Vehicle-mounted call quality test method and device, electronic equipment and storage medium

Abstract

The invention discloses a vehicle-mounted call quality testing method and device, electronic equipment and a storage medium, and relates to the technical field of voice call quality detection.The method comprises the steps that a to-be-tested included angle value formed between an environment wind direction in preset call environment wind condition information and a target vehicle direction and sound quality influence feature data are counted; analyzing the to-be-measured included angle value in an acute angle range or a right angle range or an obtuse angle range, when the obtained target vehicle speed is smaller than the minimum value in preset vehicle speed condition interval values, performing statistics to obtain a second influence degree factor, and calculating the to-be-measured included angle value according to the first influence degree factors 1, 2 and 3 and the second influence degree factor. And performing keyword extraction after tone quality adjustment on the simulated to-be-evaluated voice information, and evaluating the content integrity of the simulated to-be-evaluated voice information according to the obtained first or second to-be-tested keyword information. According to the vehicle-mounted call quality test method and device, the electronic equipment and the storage medium provided by the invention, the accuracy of the vehicle-mounted call quality test can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of voice call quality detection, and in particular to a method, device, electronic device and storage medium for testing vehicle call quality. Background Art

[0002] Good voice call quality directly affects the user's experience satisfaction in actual usage scenarios. High-quality calls can reduce users' communication barriers and improve safety and convenience during driving.

[0003] In traditional technology, the test of call voice quality mainly evaluates the distortion of voice calls in terms of timbre and pitch, but ignores the further evaluation and analysis of the completeness of the voice content. There is also no further analysis of the correlation between the completeness of the voice content and environmental factors, which leads to errors in voice content conversion, reduces user experience satisfaction and the timeliness of risk management. Summary of the invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a method, device, electronic device and storage medium for testing the quality of in-vehicle calls.

[0005] In a first aspect, the present application provides a method for testing the in-vehicle call quality, the method comprising: Perform call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, preset call environment wind condition information, and count the angle value to be measured formed between the ambient wind direction and the target vehicle direction in the call environment wind condition information and the sound quality impact characteristic data; When the angle value to be measured is in an acute angle range, a right angle range or an obtuse angle range, the influence degree factors of the sound quality influence characteristic data under three conditions are statistically obtained to obtain a first influence degree factor 1, a first influence degree factor 2 and a first influence degree factor 3; when the obtained target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, the second influence degree factor is statistically obtained; A simulation prediction scheme model is trained to predict pretreatment scheme 1, pretreatment scheme 2 and pretreatment scheme 3 according to the first influence degree factor 1, the first influence degree factor 2 and the first influence degree factor 3; According to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3, the keyword information after the sound quality adjustment of the simulated speech information to be evaluated is extracted and integrated to obtain the first keyword information to be tested; According to the second impact factor, preprocessing scheme four is predicted. According to preprocessing scheme four, the keyword information after sound quality adjustment of the simulated voice information to be evaluated is extracted to obtain the second keyword information to be tested. According to the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information belonging to the original voice information, the content completeness of the simulated voice information to be evaluated is evaluated.

[0006] Preferably, original voice information that needs to be processed for subsequent calls is obtained, and keyword information is extracted from the original voice information to obtain original keyword information; Preset call environment wind condition information, and perform call simulation on original voice information according to the call environment wind condition information to obtain simulated voice information to be evaluated; A vehicle speed condition interval value is preset to obtain a target vehicle speed of the user's vehicle, and when the target vehicle speed is within the vehicle speed condition interval value, the target vehicle direction of the user's vehicle is counted; Counting the angle data between the ambient wind direction in the call environment wind condition information and the target vehicle direction to obtain the angle value to be measured; Obtain the degree of opening of the windows of the user's vehicle and count the ambient wind speed in the call environment wind condition information; The window opening degree, ambient wind speed and target vehicle speed are combined into sound quality impact characteristic data.

