A quality testing system and method based on sound

Abstract

The present application discloses a sound-based quality testing system and method, which relates to the technical field of audio quality testing. The method includes judging whether there are at least three identical results in the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result when the first, second, third and fourth impact feature information are the same, and the other two impact feature information are the same, when the first, second, third and fourth impact feature information are the same, and when the first, second, third and fourth impact feature information are different, the audio quality test results under these four different conditions are analyzed for differentiation, and an audio qualified test result or an audio unqualified test result is output. The sound-based quality testing system and method provided by the present invention can improve the reliability of audio quality test results.

Description

Technical Field

[0001] The present application relates to the technical field of audio quality testing, and in particular to a sound-based quality testing system and method. Background Art

[0002] Audio quality assessment is one of the basic technologies in audio processing and is widely used in the field of audio processing. Therefore, how to evaluate the quality of audio is of great significance.

[0003] In the prior art, audio quality assessment does not take into account the different degrees of loss of audio signals caused by various factors in the environment, resulting in certain errors in the final test data information. In addition, diversity analysis is not performed under various situations, which leads to misjudgment of the final test results and reduces the efficiency of audio quality testing. Summary of the invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides an audio-based quality testing system and method.

[0005] In a first aspect, the present application provides a sound-based quality testing method, the method comprising:

[0006] Acquire audio emission position information, preset a circular range test area, select a first original test position point, a second original test position point, a third original test position point and a fourth original test position point from the circular range test area, and obtain first received audio feature data, second received audio feature data, third received audio feature data and fourth received audio feature data according to statistics of the four position points;

[0007] Obtain historical audio signals, historical environmental impact feature information, and historical reception test signals tested in historical periods, and train a signal loss prediction model;

[0008] The circular range test area is divided into a first area to be tested, a second area to be tested, a third area to be tested, and a fourth area to be tested, and first influencing characteristic information, second influencing characteristic information, third influencing characteristic information, and fourth influencing characteristic information are detected and obtained, and a corresponding first signal loss prediction result, a second signal loss prediction result, a third signal loss prediction result, and a fourth signal loss prediction result are predicted through a signal loss prediction model test;

[0009] According to the first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result, performing data-corrected audio quality evaluation on the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data to obtain a first evaluation result, a second evaluation result, a third evaluation result and a fourth evaluation result;

[0010] The comparison results between the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are analyzed to be the same quantity, and the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result are compared and judged accordingly, and then the audio qualified test result or the audio unqualified test result is output.

[0011] Preferably, when the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are the same, it is determined whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, and an audio qualified test result or an audio unqualified test result is output;

[0012] When two of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are the same, and the other two influencing feature information are the same, select the first category evaluation result one, the first category evaluation result two, the second category evaluation result one and the second category evaluation result two from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, judge whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same, and output an audio qualified test result or an audio unqualified test result;

[0013] When three of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are the same, selecting a third category evaluation result one, a third category evaluation result two and a third category evaluation result three from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, judging whether the third category evaluation result one, the third category evaluation result two and the third category evaluation result three are all the same, and outputting an audio qualified test result or an audio unqualified test result;

[0014] When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are all different, the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result are statistically obtained, and it is determined whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result, and the audio qualified test result or the audio unqualified test result is output.

[0015] Preferably, the audio emission position information is obtained, and a circular range test area is preset according to the audio emission position information, wherein the circular range test area is centered on the audio emission position information;

[0016] Select four audio test receiving position points from the edge of the circular range test area, and output a first original test position point, a second original test position point, a third original test position point, and a fourth original test position point, wherein the first original test position point, the second original test position point, the third original test position point, and the fourth original test position point are all in an equidistant relationship;

[0017] According to the distance between the audio emission position information and the first original test position point, audio emission parameter data is preset, and based on the audio emission parameter data, the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data received by the first original test position point, the second original test position point, the third original test position point and the fourth original test position point are respectively counted.

