagram of the connection status between a computer-readable storage medium and a processor in certain embodiments of the present invention. DETAILED DESCRIPTION
[0038] The embodiments of the present invention are further described below in conjunction with the accompanying drawings. The same or similar reference numerals in the accompanying drawings represent the same or similar elements or elements with the same or similar functions from beginning to end. In addition, the embodiments of the present invention described below in conjunction with the accompanying drawings are exemplary and are only used to explain the embodiments of the present invention, and cannot be understood as limiting the present invention.
[0039] See also Figure 1 The embodiment of the present invention provides a sound evaluation method, which includes:
[0040] Step 011: input the audio file of the audio to be evaluated into a preset sound database to obtain multiple target sample audio files;
[0041] Step 012: Establishing a sound evaluation model according to a plurality of target sample audio files;
[0042] Step 013: Output the subjective evaluation of the audio to be evaluated according to the sound evaluation model.
[0043] In the sound evaluation method of the embodiment of the present invention, after the audio file of the audio to be evaluated is input into the preset sound database, the target sample audio file corresponding to the audio file of the audio to be evaluated is determined according to the preset sound database. In this way, after the sound evaluation model is established through the target sample audio file, the sound evaluation model is more suitable for evaluating the audio to be evaluated, ensuring that when the subjective evaluation of the audio to be evaluated is output, the subjective evaluation is more accurate and more in line with the user's subjective feelings.
[0044] Specifically, the audio file of the audio to be evaluated includes not only the sound to be evaluated, but also the mechanical information and product information of the audio to be evaluated. The preset sound database stores various types of sample audio files, mainly including sample audio, mechanical information of the sample audio, product information of the sample audio, participant characteristics of the sample audio, and other information.
[0045] It can be understood that the mechanical information of the audio to be evaluated has the same meaning as the mechanical information of the sample audio, and the product information of the audio to be evaluated has the same meaning as the product information of the sample audio.
[0046] Take the mechanical information, product positioning and participant characteristics of the sample audio as an example for explanation:
[0047] The mechanical information of the sample audio refers to the mechanical parameter information of the sample audio file, which includes the type of electrical equipment corresponding to the sample audio, the power type of the electrical equipment and the power parameters. For example, the vehicle category to which the sample audio belongs (such as trucks, buses, sports cars, sedans, etc.), the power type to which the sample audio belongs (such as pure electric vehicles, hybrid vehicles, traditional internal combustion engine vehicles), and the power parameters to which the sample audio belongs (such as the number of engine cylinders of the vehicle, 4 cylinders, 6 cylinders, etc.).
[0048] The product positioning of the sample audio refers to the product positioning and style positioning of the electrical equipment to which the sample audio belongs. For example, vehicles are divided into high-end vehicles and low-end vehicles according to their value. For another example, vehicles are divided into traditional vehicles and postmodern vehicles according to their style.
[0049] The participant characteristics of the sample audio refer to the parameters of the participants who participated in the subjective evaluation of the sample audio, such as the age of the participant, the gender of the participant, the working city of the participant, etc.
[0050] More specifically, after the audio file of the audio to be evaluated is input into the sound database, multiple sample audio files corresponding to the mechanical information and product information of the audio file of the audio to be evaluated can be screened out from the preset sound database. For example, the mechanical information of the audio to be evaluated is a type A passenger car with 4 engine cylinders and a hybrid power type. The sound database can screen out multiple sample audio files of the type A passenger car with 4 engine cylinders and a hybrid power type from all sample audio files.
[0051] It should be noted that different types of vehicles also include different subtypes. For example, the bus type may include three subtypes: A, B, and C. The sound database outputs a sample audio file of the same vehicle type and subtype as the audio file of the audio to be evaluated. For example, if the mechanical information of the audio to be evaluated is a type A bus, the mechanical information of the output sample audio is also a type A bus. For another example, if the mechanical information of the audio to be evaluated is a type B bus, the mechanical information of the output sample audio is also a type B bus.
[0052] After determining the multiple sample audio files corresponding to the mechanical information of the audio to be evaluated, it is also necessary to screen out the multiple sample audio files corresponding to the mechanical information of the audio to be evaluated and the multiple sample audio files corresponding to the product information of the audio to be evaluated. The multiple sample audio files corresponding to the mechanical information of the audio to be evaluated are A, and the multiple sample audio files corresponding to the product information of the audio to be evaluated are B for explanation.
[0053] For example, if the product information of the audio to be evaluated is a high-end, traditional car, after screening out multiple sample audio files A, multiple sample audio files whose product information is also high-end, traditional will be screened out from the multiple sample audio files A as sample audio files B.
[0054] After a plurality of sample audio files B are screened out, target sample audio files may also be screened out in combination with the participant features input by the user.
[0055] For example, the user inputs the participant characteristics of age between 20 and 30, gender is male, and working city is X. Since each sample audio file B in multiple sample audio files B corresponds to a different subjective evaluation of the participant characteristics. For example, sample audio file b1 corresponds to the participant characteristics of age between 10 and 50, gender includes male and female, and working city includes X, Y, and Z. It can be understood that sample audio file b1 corresponds to the subjective evaluation of multiple participant characteristics. In this way, for sample audio file b1, only the subjective evaluation corresponding to the sample audio file with the participant characteristics of male between 20 and 30 and working location X is obtained as the target sample audio file.