[0007] Preferably, when the angle value to be measured is within the acute angle range, a first influence factor 1 of the influence on the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data; When the angle value to be measured is at a right angle, a first influence factor 2 of the influence of the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data; When the angle value to be measured is within the obtuse angle range, the first influence factor three of the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data; When the target vehicle speed is less than the minimum value in the vehicle speed condition interval value, a second influence degree factor of the sound quality of the original voice information during the call processing is extracted according to the target vehicle speed and the window opening degree.

[0008] Preferably, historical call feature information recorded by the target user in a historical period and a sound quality processing solution set corresponding to the historical call feature information are obtained, and a simulation prediction solution model is trained based on the historical call feature information and the sound quality processing solution set; Inputting the first impact factor 1 and the simulated speech information to be evaluated into a simulation prediction scheme model for testing to obtain a preprocessing scheme 1; Inputting the first impact factor 2 and the simulated speech information to be evaluated into the simulation prediction scheme model for testing to obtain the preprocessing scheme 2; The first impact factor three and the simulated speech information to be evaluated are input into the simulation prediction scheme model for testing to obtain the preprocessing scheme three.

[0009] Preferably, according to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3, the voice tones of different segments of the simulated voice information to be evaluated are adjusted to obtain the first preprocessed voice information 1, the first preprocessed voice information 2 and the first preprocessed voice information 3; Performing text conversion on the first preprocessed voice information one, the first preprocessed voice information two and the first preprocessed voice information three respectively to obtain a first preprocessed text one, a first preprocessed text two and a first preprocessed text three; The first preprocessed text one, the first preprocessed text two and the first preprocessed text three are respectively integrated to extract keyword information to obtain first keyword information to be tested.

[0010] Preferably, the second impact factor and the simulated speech information to be evaluated are input into a simulation prediction scheme model for testing to obtain a preprocessing scheme 4; According to the fourth preprocessing scheme, the voice tones of different segments of the simulated speech information to be evaluated are adjusted to obtain second preprocessed speech information; Performing text conversion on the second preprocessed voice information to obtain a second preprocessed text, and extracting keyword information from the second preprocessed text to obtain second keyword information to be tested; The first keyword information to be tested or the second keyword information to be tested is compared with the original keyword information to obtain a comparison result, and the content completeness of the simulated voice information to be evaluated is evaluated according to the comparison result.

[0011] Second, the test equipment for in-vehicle call quality includes: The associated impact data statistics module is used to perform call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, preset call environment wind condition information, and count the angle value to be measured formed between the ambient wind direction in the call environment wind condition information and the target vehicle direction, as well as the sound quality impact characteristic data; A conditional influence factor extraction module is used to obtain the first influence factor 1, the first influence factor 2 and the first influence factor 3 by performing statistics on the influence factors of the sound quality influence characteristic data under three conditions when the angle value to be measured is in the acute angle range, the right angle range or the obtuse angle range, and when the obtained target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, obtain the second influence factor by statistics; A sound quality processing scheme prediction module is used to train a simulation prediction scheme model, and predict preprocessing scheme one, preprocessing scheme two and preprocessing scheme three according to the first influence factor one, the first influence factor two and the first influence factor three; A keyword extraction module is used to extract and integrate the keyword information after the sound quality adjustment of the simulated speech information to be evaluated according to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3 to obtain the first keyword information to be tested; The content integrity evaluation module is used to predict preprocessing scheme four according to the second influence factor, extract the keyword information after sound quality adjustment of the simulated voice information to be evaluated according to the preprocessing scheme four, and evaluate the content integrity of the simulated voice information to be evaluated according to the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information belonging to the original voice information.

[0012] In a third aspect, the present invention provides an electronic device comprising: a memory for storing a computer software program; and a processor for reading and executing the computer software program, thereby implementing the method for testing the in-vehicle call quality as described in any one of the first aspects.

[0013] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the storage medium stores a computer software program, and when the computer software program is executed by a processor, the method for testing the in-vehicle call quality described in any one of the first aspects is implemented.