[0018] Preferably, historical audio emission signals, historical environmental impact feature information, and historical reception test signals of historical period tests are obtained;

[0019] Extracting a historical valid reception test signal for a valid test from the historical reception test signal;

[0020] According to the historical valid reception test signal, respectively extract the valid record information from the historical audio emission signal and the historical environmental impact characteristic information, and output the historical valid audio emission signal and the historical valid environmental impact characteristic information;

[0021] Counting the difference between the historical valid audio transmission signal and the historical valid receiving test signal to obtain the historical preprocessing effective difference;

[0022] A signal loss prediction model is trained based on the historical preprocessing effective difference and the historical effective environmental impact feature information.

[0023] Preferably, according to the first original test position point, the second original test position point, the third original test position point and the fourth original test position point, the circular range test area is divided into four equal spatial areas, and the first area to be tested, the second area to be tested, the third area to be tested and the fourth area to be tested are output, wherein the first area to be tested includes the first original test position point, the second area to be tested includes the second original test position point, the third area to be tested includes the third original test position point, and the fourth area to be tested includes the fourth original test position point;

[0024] Detecting environmental impact characteristic information in a first test area to obtain first impact characteristic information, and inputting the first impact characteristic information and audio emission parameter data into a signal loss prediction model for testing to obtain a first signal loss prediction result;

[0025] Detecting the environmental impact characteristic information in the second test area to obtain second impact characteristic information, and inputting the second impact characteristic information and the audio emission parameter data into the signal loss prediction model for testing to obtain a second signal loss prediction result;

[0026] Detecting the environmental impact characteristic information in the third test area to obtain third impact characteristic information, and inputting the third impact characteristic information and the audio emission parameter data into the signal loss prediction model for testing to obtain a third signal loss prediction result;

[0027] The environmental impact characteristic information in the fourth test area is detected to obtain fourth impact characteristic information, and the fourth impact characteristic information and audio emission parameter data are input into the signal loss prediction model for testing to obtain a fourth signal loss prediction result.

[0028] Preferably, according to the first signal loss prediction result, the first received audio feature data is corrected to obtain a first received audio correction result;

[0029] According to the second signal loss prediction result, the second received audio feature data is corrected to obtain a second received audio correction result;

[0030] According to the third signal loss prediction result, the third received audio feature data is corrected to obtain a third received audio correction result;

[0031] According to the fourth signal loss prediction result, the fourth received audio feature data is corrected to obtain a fourth received audio correction result;

[0032] Performing audio quality evaluation on the first received audio correction result, the second received audio correction result, the third received audio correction result, and the fourth received audio correction result, respectively, to obtain a first evaluation result, a second evaluation result, a third evaluation result, and a fourth evaluation result, where the evaluation results include qualified or unqualified;

[0033] When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same, determine whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result. If so, the final audio test result is determined to be qualified and the audio qualified test result is output; if not, the final audio test result is determined to be unqualified and the audio unqualified test result is output. The same result refers to qualified.

[0034] Preferably, when two of the first impact characteristic information, the second impact characteristic information, the third impact characteristic information and the fourth impact characteristic information are the same, the first type of impact characteristic information 1 and the first type of impact characteristic information 2 are output, and when the other two types of impact characteristic information are the same, the second type of impact characteristic information 1 and the second type of impact characteristic information 2 are output;

[0035] According to the first type of impact feature information 1 and the first type of impact feature information 2, select the first type of assessment result 1 and the first type of assessment result 2 from the first assessment result, the second assessment result, the third assessment result and the fourth assessment result;

[0036] According to the second-category impact feature information one and the second-category impact feature information two, correspondingly selecting the second-category assessment result one and the second-category assessment result two from the first assessment result, the second assessment result, the third assessment result, and the fourth assessment result;

[0037] Determine whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same. If the judgment results are both yes, then the final audio test result is determined to be qualified, and the audio qualified test result is output. If one of the judgment results is no, then the final audio test result is determined to be unqualified, and the audio unqualified test result is output.