[0056] It can be understood that the target sample audio file finally obtained has the same label information, product positioning and participant characteristics as the audio to be evaluated output by the user.
[0057] In this way, when a sound evaluation model is established based on multiple target sample audio files and the subjective evaluation of the audio to be evaluated is output, the label information, product positioning, and participant characteristics of the target sample audio files contained in the sound evaluation model are the same as those of the audio to be evaluated, ensuring that the output subjective evaluation is closer to the user's feelings, thereby ensuring that the output subjective evaluation is more accurate and more in line with the user's subjective feelings.
[0058] In addition, before inputting the audio file to be evaluated into the sound database, the audio file to be evaluated can also be preprocessed, such as filtering, editing, etc., to eliminate the noise of the audio file to be evaluated, thereby ensuring the accuracy of establishing the sound evaluation model and ensuring the accuracy of the output subjective feeling.
[0059] See also Figure 2 In some embodiments, the audio file includes mechanical information and product information of the audio to be evaluated. Step 011: input the audio file of the audio to be evaluated into a preset sound database to obtain a plurality of target sample audio files, including:
[0060] Step 0111: according to the mechanical information, determining a plurality of sample audio files corresponding to the mechanical information in the sound database as first sample audio files;
[0061] Step 0112: According to the product information, a plurality of sample audio files corresponding to the product information in the first sample audio files are determined as second sample audio files;
[0062] Step 0113: Determine the target sample audio file based on the second sample audio file and the participant characteristics.
[0063] Specifically, the mechanical information of the audio to be evaluated is used to characterize the mechanical parameters of the audio to be tested, which includes the type of electrical equipment corresponding to the audio to be evaluated, the power type of the electrical equipment, and the power parameters. For example, the car category (truck, bus, sports car, sedan, etc.) to which the audio to be evaluated belongs, the power type (pure electric vehicle, hybrid vehicle, traditional internal combustion engine vehicle) to which the audio to be evaluated belongs, and the power parameters of the vehicle (4-cylinder, 6-cylinder, etc.) to which the audio to be evaluated belongs.
[0064] The sample audio files in the sound database also include mechanical information of the sample audio.
[0065] In this way, after the audio file of the audio to be evaluated is input into the sound database, the sound database can determine multiple sample audio files corresponding to the type of electrical equipment, power type and power parameters of the audio file of the audio to be evaluated as the first sample audio file.
[0066] For example, if the mechanical information of the audio to be evaluated is a sports car of type A with 6 cylinders and electric power, the sound database can output multiple sample audio files of the sports car of type A with 6 cylinders and electric power.
[0067] It can be understood that the first sample audio file is a screening of the sound database from a mechanical dimension according to the mechanical information of the audio to be evaluated.
[0068] The product information of the audio to be evaluated is used to characterize the product positioning and style positioning of the electrical equipment to which the audio to be evaluated belongs. For example, vehicles are divided into high-end vehicles and low-end vehicles according to their value. For another example, vehicles are divided into traditional vehicles and postmodern vehicles according to their style.
[0069] Specifically, after obtaining a plurality of first sample audio files, a plurality of sample audio files corresponding to the product positioning and style positioning of the audio to be evaluated may be screened out from the plurality of first sample audio files to serve as second sample audio files.
[0070] For example, if the product information of the audio to be evaluated is high-end and traditional, then multiple sample audio files with product information of high-end and traditional can be screened out from multiple first sample audio files to serve as second sample audio files.
[0071] It can be understood that the first sample audio file is a sound database that is screened for the first time from a mechanical dimension based on the mechanical information of the audio to be evaluated, and then screened for the second time from an emotional dimension.
[0072] In one embodiment, if Figure 3 As shown, an audio file of the audio to be evaluated is input into the sound database and displayed at P in the sound data model as an example for explanation. Each point in the model represents a sample audio file in the sound database.
[0073] The user can select the "Mechanical Dimension" button to input the mechanical information of the audio to be evaluated, such as vehicle type, power type, power parameters, etc., to determine the mechanical information of the audio to be evaluated. At this time, the model can display sample audio files with the same mechanical information as the audio to be evaluated, for example, displayed as dark gray dots.
[0074] After selecting the sample audio files with the same mechanical information as the audio to be evaluated, the user can also select the "emotional dimension" button to select whether the audio to be evaluated is "high-end" or "low-end", "traditional" or "post-modern", so as to determine the product information of the audio to be evaluated. At this time, the model can display the sample audio files with the same mechanical information and product information as the audio to be evaluated, for example, as gray dots in the elliptical area. In addition, the user can also select the vehicle's motion state, such as starting, idling, low-speed uniform speed, high-speed uniform speed, acceleration, rapid acceleration, deceleration, rapid deceleration, etc.
[0075] After filtering out the second sample audio file, please combine Figure 3 The user can also select the "Participant Characteristics" button to enter the characteristics of the participant, such as age, gender, and region, to filter the second sample audio file to obtain the target audio file.