[0014] Compared with the prior art, the present invention has the following characteristics and beneficial effects: Through preliminary call simulation processing of the original voice information, and the user's driving vehicle exists in this simulated call processing process, and it is affected by the interference of external environmental factors. In order to reduce the misjudgment of keyword extraction in the voice information after the simulated call, the external influence environment in the simulated call is analyzed in detail, and the wind condition information of the call environment is preset, which mainly includes the noise influence of wind speed and wind direction in the external environment on the voice during the call, and the angle value formed between the wind direction and the target vehicle direction is distinguished. Because the different angle values ​​​​formed have different degrees of noise interference to the voice call, a more comprehensive and diversified processing is carried out to avoid the large errors caused by the single processing method in the traditional technology. Three conditions are judged, one is One case is that the angle formed is within the range of acute angles, one case is that the angle formed is a right angle, and one case is that the angle formed is within the range of obtuse angles, and the sound quality influencing feature data in these three cases are respectively extracted to extract the influence degree factor of the simulated voice information to be evaluated, so as to predict the voice information processing plan according to the extracted influence degree factor, so as to obtain the processed voice information, and perform keyword extraction on the processed voice information, and compare the extracted integrated keyword information with the original keyword information to evaluate the content completeness of the simulated voice information to be evaluated. Through the above processing method, the accuracy and reliability of the evaluation results can be improved, and the data of the related environmental influencing factors are analyzed in a differentiated and diversified manner, thereby improving the rigor of the entire processing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flowchart of the steps of the method for testing the in-vehicle call quality mainly embodied in this embodiment.

[0016] Figure 2 It is a structural block diagram of the vehicle call quality testing device mainly embodied in this embodiment.

[0017] Figure 3 This is a functional block diagram of the electronic device provided by the present invention.

[0018] Figure 4 A functional block diagram of a computer-readable storage device provided by the present invention. DETAILED DESCRIPTION

[0019] The present invention is further described in detail below in conjunction with the following examples.

[0020] Reference Figure 1 , a method for testing the in-vehicle call quality, the method comprising the following steps: S1. Perform call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, preset call environment wind condition information, and count the angle value to be measured formed between the ambient wind direction in the call environment wind condition information and the target vehicle direction as well as the sound quality impact characteristic data.

[0021] S2. The influence factors of the sound quality influence characteristic data under three conditions are statistically analyzed when the measured angle value is in the acute angle range, the right angle range or the obtuse angle range to obtain the first influence factor one, the first influence factor two and the first influence factor three. When the obtained target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, the second influence factor is statistically obtained.

[0022] S3. Train a simulation prediction scheme model, and predict pretreatment scheme one, pretreatment scheme two, and pretreatment scheme three according to the first influence factor one, the first influence factor two, and the first influence factor three.

[0023] S4. According to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3, the keyword information after the sound quality adjustment of the simulated speech information to be evaluated is extracted and integrated to obtain the first keyword information to be tested.

[0024] S5. Based on the second influence factor, preprocessing scheme four is predicted. According to preprocessing scheme four, the keyword information after sound quality adjustment of the simulated voice information to be evaluated is extracted to obtain the second keyword information to be tested. Based on the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information belonging to the original voice information, the content completeness of the simulated voice information to be evaluated is evaluated.

[0025] Specifically, by preliminarily performing call simulation processing on the original voice information, and in this simulated call processing process, there is a user driving a vehicle, and it is affected by the interference of external environmental factors. In order to reduce the misjudgment of keyword extraction in the voice information after the simulated call, the external environment in which the simulated call is located is analyzed in detail, and the wind condition information of the call environment is preset, which mainly includes the noise influence of wind speed and wind direction in the external environment on the voice during the call, and the angle value formed between the wind direction and the target vehicle direction is differentiated. Because the different angle values ​​have different noise interference degrees on the voice call, a more comprehensive and diversified processing is performed to avoid the large errors caused by the single processing method in the traditional technology, in which three conditions are judged. In one case, the angle formed is within the acute angle range, in one case, the angle formed is a right angle, and in one case, the angle formed is within the obtuse angle range. The sound quality influencing feature data in these three cases are respectively extracted to extract the influence degree factor of the simulated voice information to be evaluated, so as to predict the voice information processing scheme according to the extracted influence degree factor, so as to obtain the processed voice information, and perform keyword extraction on the processed voice information, and compare the extracted integrated keyword information with the original keyword information to evaluate the content completeness of the simulated voice information to be evaluated. Through the above processing method, the accuracy and reliability of the evaluation results can be improved, and the data of the correlation environmental influencing factors are analyzed in a differentiated and diversified manner, thereby improving the rigor of the entire processing process.