[0038] Preferably, when three of the first impact characteristic information, the second impact characteristic information, the third impact characteristic information and the fourth impact characteristic information are the same, the impact characteristic information of the three same impact characteristic information is screened out to obtain third-category impact characteristic information one, third-category impact characteristic information two and third-category impact characteristic information three;

[0039] According to the third-category impact characteristic information one, the third-category impact characteristic information two, and the third-category impact characteristic information three, correspondingly selecting the third-category assessment result one, the third-category assessment result two, and the third-category assessment result three from the first assessment result, the second assessment result, the third assessment result, and the fourth assessment result;

[0040] Determine whether the third category evaluation result one, the third category evaluation result two and the third category evaluation result three are all the same. If so, determine that the final audio test result is qualified and output the audio qualified test result. If not, determine that the final audio test result is unqualified and output the audio unqualified test result.

[0041] Preferably, when the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are all different, it is determined whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result. If so, it is determined that the final audio test result is qualified, and the audio qualified test result is output; if not, and the judgment result is that there are two identical results, a new test is performed and secondary test information is output;

[0042] According to the secondary test information, a first new evaluation result, a second new evaluation result, a third new evaluation result and a fourth new evaluation result are obtained by statistics;

[0043] Determine whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result. If so, determine that the final audio test result is qualified and output the qualified audio test result. If not, determine whether there are still two identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result. If so, determine that the final audio test result is unqualified and output the unqualified audio test result.

[0044] In a second aspect, a sound-based quality testing system includes:

[0045] A test position point selection unit is used to obtain audio emission position information, preset a circular range test area, select a first original test position point, a second original test position point, a third original test position point and a fourth original test position point from the circular range test area, and obtain first received audio feature data, second received audio feature data, third received audio feature data and fourth received audio feature data according to statistics of the four position points;

[0046] A prediction model training unit, used to obtain historical audio transmission signals, historical environmental impact feature information and historical reception test signals tested in historical periods, and train a signal loss prediction model;

[0047] A signal loss prediction unit is used to divide the circular range test area into a first test area, a second test area, a third test area and a fourth test area, and detect the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information, and predict the corresponding first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result through a signal loss prediction model test;

[0048] A test quality evaluation unit 1 is used to perform a data-corrected audio quality evaluation on the first received audio feature data, the second received audio feature data, the third received audio feature data, and the fourth received audio feature data according to the first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result, and the fourth signal loss prediction result to obtain a first evaluation result, a second evaluation result, a third evaluation result, and a fourth evaluation result;

[0049] Test quality assessment unit two is used to analyze the comparison results of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information to find that the results are the same, and to output an audio qualified test result or an audio unqualified test result after making corresponding comparisons and judgments on the first assessment result, the second assessment result, the third assessment result and the fourth assessment result.

[0050] Compared with the prior art, the present invention has the following characteristics and beneficial effects:

[0051] By selecting four original test position points in a preset circular range test area, and the four original test position points are equidistant, and statistically obtaining four received audio feature data received by the four original test position points, and these four received audio feature data are very likely to be different, in order to improve the accuracy of the audio test quality evaluation results, it is necessary to evaluate the degree of signal loss caused by environmental factors in the process of the audio signal being sent from the audio position information to the four original test position points, so as to achieve data correction of the four received audio feature data and a preliminary evaluation of the corrected audio quality to obtain four evaluation results, reduce the error probability of subsequent data processing, and in order to analyze the signal loss in various situations, a diverse and more comprehensive analysis is conducted, that is, the circular The range test area is divided into four areas to be tested, and the environmental impact characteristic information in the four areas to be tested are counted respectively, and four situations are judged on the four detected environmental impact characteristic information, namely, the four environmental impact characteristic information are the same, three of the four environmental impact characteristic information are the same, two of the four environmental impact characteristic information are the same and the other two are the same, and the four environmental impact characteristic information are different. The final result of the judgment is that the final audio test result is qualified or the final audio test result is unqualified. Through the above-mentioned diversity processing process, the analysis of the mutual influence of various information is enhanced, and the utilization of resources and the accuracy of the audio quality test results are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flowchart of a sound-based quality testing method mainly embodied in this embodiment.