[0076] For example, the user inputs the participant characteristics as age 25 to 30 years old, gender as male, and region as X. In this way, multiple sample audio files with the same participant characteristics as the user input can be obtained from the second sample audio file as target sample audio files. The sample audio files in the sound database also include subjective evaluations of the sample audio files, such as evaluations of a certain parameter of the sample audio files.
[0077] More specifically, for example, sample audio file A includes loudness evaluations by participants who are 10 years old, male, and from X; loudness evaluations by participants who are 20 years old, female, and from X; loudness evaluations by participants who are 30 years old, male, and from X; loudness evaluations by participants who are 40 years old, male, and from X; and loudness evaluations by participants who are 50 years old, female, and from Y. If the user inputs participant characteristics of ages 20 to 50, male, and from X, when filtering sample audio file A, loudness evaluations a1 by participants who are 30 years old, male, and from X; and loudness evaluations a2 by participants who are 40 years old, male, and from X can be filtered out, so as to obtain the average of a1 and a2 as the subjective evaluation corresponding to the target sample audio file.
[0078] In this way, when a sound evaluation model is established based on the target sample audio file to output the subjective evaluation of the audio to be evaluated, since the subjective evaluation corresponding to the target sample audio file is more consistent with the user's subjective evaluation, the output subjective evaluation is ensured to be more consistent with the user's subjective evaluation, thereby ensuring the accuracy of the subjective evaluation.
[0079] In addition, weights may be assigned to the participant characteristics input by the user. For example, the first weight of gender is set to a, the second weight of age is set to b, and the third weight of region is set to c.
[0080] When the second audio file is screened, the second sample audio files may be screened based on at least one of the first weight, the second weight, and the third weight to determine the target sample audio file.
[0081] For example, the user sets the first weight of the age of 25 to 30 to 2, the second weight of the gender of male to 4, and the third weight of the region of X to 3. Then, when screening the second sample audio files to obtain the target sample audio files, it can be a weighted sum of the subjective evaluations of each second sample audio file corresponding to the characteristics of multiple participants.
[0082] For example, for a second sample audio file, the subjective evaluation of loudness by a 15-year-old is 90 points, the subjective evaluation of loudness by a 20-year-old is 91 points, the subjective evaluation of loudness by a 25-year-old is 94 points, the subjective evaluation of loudness by a 30-year-old is 95 points, and the subjective evaluation of loudness by a 35-year-old is 97 points. If the user sets the first weight of the ages of 25 to 30 to 2, then the subjective evaluation of loudness corresponding to the target sample audio file screened out from the second sample audio file should be (90*1+91*1+94*2+95*2+97*1) / 7.
[0083] Similarly, for the second weight and the third weight, when determining the subjective evaluation corresponding to the target sample audio file, the weighted values are set in accordance with the gender and region input by the user, so that the subjective evaluation corresponding to the target sample audio file obtained is more in line with the user's preferences, thereby ensuring the accuracy of the subjective evaluation.
[0084] See also Figure 4 In some embodiments, the sound database includes tag information of a plurality of sample audio files and product positioning of a plurality of sample audio files, and the sound evaluation method includes:
[0085] Step 014: establishing a mechanical classification model based on the mechanical information of the plurality of sample audio files;
[0086] Step 015: Establish a product classification model based on the product information of multiple sample audio files;
[0087] Step 016: when the mechanical information of the audio to be evaluated is empty, determining the mechanical information of the audio to be evaluated based on the mechanical classification model;
[0088] Step 017: When the product information of the audio to be evaluated is empty, determine the product information of the audio to be evaluated based on the product classification model.
[0089] Specifically, the sound database contains a plurality of sample audio files, and each sample audio file contains machine information and product information. Thus, a machine classification model can be established based on the machine information of the plurality of sample audio files in the sound database, and a product classification model can also be established based on the product information of the plurality of sample audio files in the sound database.
[0090] In this way, when the mechanical information of the audio file to be evaluated is empty, the mechanical information of the audio file to be evaluated can be determined according to the mechanical classification model, thereby screening out a plurality of first sample audio files corresponding to the mechanical information of the audio file to be evaluated.
[0091] When the product information of the audio file to be evaluated is empty, the product information of the audio file to be evaluated can be determined according to the product classification model, thereby screening out a plurality of second sample audio files corresponding to the product positioning of the audio file to be evaluated.
[0092] Specifically, after inputting the audio file of the audio to be evaluated into the sound database, if the mechanical information and product information of the audio file to be evaluated are unknown, the mechanical classification model can obtain the time domain and frequency domain features of the audio to be evaluated to obtain a sample audio file similar to the time domain and frequency domain features of the audio to be evaluated, and use the mechanical information of the sample audio file as the mechanical information of the audio to be evaluated. Similarly, the product classification model can obtain the time domain and frequency domain features of the audio to be evaluated to obtain a sample audio file similar to the time domain and frequency domain features of the audio to be evaluated, and use the product information of the sample audio file as the product information of the audio to be evaluated.