[0026] The specific step S1 includes the following sub-steps: The original voice information that needs to be processed for subsequent calls is obtained, and keyword information is extracted from the original voice information to obtain original keyword information.

[0027] The wind condition information of the call environment is preset, and according to the wind condition information of the call environment, a call simulation is performed on the original voice information to obtain the simulated voice information to be evaluated.

[0028] A speed condition interval value is preset to obtain a target speed of the user's vehicle. When the target speed is within the speed condition interval value, the target direction of the user's vehicle is counted.

[0029] The angle data formed between the ambient wind direction in the call environment wind condition information and the target vehicle direction are counted to obtain the angle value to be measured.

[0030] Obtain the degree of window opening of the user's vehicle and count the ambient wind speed in the call environment wind condition information.

[0031] The window opening degree, ambient wind speed and target vehicle speed are combined into the sound quality impact characteristic data.

[0032] Specifically, such as the wind condition information of the call environment (for example, the user's vehicle is making a call while driving, and the voice of the call is interfered by external noise, for example, the window of the user's vehicle is opened while driving, then the wind speed of the external environment itself will have a greater noise impact on the voice, which needs to be considered), the preset vehicle speed condition interval value (because the speed of the user's vehicle is within a certain interval value, the impact of the wind speed in the external environment can be ignored: the wind speed value is small at this time, if the speed of the user's vehicle is fast, then the wind speed value in the external environment will be small. There will be greater wind noise interference, so it is necessary to make a differentiated judgment on the vehicle speed value. If it is V1-V3, V1 is the minimum critical value. If the vehicle speed reaches V1, the wind speed interference in the external environment needs to be considered), target vehicle speed (if it is V2), target vehicle direction (if it is y, and y is the vertical axis, and the horizontal x-axis is perpendicular to the y-axis, so as to judge the angle value formed between the ambient wind direction and the target vehicle direction in the subsequent call environment wind condition information), the angle value to be measured (that is, the angle value formed by the formed x and y axes is judged, if it is a).

[0033] The specific step S2 includes the following sub-steps: When the angle value to be measured is within the acute angle range, a first influence factor 1 of the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data.

[0034] When the angle value to be measured is at a right angle, a first influence factor 2 of the influence on the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data.

[0035] When the angle value to be measured is within the obtuse angle range, the first influence factor three of the influence on the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data.

[0036] When the target vehicle speed is less than the minimum value in the vehicle speed condition interval value, a second influence factor of the sound quality of the original voice information during the call processing is extracted according to the target vehicle speed and the window opening degree.

[0037] Specifically, when the angle value to be measured is within the acute angle range (such as the ambient wind direction is opposite to the y-axis direction, it can be judged that the ambient wind direction is the same as the wind speed direction generated by the user's vehicle, so the degree of wind noise interference is most likely to be small, then the angle value of a is an acute angle), the first impact factor one (that is, according to the regular statistics of the call simulation processing record data in the historical period, such as the change value of the voice noise data generated under the same circumstances analyzed in multiple historical periods, that is, the angle value to be measured and the various feature item data in the sound quality impact feature data are respectively calculated with the generated noise data, and the multiple ratios calculated are averaged to obtain a mean, which is the first impact factor one, if it is b1), when the angle value to be measured is at a right angle (the ambient wind direction is perpendicular to the y-axis direction, that is, parallel to the x-axis, then the interference noise caused under this condition needs to be considered, and the degree is relatively large), the first impact factor Factor two (same explanation as the first influence factor one, no further explanation here), when the target vehicle speed is less than the minimum value in the vehicle speed condition interval value (such as the ambient wind direction is the same as the y-axis direction, it can be judged that the ambient wind direction is opposite to the wind speed direction generated by the user's vehicle, so the degree of wind noise interference is most likely to be the largest, then the angle value of a is an obtuse angle), the first influence factor three (same explanation as the first influence factor one, no further explanation here), when the target vehicle speed is less than the minimum value in the vehicle speed condition interval value (if the target vehicle speed is V0, the limiting conditions for generating a measurable ambient wind condition impact value are not met, at this time, what is considered is the wind noise impact value caused by the user's vehicle's own conditions during driving), the second influence factor (same explanation as the first influence factor one, the difference is that the analysis factors are different, here only the target vehicle speed and the window opening degree are analyzed).