[0053] Figure 2 It is a structural block diagram of an audio-based quality testing system mainly embodied in this embodiment. DETAILED DESCRIPTION

[0054] The present invention is further described in detail below in conjunction with the following examples.

[0055] Reference Figure 1 , a sound-based quality testing system and method, the method comprising the following steps:

[0056] S1. Obtain the audio emission position information, preset a circular range test area, select the first original test position point, the second original test position point, the third original test position point and the fourth original test position point from the circular range test area, and obtain the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data based on statistics of the four position points.

[0057] S2. Obtain historical audio transmission signals, historical environmental impact feature information, and historical reception test signals tested in historical periods, and train a signal loss prediction model.

[0058] S3. Divide the circular range test area into a first test area, a second test area, a third test area and a fourth test area, and detect the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information, and predict the corresponding first signal loss prediction result, second signal loss prediction result, third signal loss prediction result and fourth signal loss prediction result through signal loss prediction model testing.

[0059] S4. Based on the first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result, the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data are subjected to audio quality evaluation after data correction to obtain a first evaluation result, a second evaluation result, a third evaluation result and a fourth evaluation result.

[0060] S5. Analyze the comparison results of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information to find that they are the same quantity, and output an audio qualified test result or an audio unqualified test result after making corresponding comparisons and judgments on the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result.

[0061] The specific step S5 includes the following sub-steps:

[0062] When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same, determine whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, and output an audio qualified test result or an audio unqualified test result.

[0063] When two of the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same, and the other two influencing characteristic information are the same, select the first category evaluation result one, the first category evaluation result two, the second category evaluation result one and the second category evaluation result two from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, judge whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same, and output the audio qualified test result or the audio unqualified test result.

[0064] When three of the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same, select the third category evaluation result one, the third category evaluation result two and the third category evaluation result three from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, judge whether the third category evaluation result one, the third category evaluation result two and the third category evaluation result three are all the same, and output the audio qualified test result or the audio unqualified test result.

[0065] When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are all different, the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result are statistically obtained, and it is determined whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result, and the audio qualified test result or the audio unqualified test result is output.

[0066] Specifically, four original test position points are selected within a preset circular range test area, and the four original test position points are equidistant, and four received audio feature data received by the four original test position points are statistically obtained, and these four received audio feature data are very likely to be different. In order to improve the accuracy of the audio test quality evaluation results, it is necessary to evaluate the degree of signal loss caused by environmental factors in the process of the audio signal being sent from the audio position information to the four original test position points, so as to achieve data correction of the four received audio feature data and a preliminary evaluation of the corrected audio quality to obtain four evaluation results, thereby reducing the error probability of subsequent data processing. In order to analyze the signal loss in various situations, a diverse and more comprehensive analysis is conducted, that is, The circular range test area is divided into four areas to be tested, and the environmental impact characteristic information in the four areas to be tested is counted respectively, and four situations are judged on the four detected environmental impact characteristic information, namely, the four environmental impact characteristic information are the same, three of the four environmental impact characteristic information are the same, two of the four environmental impact characteristic information are the same and the other two are the same, and the four environmental impact characteristic information are different. The final result of the judgment is that the final audio test result is qualified or the final audio test result is unqualified. Through the above-mentioned diversity processing process, the analysis of the mutual influence between the various information is enhanced, and the utilization of resources and the accuracy of the audio quality test results are improved.

[0067] The specific step S1 includes the following sub-steps:

[0068] The audio source location information is obtained, and a circular range test area is preset based on the audio source location information. The circular range test area is centered on the audio source location information.