[0093] It is understandable that users often cannot accurately determine the mechanical information and product information of the audio to be evaluated. That is, when the mechanical information and product information of the audio to be evaluated are unknown, the mechanical information and product information of the audio to be evaluated can be predicted through a classification model to obtain the mechanical information and product information of the audio to be evaluated, thereby achieving the output of a subjective evaluation of the audio to be evaluated.
[0094] See also Figure 5 In some embodiments, the sound evaluation method further comprises:
[0095] Step 018: Recording sound samples in batches and storing mechanical information of multiple sound samples;
[0096] Step 019: Obtain product information, subjective evaluation and language description of each participant for the sound sample, and record the participant's participant characteristics;
[0097] Step 020: Establish the sound database based on the sound samples, mechanical information of the sound samples, product information of the sound samples, subjective evaluations, language descriptions, and participant characteristics of the participants.
[0098] Specifically, when outputting the subjective evaluation of the audio to be evaluated, a preset sound database is indispensable, and therefore, a sound database needs to be established first.
[0099] First, the sound samples can be recorded in batches, and the mechanical information (i.e., mechanical parameters) of the recorded sound samples can be recorded. After the sound samples are recorded, the sound samples can be filtered or edited to reduce the noise in the sound samples, thereby ensuring the quality of the sound samples and the accuracy of the subjective evaluation of the sound samples.
[0100] Next, more participants can be recruited to ensure the diversity of participants, and the participants can conduct subjective evaluations of the sound samples to obtain the product positioning, subjective evaluations and language descriptions obtained by each participant when conducting subjective evaluations of the sound samples. Among them, the language description can be the participant's subjective feelings about the sound sample, such as the participant describing the sound sample as a relatively special sound emitted by a high-end traditional sports car, which is strong, stable, soft, low-key, etc.
[0101] In this way, a sound database can be established for each sound sample and the label information, product positioning, subjective evaluation, language description and participant characteristics corresponding to the sound sample.
[0102] When recording sound samples, the recording conditions of the sound samples can also be recorded, such as the recording environment, recording method (rating method, numerical estimation method, paired comparison method, etc.), so as to conduct repeatable experiments by changing one or more of the recording conditions, thereby facilitating the later verification of the reliability of the participants corresponding to the sound samples.
[0103] See also Figure 6 In some embodiments, step 020: establishing a sound database includes:
[0104] Step 0201: Obtain the product positioning, subjective evaluation and language description of the sound samples from the participants multiple times to verify whether each participant is reliable;
[0105] Step 0202: If yes, establish a sound database.
[0106] Specifically, when establishing a sound database, it is also necessary to verify whether the data output by the participants in the sound database (i.e., the participants' product positioning, subjective evaluation and language description of the sound samples) are reliable in order to ensure the reliability of the sound database.
[0107] Therefore, after obtaining each participant's product positioning, subjective evaluation and language description of the sound sample each time, it is stored, and the participant's product positioning, subjective evaluation and language description of the sound sample are repeatedly obtained multiple times to compare the data output by the participants stored multiple times.
[0108] In this way, by comparing whether the data output by the participants stored multiple times are consistent or similar, it is possible to determine whether the participants are reliable. It can be understood that when the data output by the participants multiple times is consistent or similar, it means that the reliability of the participant's evaluation of the sound sample is high, and the data output by the participant is valid data. Then, a sound database can be established based on the data output by the participant, the sound sample, and the label information of the sound sample to ensure that the reliability of the sound database is high.
[0109] See also Figure 7 In some embodiments, after establishing the sound database, the sound evaluation method further includes:
[0110] Step 021: Re-obtain the product information, subjective evaluation and language description of the sound sample from the participant to verify whether the participant is reliable;
[0111] Step 022: If yes, keep the sound database; otherwise, delete the product information, subjective evaluation and language description of the sound sample corresponding to the participant in the sound database;
[0112] Step 023: Obtain product information, subjective evaluation and language description of the sound sample from the new participant, and verify whether the new participant is reliable;
[0113] Step 024: If yes, update the sound database; otherwise, keep the sound database.
[0114] Specifically, due to different conditions, the product information, subjective evaluation and language description of the sound sample may change. For example, the change of the participant's mood may cause the output product information, subjective evaluation and language description to be inconsistent when evaluating the sound sample.
[0115] Therefore, in order to ensure that the data in the sound database is always highly reliable, after the sound database is established, it is also possible to re-obtain the participants' product positioning, subjective evaluation and language description of the sound samples, and verify whether the data output by the participants is reliable.
[0116] If the data output by the participant is still reliable, the current sound database is maintained unchanged, that is, the sound database is kept. If the data output by the participant is unreliable, the data corresponding to the participant recorded in the sound database is considered unreliable, and the product information, subjective evaluation and language description of the sound sample corresponding to the participant in the sound database need to be deleted.
[0117] Furthermore, to ensure the comprehensiveness of the sound database, the product information, subjective evaluation and language description of the sound samples by the new participants can be obtained and stored, so as to compare the data output by the new participants multiple times to verify the reliability of the new participants.