[0038] The specific step S3 includes the following sub-steps: The historical call feature information of the target user recorded in the historical period and the sound quality processing solution set corresponding to the historical call feature information are obtained, and a simulation prediction solution model is trained based on the historical call feature information and the sound quality processing solution set.

[0039] The first influence factor 1 and the simulated speech information to be evaluated are input into the simulation prediction scheme model for testing to obtain the preprocessing scheme 1.

[0040] The first influence factor 2 and the simulated speech information to be evaluated are input into the simulation prediction scheme model for testing to obtain the preprocessing scheme 2.

[0041] The first influence factor three and the simulated speech information to be evaluated are input into the simulation prediction scheme model for testing to obtain the preprocessing scheme three.

[0042] Specifically, such as the simulation prediction scheme model (the historical call feature information here includes the call environment wind condition information, target vehicle speed, target vehicle direction, window opening degree, and the angle value to be measured mentioned above, the sound quality processing scheme set refers to the call voice information adjustment processing scheme corresponding to each historical call feature information situation: such as strengthening or lowering the tone of a certain voice and other similar voice tone processing, so as to achieve noise removal or weakening processing, simulation prediction scheme model: a plurality of historical call feature information and corresponding voice information adjustment processing schemes are matched and constructed in a table to facilitate the matching of subsequent feature information, for example, y=Ax+Bz+Cj, where A refers to the target vehicle speed, B refers to the target vehicle direction, C refers to the window opening degree, and x, z, and j refer to their respective effects. The difference rate of the impact factor data, y refers to the call voice information adjustment processing scheme, here only some of the impact factor data are used as examples, and the same can be said), pre-processing scheme one (the first impact factor one and the simulated voice information to be evaluated, where the first impact factor one can be A, and the simulated voice information to be evaluated can be a matching restriction condition, such as performing subsequent feature information matching from some historical call voice information that is very similar to the simulated voice information conditions to be evaluated. If the matchable call voice information adjustment processing schemes are F1, F2, and F3, the final decision on which scheme to select is determined by the first impact factor one, if it is F1), pre-processing scheme two and pre-processing scheme three (both have the same explanation as pre-processing scheme one, no further explanation is given here, if they are F2 and F3 respectively).

[0043] The specific step S4 includes the following sub-steps: According to preprocessing scheme one, preprocessing scheme two and preprocessing scheme three, the voice tones of different segments of the simulated voice information to be evaluated are adjusted to obtain first preprocessed voice information one, first preprocessed voice information two and first preprocessed voice information three.

[0044] The first preprocessed voice information one, the first preprocessed voice information two and the first preprocessed voice information three are respectively converted into text to obtain the first preprocessed text one, the first preprocessed text two and the first preprocessed text three.

[0045] The keyword information extraction is performed on the first preprocessed text one, the first preprocessed text two and the first preprocessed text three respectively, and the first keyword information to be tested is obtained by integrating them.