[0069] Four audio test receiving position points are selected from the edge of the circular range test area, and the first original test position point, the second original test position point, the third original test position point and the fourth original test position point are output. The first original test position point, the second original test position point, the third original test position point and the fourth original test position point are all equidistant.

[0070] According to the distance between the audio emission position information and the first original test position point, the audio emission parameter data is preset, and based on the audio emission parameter data, the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data received by the first original test position point, the second original test position point, the third original test position point and the fourth original test position point are respectively counted.

[0071] Specifically, such as the circular range test area (which is set according to ideal conditions, if the radius distance is 5 meters, the ideal condition is that the audio signal detected at this distance is relatively good, and the surrounding is quiet and free of obstacles, etc.), the first original test position point, the second original test position point, the third original test position point and the fourth original test position point (if they are A1, A2, A3, A4, and the straight-line distances between A1-A2, A2-A3, A3-A4, and A1-A4 are equal), preset audio emission parameter data (parameter categories include multiple parameter data such as level, frequency response, total harmonic distortion plus noise, phase, signal-to-noise ratio, amplitude-frequency characteristics, etc., some of the data information is used as an example, If they are a0, b0, c0, d0, e0 respectively), the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data (that is, the actual signal data information received when the signal reaches the four original test positions after a certain distance of signal transmission, it is very likely that there is a difference between the audio parameter data emitted, if the first received audio feature data is a1, b1, c1, d1, e1, if the second received audio feature data is a2, b2, c2, d2, e2, if the third received audio feature data is a3, b3, c3, d3, e3, if the fourth received audio feature data is a4, b4, c4, d4, e4).

[0072] The specific step S2 includes the following sub-steps:

[0073] Obtain historical audio emission signals, historical environmental impact feature information, and historical reception test signals tested during historical periods.

[0074] A historical effective reception test signal for effective testing is extracted from the historical reception test signal.

[0075] According to the historical effective reception test signal, the effective record information is extracted from the historical audio emission signal and the historical environmental impact characteristic information respectively, and the historical effective audio emission signal and the historical effective environmental impact characteristic information are output.

[0076] The difference between the historical effective audio transmission signal and the historical effective receiving test signal is counted to obtain the historical pre-processing effective difference.

[0077] The signal loss prediction model is trained based on the historical preprocessing effective difference and the historical effective environmental impact feature information.

[0078] Specifically, such as historical audio signals (if b0, b1, b2, b3, b4, each represents a parameter of a genus, and the locations of historical tests are different, only one genus parameter is used here for illustration), historical environmental impact characteristic information (including multiple environmental impact characteristic information such as test distance, temperature, humidity, and atmospheric pressure. If the characteristic data of the environmental impact of the surrounding area of ​​the location corresponding to the b0 test is test distance H1, temperature H2, humidity H3, and atmospheric pressure H4, and the characteristic data of the environmental impact of the surrounding area of ​​the location corresponding to the b1 test is test distance M1, temperature M2, humidity M3, and atmospheric pressure M4), historical effective reception test signals (if corresponding to b0, b1, b2, b3, b4's respective effective received test signals are B0, B1, B2, B3, and B4), historical pre-processing effective differences (b0-B0, b1-B1, b2-B2, b3-B3, and b4-B4), and a signal loss prediction model (i.e., a pairing table is formed based on the proportional matching between the historical pre-processing effective differences and the historical effective environmental impact feature information, such as calculating the ratio between the b0-B0 difference and the test distance H1 as the loss ratio value, and so on. In order to improve the scientific analysis of the sustainable development and researchability of regular characteristics, it is necessary to perform repeated steps on multiple types of historical data, i.e., perform the same steps on b1-B1, b2-B2, b3-B3, and b4-B4 in turn).