[0118] If the data output by the new participant is reliable, the sound database can be updated, that is, the data output by the new participant is added to the sound database to ensure the comprehensiveness of the sound database, thereby further ensuring the accuracy of the output subjective evaluation. If the data output by the new participant is unreliable, the current sound database is maintained unchanged, that is, the sound database is kept.
[0119] See also Figure 8 and Fig. 9 In some implementations, step 012: establishing a sound evaluation model based on a plurality of target sample audio files, comprises:
[0120] Step 0121: Establishing an initial sound evaluation model based on the target sample audio file;
[0121] Step 0122: Verify whether the initial sound evaluation model meets the preset requirements;
[0122] Step 0123: If yes, use the initial sound evaluation model as the sound evaluation model;
[0123] Step 0124: If not, prompt the user.
[0124] The target sample audio file includes an actual evaluation value. Step 0122: verifying whether the initial sound evaluation model meets the preset requirements, specifically including:
[0125] Step 01221: Obtain the predicted value of the initial sound evaluation model for the audio to be evaluated, and the actual value corresponding to the predicted value in the target sample audio file;
[0126] Step 01222: Determine whether the difference between the predicted value and the actual value is greater than or equal to a preset threshold;
[0127] Step 01223: If yes, determine whether the initial evaluation model meets the preset requirements.
[0128] Specifically, when establishing a sound evaluation model according to a target sample audio file, an initial sound evaluation model may be first established based on the target sample audio file.
[0129] For example, based on the target sample audio file, an initial sound evaluation model is established in a linear regression manner. At this point, a regression line of the subjective evaluation relative to each parameter in the target sample audio file can be obtained. For example, the loudness of the target sample audio file relative to the score of the subjective evaluation, and the tone of the target sample audio file relative to the score of the subjective evaluation.
[0130] In this way, whether the initial sound evaluation model meets the preset requirements can be verified based on whether the difference between the predicted value of each parameter of the audio to be evaluated and the actual value of the target sample audio file is less than the preset threshold.
[0131] For example, if the loudness of the audio to be evaluated is 85, the initial sound evaluation model outputs a subjective evaluation value of 20 points for the loudness of the audio to be evaluated, while the subjective evaluation value of the same loudness in the target sample audio file is 20.1 points. That is, the difference between the predicted value of each parameter of the audio to be evaluated and the actual value of the target sample audio file is 0.1 points.
[0132] In verifying the initial sound evaluation model, there is a determination coefficient R, which ranges from 0 to 1. The closer R is to 1, the more accurate the subjective evaluation output by the initial sound evaluation model is. And the determination coefficient R is inversely proportional to the square of the difference between the predicted value of each parameter of the audio to be evaluated and the actual value of the target sample audio file. That is, the smaller the difference between the predicted value of each parameter of the audio to be evaluated and the actual value of the target sample audio file, the larger the determination coefficient R, and the more accurate the subjective evaluation output by the initial sound evaluation model is.
[0133] Therefore, when the difference between the predicted value of each parameter of the audio to be evaluated and the actual value of the target sample audio file is less than or equal to the preset threshold, it means that the determination coefficient R is close to 1, and the subjective evaluation output by the initial sound evaluation model is more accurate, that is, the initial sound evaluation model meets the preset requirements.
[0134] When the initial sound evaluation model does not meet the preset requirements, the user can be reminded, such as reminding the user to expand the sound database, or reminding the user to change the way of establishing the sound evaluation model. For example, the default way of establishing the sound evaluation model is linear regression, which can be replaced by a neural network way to establish the sound evaluation model.
[0135] In summary, the sound evaluation model can output the subjective evaluation of the audio to be evaluated based on the subjective evaluation of the target audio file.
[0136] Please combine Fig.10The sound evaluation model can evaluate the audio to be tested from 10 aspects, including high-low pitch, fluctuation-stable, soft-sharp, bold-low-key, strong-weak, loud-low volume, special-normal, roaring-quiet, total score, and ranking in the current sample library (i.e., ranking in similar sample audio files in the sound database). Among them, subjective evaluation is not limited to 10 aspects, and can also be evaluation of more than 10 aspects.
[0137] In one embodiment, please combine Fig.10 and Fig.11 When the sound evaluation model outputs the subjective evaluation of the audio to be evaluated, it can also output the subjective evaluation of the sample audio of the same type in the current sample library with any ranking, and display it in the form of a chart to more intuitively show the subjective evaluation of the audio to be evaluated to the user. For example, the subjective evaluation of the audio to be evaluated and the subjective evaluation of the sample audio of the same type ranked first in the current sample library are displayed separately, so that the user can more intuitively know the comprehensive quality of the audio to be evaluated.
[0138] In addition, after the sound evaluation model is established, the correlation between the average value of the subjective evaluation of the sample audio file and a certain parameter of the sample audio file can be calculated to analyze the influence of a certain parameter in the sample audio file on the subjective evaluation.
[0139] For example, there are 5 target sample audio files, and each target sample audio file corresponds to N participants. The loudness score evaluation of each sample audio file is X1-Xn, and the average value of the subjective evaluation of the loudness parameter of each sample audio file is (X1+Xn) / N. That is, in this way, the influence of the loudness parameter on the subjective evaluation can be obtained, so as to understand the influence of each parameter of the target sample audio file on the subjective evaluation value.