[0046] Specifically, such as the first preprocessed voice information one, the first preprocessed voice information two and the first preprocessed voice information three (if the simulated voice information to be evaluated is divided into three noisy voice content segments: n1, n2, n3, the wind noise contained in these three segments is different, if the first preprocessed voice information one: is to perform tone enhancement processing on the n1 content segment to obtain the N1 content voice segment, the first preprocessed voice information two: is to perform tone enhancement processing on the n2 content segment to obtain the N2 content voice segment, the first preprocessed voice information three: is to perform tone enhancement processing on the n3 content segment to obtain the N3 content voice segment), the first preprocessed text one, the first preprocessed text two and the first preprocessed text three (if they are W1, W2, W3 respectively), the first keyword information to be tested (wherein the keyword information extracted from W1 is: g1, g2, the keyword information extracted from W2 is: g0, g2, g3, the keyword information extracted from W3 is: g3, g4, then the final total extracted keyword information is g0, g1, g2, g3, g4).

[0047] The specific step S5 includes the following sub-steps: The second impact factor and the simulated speech information to be evaluated are input into the simulation prediction scheme model for testing to obtain preprocessing scheme four.

[0048] According to the fourth preprocessing scheme, the voice tones of different segments of the simulated speech information to be evaluated are adjusted to obtain the second preprocessed speech information.

[0049] The second preprocessed voice information is converted into text to obtain a second preprocessed text, and the second preprocessed text is extracted into keyword information to obtain second keyword information to be tested.

[0050] The first keyword information to be tested or the second keyword information to be tested is compared with the original keyword information to obtain a comparison result, and the content completeness of the simulated speech information to be evaluated is evaluated according to the comparison result.

[0051] Specifically, such as preprocessing scheme four, the second preprocessed voice information (the explanations are the same as those of preprocessing scheme one and the first preprocessed voice information one, and no further explanation is given here), the second preprocessed text (if it is W4), the second keyword information to be tested (if it is g0, g1, g3, g5, g6), the original keyword information (if it is g0, g1, g2, g3, g4, g5, g6, g7), the comparison results (there are five identical keyword information), and the evaluation of the content completeness of the simulated voice information to be evaluated (that is, the completeness is 62.5%).

[0052] The vehicle call quality testing device, by applying the vehicle call quality testing method as described above, includes a correlation influence data statistics module, a condition influence factor extraction module, a sound quality processing solution prediction module, a keyword extraction module and a content integrity evaluation module, referring to Figure 2 , the associated impact data statistics module is used to perform call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, the call environment wind condition information is preset, and the angle value to be measured formed between the environmental wind direction and the target vehicle direction in the call environment wind condition information and the sound quality impact characteristic data are counted; the conditional impact factor extraction module is used to perform statistics on the impact degree factors of the sound quality impact characteristic data under three conditions when the measured angle value is in the acute angle range, the right angle range or the obtuse angle range to obtain the first impact degree factor one, the first impact degree factor two and the first impact degree factor three; when the acquired target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, the second impact degree factor is obtained by statistics; the sound quality processing solution prediction module is used to train a simulation prediction solution model, and according to the first impact degree factor 1. Preprocessing scheme 1, preprocessing scheme 2 and preprocessing scheme 3 are predicted by the first influence factor 2 and the first influence factor 3; the keyword extraction module extracts and integrates the keyword information after the sound quality adjustment of the simulated voice information to be evaluated according to the preprocessing scheme 1, preprocessing scheme 2 and preprocessing scheme 3 to obtain the first keyword information to be tested; the content integrity evaluation module predicts preprocessing scheme 4 according to the second influence factor, and extracts the keyword information after the sound quality adjustment of the simulated voice information to be evaluated according to the preprocessing scheme 4 to obtain the second keyword information to be tested; the content integrity of the simulated voice information to be evaluated is evaluated according to the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information belonging to the original voice information.

[0053] Corresponds to Figure 1 The present disclosure also provides an electronic device 500, such as Figure 3 FIG. 5 is a schematic diagram of the structure of an electronic device 500 provided in an embodiment of the present disclosure, including: The processor 510, the memory 520, and the computer program 511; the memory 520 is used to store execution instructions, including the computer program 511, and is used to temporarily store the operation data in the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 so that the processor 510 executes Figure 1 Steps of the test method for in-vehicle call quality.

[0054] The present disclosure also provides a computer-readable storage medium 600. Figure 4As shown, the computer-readable storage medium stores a computer program 611, and when the computer program 611 is executed by the processor 510, the steps of the method for testing the in-vehicle call quality described in the above method embodiment are executed.