[0079] The specific step S3 includes the following sub-steps:

[0080] According to the first original test position point, the second original test position point, the third original test position point and the fourth original test position point, the circular range test area is divided into four equal spatial areas, and the first area to be tested, the second area to be tested, the third area to be tested and the fourth area to be tested are output, wherein the first area to be tested includes the first original test position point, the second area to be tested includes the second original test position point, the third area to be tested includes the third original test position point, and the fourth area to be tested includes the fourth original test position point.

[0081] The environmental impact characteristic information in the first test area is detected to obtain first impact characteristic information, and the first impact characteristic information and audio emission parameter data are input into the signal loss prediction model for testing to obtain a first signal loss prediction result.

[0082] The environmental impact characteristic information in the second test area is detected to obtain second impact characteristic information, and the second impact characteristic information and audio emission parameter data are input into the signal loss prediction model for testing to obtain a second signal loss prediction result.

[0083] The environmental impact characteristic information in the third test area is detected to obtain third impact characteristic information, and the third impact characteristic information and audio emission parameter data are input into the signal loss prediction model for testing to obtain a third signal loss prediction result.

[0084] The environmental impact characteristic information in the fourth test area is detected to obtain fourth impact characteristic information, and the fourth impact characteristic information and the audio emission parameter data are input into the signal loss prediction model for testing to obtain a fourth signal loss prediction result.

[0085] Specifically, the first test area, the second test area, the third test area and the fourth test area (if they are O1, O2, O3, O4 respectively, and the areas covered by O1, O2, O3, O4 are equal, the circular range test area is divided into four equal parts, that is, the area is 6.25π), the first influencing characteristic information (if it is the test distance h1, temperature h2, humidity h3, atmospheric pressure h4), the first signal loss prediction result (that is, a0, b0, c0, d0, e0 and the test distance h1, temperature h2, humidity h3, atmospheric pressure h4 are input into the model for comparison with the originally formed The information matching table is matched to obtain the first signal loss prediction result: the same as the historical preprocessing effective difference mentioned above, if it is N0, N1, N2, N3, N4), the second signal loss prediction result (if it is P0, P1, P2, P3, P4), the third signal loss prediction result (if it is K0, K1, K2, K3, K4) and the fourth signal loss prediction result (if it is L0, L1, L2, L3, L4) (the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result have the same interpretation as the first signal loss prediction result, so no further explanation is given here).

[0086] The specific step S4 includes the following sub-steps:

[0087] According to the first signal loss prediction result, the first received audio feature data is corrected to obtain a first received audio correction result.

[0088] According to the second signal loss prediction result, the second received audio feature data is corrected to obtain a second received audio correction result.

[0089] According to the third signal loss prediction result, the third received audio feature data is corrected to obtain a third received audio correction result.

[0090] According to the fourth signal loss prediction result, the fourth received audio feature data is corrected to obtain a fourth received audio correction result.

[0091] Audio quality evaluation is performed on the first received audio correction result, the second received audio correction result, the third received audio correction result and the fourth received audio correction result to obtain a first evaluation result, a second evaluation result, a third evaluation result and a fourth evaluation result, respectively, and the evaluation results include qualified or unqualified.

[0092] When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same, determine whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result. If so, the final audio test result is determined to be qualified and the audio qualified test result is output; if not, the final audio test result is determined to be unqualified and the audio unqualified test result is output. The same result refers to qualified.

[0093] Specifically, the first received audio correction result is mainly used for subsequent illustration, such as the first received audio correction result (a1+N0, b1+N1, c1+N2, d1+N3, e1+N4), the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result (that is, according to the difference comparison of a1+N0, b1+N1, c1+N2, d1+N3, e1+N4, the difference comparison of the five categories of data is performed respectively with a0, b0, c0, d0, e0, and the difference comparison results of the five categories of data are obtained. The difference comparison results of the five categories of data are statistically analyzed to obtain a comprehensive difference comparison result. If it is W1, and the final audio quality evaluation of W1 is qualified, the first evaluation result is qualified, and the second evaluation result, the third evaluation result and the fourth evaluation result are based on this By analogy), when the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same (that is, the four influencing characteristic information are: test distance h1, temperature h2, humidity h3, atmospheric pressure h4, which means that the loss degree values ​​of the audio signal in the four test areas are relatively stable), it is determined whether there are at least three identical results in the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result (for example, if the first evaluation result, the second evaluation result and the third evaluation result are all W1), then it can be determined that the test results of the corresponding three original test positions are more accurate, that is, the probability of misjudgment is reduced (if only the first evaluation result and the second evaluation result are both W1, or only the first evaluation result is W1), then it means that the audio quality is indeed unqualified.