[0140] See also Fig.12 In some embodiments, the sound evaluation device 10 includes an acquisition module 11, a building module 12, and an output module 13. The acquisition module 11 is used to input the audio file of the audio to be evaluated into a preset sound database to obtain multiple target sample audio files. The building module 12 is used to establish a sound evaluation model based on the multiple target sample audio files. The output module 13 is used to output the subjective evaluation of the audio to be evaluated based on the sound evaluation model.
[0141] In some embodiments, the audio file includes mechanical information and product information of the audio to be evaluated, and the acquisition module 11 is specifically used to determine, based on the mechanical information, multiple sample audio files corresponding to the mechanical information in the sound database as the first sample audio file; based on the product information, determine multiple sample audio files corresponding to the product information in the first sample audio file as the second sample audio file; based on the second sample audio file and the characteristics of the participants, determine the target sample audio file.
[0142] In some embodiments, the mechanical information includes the type of electrical equipment corresponding to the audio to be evaluated, the power type and power parameters of the electrical equipment, and the acquisition module 11 is specifically used to screen multiple sample audio files corresponding to the type, power type and power parameters of the electrical equipment in the sound database to determine the first sample audio file.
[0143] In some embodiments, product positioning includes product positioning and style positioning of the electrical equipment corresponding to the audio to be evaluated, and the acquisition module 11 is specifically used to screen multiple sample audio files corresponding to the product positioning and style positioning of the electrical equipment in the first sample audio file to determine the second sample audio file.
[0144] The sound evaluation device 10 further includes a determination module 14 , a recording module 15 , a verification module 16 , a deletion module 17 , an update module 18 and a processing module 19 .
[0145] In some embodiments, the determination module 14 is used to determine the participant characteristics of the audio to be evaluated based on the user's input, the participant characteristics include gender, age and region; set a first weight for age, a second weight for gender and a third weight for region. The acquisition module is also specifically used to screen multiple sample audio files corresponding to the participant characteristics in the second sample audio file to determine the target sample audio file; or screen the second sample audio file based on at least one of the first weight, the second weight and the third weight to determine the target sample audio file.
[0146] In some embodiments, the sound database includes mechanical information of multiple sample audio files and product information of multiple sample audio files. The establishment module 12 is specifically used to establish a mechanical classification model based on the mechanical information of the multiple sample audio files; and to establish a product classification model based on the product information of the multiple sample audio files. The determination module 14 is specifically used to determine the mechanical information of the audio to be evaluated based on the mechanical classification model when the mechanical information of the audio to be evaluated is empty; and to determine the product information of the audio to be evaluated based on the product classification model when the product information of the audio to be evaluated is empty.
[0147] The recording module 15 is used to record sound samples in batches and store the mechanical information of multiple sound samples. The acquisition module 11 is used to obtain the product information, subjective evaluation and language description of each participant for the sound sample, and record the participant characteristics of the participant. The establishment module 12 establishes a sound database based on the sound samples, the mechanical information of the sound samples, the product information, subjective evaluation, language description of the sound samples and the participant characteristics of the participants.
[0148] The verification module 16 is used to obtain the product information, subjective evaluation and language description of the sound samples from the participants for multiple times to verify whether each participant is reliable. If the participants are reliable, the establishment module 13 is used to establish a sound database.
[0149] The verification module 17 is also used to retrieve the product information, subjective evaluation and language description of the sound sample by the participant to verify whether the participant is reliable. If the participant is unreliable, the deletion module 17 is used to delete the product information, subjective evaluation and language description of the sound sample corresponding to the participant in the sound database. If the participant is reliable, the sound database is maintained.
[0150] The verification module 17 is also used to obtain the product information, subjective evaluation and language description of the sound sample from the new participant, and verify whether the new participant is reliable. In the case that the participant is unreliable, the sound database is maintained. In the case that the participant is reliable, the update module is used to update the sound database.
[0151] The establishment module 12 is specifically used to establish an initial sound evaluation model based on the target sample audio file; the verification module 16 is used to verify whether the initial sound evaluation model meets the preset requirements; if so, the initial sound evaluation model is used as the sound evaluation model; if not, the user is prompted.
[0152] The target sample audio file includes an actual evaluation value. The verification module 16 is specifically used to obtain the predicted value of the initial sound evaluation model for the audio to be evaluated, and the actual value corresponding to the predicted value in the target sample audio file; determine whether the difference between the predicted value and the actual value is less than or equal to a preset threshold; if so, determine that the initial evaluation model meets the preset requirements.
[0153] The processing module 19 is specifically used to obtain the audio file to be evaluated and pre-process the audio file.
[0154] It should be noted that the explanation of the sound evaluation method in the aforementioned embodiment is also applicable to the sound evaluation device 10 in the embodiment of the present invention, and will not be further described here.
[0155] See also Fig.13The embodiment of the present invention further provides a sound evaluation system 100, which includes a processor 110, a memory 120, and a computer program 130. The computer program 130 is stored in the memory 120 and executed by the processor 110. The computer program 130 includes a method for executing the sound evaluation method in the above embodiment.