[0055] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for testing the quality of in-vehicle calls, characterized in that: The following steps are involved: Perform call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, preset call environment wind condition information, and count the angle value to be measured formed between the ambient wind direction and the target vehicle direction in the call environment wind condition information and the sound quality impact characteristic data; When the angle value to be measured is in an acute angle range, a right angle range or an obtuse angle range, the influence degree factors of the sound quality influence characteristic data under three conditions are statistically obtained to obtain a first influence degree factor 1, a first influence degree factor 2 and a first influence degree factor 3; when the obtained target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, the second influence degree factor is statistically obtained; A simulation prediction scheme model is trained to predict pretreatment scheme 1, pretreatment scheme 2 and pretreatment scheme 3 according to the first influence degree factor 1, the first influence degree factor 2 and the first influence degree factor 3; According to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3, the keyword information after the sound quality adjustment of the simulated speech information to be evaluated is extracted and integrated to obtain the first keyword information to be tested; According to the second impact factor, preprocessing scheme four is predicted. According to preprocessing scheme four, the keyword information after sound quality adjustment of the simulated voice information to be evaluated is extracted to obtain the second keyword information to be tested. According to the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information belonging to the original voice information, the content completeness of the simulated voice information to be evaluated is evaluated.

2. The method for testing the in-vehicle call quality according to claim 1, characterized in that: The steps of performing call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, presetting call environment wind condition information, and counting the angle value to be measured formed between the ambient wind direction and the target vehicle direction in the call environment wind condition information and the sound quality impact characteristic data are specifically as follows: Acquire original voice information that needs to be processed for subsequent calls, and extract keyword information from the original voice information to obtain original keyword information; Preset call environment wind condition information, and perform call simulation on original voice information according to the call environment wind condition information to obtain simulated voice information to be evaluated; A vehicle speed condition interval value is preset to obtain a target vehicle speed of the user's vehicle, and when the target vehicle speed is within the vehicle speed condition interval value, the target vehicle direction of the user's vehicle is counted; Counting the angle data between the ambient wind direction in the call environment wind condition information and the target vehicle direction to obtain the angle value to be measured; Obtain the degree of opening of the windows of the user's vehicle and count the ambient wind speed in the call environment wind condition information; The window opening degree, ambient wind speed and target vehicle speed are combined into sound quality impact characteristic data.

3. The method for testing the in-vehicle call quality according to claim 2, characterized in that: The step of performing statistics on the influence degree factors of the sound quality influence characteristic data under three conditions when the angle value to be measured is in the acute angle range, the right angle range or the obtuse angle range to obtain the first influence degree factor 1, the first influence degree factor 2 and the first influence degree factor 3, and when the obtained target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, the step of obtaining the second influence degree factor by statistics is specifically as follows: When the angle value to be measured is within the acute angle range, extracting a first influence factor 1 of the sound quality of the original voice information during the call processing according to the angle value to be measured and the sound quality influence characteristic data; When the angle value to be measured is at a right angle, a first influence factor 2 of the influence of the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data; When the angle value to be measured is within the obtuse angle range, the first influence factor three of the sound quality of the original voice information during the call processing is extracted according to the angle value to be measured and the sound quality influence characteristic data; When the target vehicle speed is less than the minimum value in the vehicle speed condition interval value, a second influence degree factor of the sound quality of the original voice information during the call processing is extracted according to the target vehicle speed and the window opening degree.

4. The method for testing the in-vehicle call quality according to claim 3, characterized in that: The steps of training a simulation prediction scheme model and predicting pretreatment scheme one, pretreatment scheme two and pretreatment scheme three according to the first influence degree factor one, the first influence degree factor two and the first influence degree factor three are specifically as follows: Acquire historical call feature information of the target user recorded in a historical period and a sound quality processing solution set corresponding to the historical call feature information, and train a simulation prediction solution model based on the historical call feature information and the sound quality processing solution set; Inputting the first impact factor 1 and the simulated speech information to be evaluated into a simulation prediction scheme model for testing to obtain a preprocessing scheme 1; Inputting the first impact factor 2 and the simulated speech information to be evaluated into the simulation prediction scheme model for testing to obtain the preprocessing scheme 2; The first impact factor three and the simulated speech information to be evaluated are input into the simulation prediction scheme model for testing to obtain the preprocessing scheme three.