[0156] For example, the computer program 130 is used to execute the following sound evaluation method:
[0157] 011: Input the audio file of the audio to be evaluated into the preset sound database to obtain multiple target sample audio files;
[0158] 012: Establish a sound evaluation model based on multiple target sample audio files;
[0159] 013: Output the subjective evaluation of the audio to be evaluated based on the sound evaluation model.
[0160] For another example, the computer program 130 is used to execute the following sound evaluation method:
[0161] 0111: determining, according to the mechanical information, a plurality of sample audio files corresponding to the mechanical information in the sound database as first sample audio files;
[0162] 0112: According to the product information, determine multiple sample audio files corresponding to the product information in the first sample audio files as second sample audio files;
[0163] 0113: Determine a target sample audio file based on the second sample audio file and the participant characteristics.
[0164] It should be noted that the explanations and descriptions of the sound evaluation method and the sound evaluation device 10 in the aforementioned embodiments are also applicable to the sound evaluation system 100 in the embodiments of the present invention, and will not be further described here.
[0165] See also Fig.14 The embodiment of the present invention further provides an electric device 300, which includes a body 310, a sound evaluation system 100, and one or more processors 110. The sound evaluation system 100 and the one or more processors 110 are arranged in the body 310. The one or more processors 110 are used to execute the sound evaluation method in the above embodiment.
[0166] For example, one or more processors 110 are used to implement the following sound evaluation method:
[0167] 011: Input the audio file of the audio to be evaluated into the preset sound database to obtain multiple target sample audio files;
[0168] 012: Establish a sound evaluation model based on multiple target sample audio files;
[0169] 013: Output the subjective evaluation of the audio to be evaluated based on the sound evaluation model.
[0170] For another example, the one or more processors 110 are used to implement the following sound evaluation method:
[0171] 0111: determining, according to the mechanical information, a plurality of sample audio files corresponding to the mechanical information in the sound database as first sample audio files;
[0172] 0112: According to the product information, determine multiple sample audio files corresponding to the product information in the first sample audio files as second sample audio files;
[0173] 0113: Determine a target sample audio file based on the second sample audio file and the participant characteristics.
[0174] It should be noted that the explanations and descriptions of the sound evaluation method, the sound evaluation device 10, and the sound evaluation system 100 in the aforementioned embodiments are also applicable to the electrical equipment 300 in the embodiments of the present invention, and will not be further described here.
[0175] See also Fig.15 The embodiment of the present invention further provides a computer-readable storage medium 500 on which a computer program 510 is stored. When the computer program 510 is executed by the processor 110, the sound evaluation method of any of the above embodiments is implemented.
[0176] For example, when the program is executed by the processor 110, the following sound evaluation method is implemented:
[0177] 011: Input the audio file of the audio to be evaluated into the preset sound database to obtain multiple target sample audio files;
[0178] 012: Establish a sound evaluation model based on multiple target sample audio files;
[0179] 013: Output the subjective evaluation of the audio to be evaluated based on the sound evaluation model.
[0180] For another example, when the program is executed by the processor 110, the following sound evaluation method is implemented:
[0181] 0111: determining, according to the mechanical information, a plurality of sample audio files corresponding to the mechanical information in the sound database as first sample audio files;
[0182] 0112: According to the product information, determine multiple sample audio files corresponding to the product information in the first sample audio files as second sample audio files;
[0183] 0113: Determine a target sample audio file based on the second sample audio file and the participant characteristics.
[0184] It should be noted that the explanations and descriptions of the sound evaluation method, the sound evaluation device 10, the sound evaluation system 100, and the electrical equipment 300 in the aforementioned embodiments are also applicable to the computer-readable storage medium 500 of the embodiments of the present invention, and will not be elaborated here.
[0185] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.
[0186] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.
[0187] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, a computer-readable storage medium can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable storage media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable storage medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.
[0188] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0189] A person of ordinary skill in the art can understand that all or part of the steps carried by the method of the above-mentioned embodiment can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, and the program, when executed, includes one or a combination of the steps of the method embodiment. In addition, each functional unit in each embodiment of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into one module. The above-mentioned integrated module can be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.
[0190] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention. The scope of the present invention is defined by the claims and their equivalents.
Claims
1. A sound evaluation method, characterized in that: include: Input the audio file of the audio to be evaluated into a preset sound database to obtain multiple target sample audio files; Establishing a sound evaluation model according to the plurality of target sample audio files; Outputting a subjective evaluation of the audio to be evaluated according to the sound evaluation model; The audio file includes mechanical information and product information of the audio to be evaluated, and the audio file of the audio to be evaluated is input into a preset sound database to obtain a plurality of target sample audio files, including: According to the mechanical information, determining a plurality of sample audio files corresponding to the mechanical information in the sound database as first sample audio files; According to the product information, determining a plurality of sample audio files corresponding to the product information in the first sample audio files as second sample audio files; The target sample audio file is determined according to the second sample audio file and the participant characteristics.