5. The method for testing the in-vehicle call quality according to claim 4, characterized in that: According to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3, the steps of extracting and integrating the keyword information after the sound quality adjustment of the simulated speech information to be evaluated to obtain the first keyword information to be tested are specifically: According to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3, the voice tones of different segments of the simulated voice information to be evaluated are adjusted to obtain the first preprocessed voice information 1, the first preprocessed voice information 2 and the first preprocessed voice information 3; Performing text conversion on the first preprocessed voice information one, the first preprocessed voice information two and the first preprocessed voice information three respectively to obtain a first preprocessed text one, a first preprocessed text two and a first preprocessed text three; The first preprocessed text one, the first preprocessed text two and the first preprocessed text three are respectively integrated to extract keyword information to obtain first keyword information to be tested.

6. The method for testing the in-vehicle call quality according to claim 5, characterized in that: According to the second impact factor, a preprocessing scheme 4 is predicted. According to the preprocessing scheme 4, the keyword information after the sound quality adjustment of the simulated voice information to be evaluated is extracted to obtain the second keyword information to be tested. According to the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information to which the original voice information belongs, the steps of evaluating the content integrity of the simulated voice information to be evaluated are specifically as follows: Inputting the second impact factor and the simulated speech information to be evaluated into the simulation prediction scheme model for testing to obtain preprocessing scheme 4; According to the fourth preprocessing scheme, the voice tones of different segments of the simulated speech information to be evaluated are adjusted to obtain second preprocessed speech information; Performing text conversion on the second preprocessed voice information to obtain a second preprocessed text, and extracting keyword information from the second preprocessed text to obtain second keyword information to be tested; The first keyword information to be tested or the second keyword information to be tested is compared with the original keyword information to obtain a comparison result, and the content completeness of the simulated voice information to be evaluated is evaluated according to the comparison result.

7. A device for testing the quality of in-vehicle calls, characterized in that: The device is used to implement the vehicle call quality testing method according to any one of claims 1 to 6, comprising: The associated impact data statistics module is used to perform call simulation on the acquired original voice information to obtain simulated voice information to be evaluated, preset call environment wind condition information, and count the angle value to be measured formed between the ambient wind direction in the call environment wind condition information and the target vehicle direction, as well as the sound quality impact characteristic data; A conditional influence factor extraction module is used to obtain the first influence factor 1, the first influence factor 2 and the first influence factor 3 by performing statistics on the influence factors of the sound quality influence characteristic data under three conditions when the angle value to be measured is in the acute angle range, the right angle range or the obtuse angle range, and when the obtained target vehicle speed is less than the minimum value in the preset vehicle speed condition interval value, obtain the second influence factor by statistics; A sound quality processing scheme prediction module is used to train a simulation prediction scheme model, and predict pre-processing scheme one, pre-processing scheme two and pre-processing scheme three according to the first influence factor one, the first influence factor two and the first influence factor three; A keyword extraction module is used to extract and integrate the keyword information after the sound quality adjustment of the simulated speech information to be evaluated according to the preprocessing scheme 1, the preprocessing scheme 2 and the preprocessing scheme 3 to obtain the first keyword information to be tested; The content integrity evaluation module is used to predict preprocessing scheme four according to the second influence factor, extract the keyword information after sound quality adjustment of the simulated voice information to be evaluated according to the preprocessing scheme four, and evaluate the content integrity of the simulated voice information to be evaluated according to the comparison result of the first keyword information to be tested or the second keyword information to be tested with the original keyword information belonging to the original voice information.

8. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the method for testing the in-vehicle call quality as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that: The storage medium stores a computer software program, which, when executed by a processor, implements the method for testing the in-vehicle call quality as described in any one of claims 1-6.

Images