2. The sound evaluation method according to claim 1, characterized in that: The mechanical information includes the type of electrical equipment corresponding to the audio to be evaluated, the power type and power parameters of the electrical equipment, and the determining, according to the mechanical information, a plurality of sample audio files corresponding to the mechanical information in the sound database as first sample audio files, includes: A plurality of sample audio files corresponding to the type, power type and power parameter of the electrical equipment in the sound database are screened to determine the first sample audio file.
3. The sound evaluation method according to claim 1, characterized in that: The product information includes the product positioning and style positioning of the electrical device corresponding to the audio to be evaluated, and the determining, according to the product information, a plurality of sample audio files corresponding to the product information in the first sample audio file as the second sample audio file includes: A plurality of sample audio files corresponding to the product positioning and style positioning of the electrical device are screened in the first sample audio files to determine the second sample audio file.
4. The sound evaluation method according to claim 1, characterized in that: include: Determine, based on the user's input, the participant characteristics of the audio to be evaluated, wherein the participant characteristics include gender, age, and region; Setting a first weight for the age, a second weight for the gender, and a third weight for the region; The step of determining the target sample audio file according to the second sample audio file and the participant characteristics includes: Screening a plurality of the sample audio files corresponding to the participant features in the second sample audio files to determine the target sample audio file; or Based on at least one of the first weight, the second weight, and the third weight, the second sample audio file is screened to determine the target sample audio file.
5. The sound evaluation method according to claim 1, characterized in that: The sound database includes mechanical information of a plurality of sample audio files and product information of a plurality of sample audio files, and the sound evaluation method further includes: Establishing a mechanical classification model according to the mechanical information of the plurality of sample audio files; Establishing a product classification model based on the product information of the plurality of sample audio files; When the mechanical information of the audio to be evaluated is empty, determining the mechanical information of the audio to be evaluated based on the mechanical classification model; When the product information of the audio to be evaluated is empty, the product information of the audio to be evaluated is determined based on the product classification model.
6. The sound evaluation method according to claim 1, characterized in that: The sound evaluation method further comprises: Recording sound samples in batches and storing mechanical information of a plurality of the sound samples; Obtain product information, subjective evaluation and language description of the sound sample from each participant, and record the participant characteristics of the participant; The sound database is established based on the sound samples, mechanical information of the sound samples, product information of the sound samples, subjective evaluations, language descriptions, and participant characteristics of the participants.
7. The sound evaluation method according to claim 6, characterized in that: The step of establishing the sound database comprises: Obtaining product information, subjective evaluations and language descriptions of the sound samples from the participants multiple times to verify whether each participant is reliable; If so, the sound database is established.
8. The sound evaluation method according to claim 7, characterized in that: After establishing the sound database, the sound evaluation method further includes: Re-obtaining the participant's product information, subjective evaluation and language description of the sound sample to verify whether the participant is reliable; If yes, keep the voice database; otherwise, delete the product information, subjective evaluation and language description of the voice sample corresponding to the participant in the voice database; Obtain product information, subjective evaluation and language description of the sound sample from the new participant, and verify whether the new participant is reliable; If so, update the sound database; otherwise, keep the sound database.
9. The sound evaluation method according to claim 1, characterized in that: The step of establishing a sound evaluation model according to the plurality of target sample audio files comprises: Based on the target sample audio file, establishing an initial sound evaluation model; Verifying whether the initial sound evaluation model meets preset requirements; If so, using the initial sound evaluation model as the sound evaluation model; If not, the user is prompted.
10. The sound evaluation method according to claim 9, characterized in that: The target sample audio file includes an actual evaluation value, and the verifying whether the initial sound evaluation model meets the preset requirements includes: Obtaining a predicted value of the initial sound evaluation model for the audio to be evaluated, and an actual value in the target sample audio file corresponding to the predicted value; Determine whether the difference between the predicted value and the actual value is less than or equal to a preset threshold; If so, it is determined that the initial sound evaluation model meets the preset requirements.
11. The sound evaluation method according to claim 1, characterized in that: Before inputting the audio file of the audio to be evaluated into the preset sound database, the sound evaluation method further includes: The audio to be evaluated is obtained, and the audio to be evaluated is preprocessed.
12. A sound evaluation device, characterized in that: The device is used to implement the sound evaluation method according to any one of claims 1 to 11, and the device comprises: An acquisition module, wherein the acquisition module is used to input an audio file of the audio to be evaluated into a preset sound database to obtain a plurality of target sample audio files; A building module, the building module is used to build a sound evaluation model according to a plurality of target sample audio files; An output module is used to output a subjective evaluation of the audio to be evaluated based on the sound evaluation model.
13. A sound evaluation system, characterized in that: include: Processor, memory; and A computer program, wherein the computer program is stored in the memory and executed by the processor, and the computer program includes instructions for executing the sound evaluation method according to any one of claims 1 to 11.
14. An electrical device, characterized in that: include ontology; A sound evaluation system, wherein the sound evaluation system is arranged on the main body; One or more processors, one or more of the processors are arranged in the main body, and the one or more processors are used to execute instructions of the sound evaluation method described in any one of claims 1-11.
15. A non-volatile computer-readable storage medium containing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to execute the sound evaluation method according to any one of claims 1 to 11.