[0094] The specific step S5 includes the following sub-steps:

[0095] When two of the first impact characteristic information, the second impact characteristic information, the third impact characteristic information and the fourth impact characteristic information are the same, the first type of impact characteristic information one and the first type of impact characteristic information two are output, and when the other two types of impact characteristic information are the same, the second type of impact characteristic information one and the second type of impact characteristic information two are output.

[0096] According to the first type of impact characteristic information one and the first type of impact characteristic information two, the first type of evaluation result one and the first type of evaluation result two are correspondingly selected from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result.

[0097] According to the second type of impact characteristic information one and the second type of impact characteristic information two, the second type of evaluation result one and the second type of evaluation result two are correspondingly selected from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result.

[0098] Determine whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same. If the judgment results are both yes, then the final audio test result is determined to be qualified, and the audio qualified test result is output. If one of the judgment results is no, then the final audio test result is determined to be unqualified, and the audio unqualified test result is output.

[0099] Specifically, when two of the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information, and the fourth influencing characteristic information are the same (for example, the first influencing characteristic information and the second influencing characteristic information are both the test distance h1, the temperature h2, the humidity h3, and the atmospheric pressure h4, that is, the first influencing characteristic information is the first type of influencing characteristic information one and the second influencing characteristic information is the first type of influencing characteristic information two), the second type of influencing characteristic information one and the second type of influencing characteristic information two (that is, the third influencing characteristic information is the second type of influencing characteristic information one and the fourth influencing characteristic information is the second type of influencing characteristic information 2, and the environmental impact characteristic information possessed are test distance g1, temperature g2, humidity g3, and atmospheric pressure g4), the second type of evaluation result one and the second type of evaluation result two (i.e. corresponding to the first evaluation result and the second evaluation result, if both are W1 qualified), the second type of evaluation result one and the second type of evaluation result two (i.e. corresponding to the third evaluation result and the fourth evaluation result, if both are W2 qualified), then the final audio test result is determined to be qualified, (if corresponding to the third evaluation result and the fourth evaluation result, one is W2 and the other is W3, and at most one of them is qualified), then the final audio test result is determined to be unqualified.

[0100] The specific step S5 also includes the following sub-steps:

[0101] When three of the first impact characteristic information, the second impact characteristic information, the third impact characteristic information and the fourth impact characteristic information are the same, the impact characteristic information of which the three are the same is filtered out to obtain third category impact characteristic information one, third category impact characteristic information two and third category impact characteristic information three.

[0102] According to the third category of impact characteristic information one, the third category of impact characteristic information two and the third category of impact characteristic information three, the third category of assessment result one, the third category of assessment result two and the third category of assessment result three are selected accordingly from the first assessment result, the second assessment result, the third assessment result and the fourth assessment result.

[0103] Determine whether the third-category evaluation result one, the third-category evaluation result two and the third-category evaluation result three are all the same. If so, determine that the final audio test result is qualified and output the audio qualified test result. If not, determine that the final audio test result is unqualified and output the audio unqualified test result.

[0104] Specifically, for example, when three of the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same (the first influencing characteristic information, the second influencing characteristic information and the third influencing characteristic information are all test distance h1, temperature h2, humidity h3 and atmospheric pressure h4), the third type of influencing characteristic information one (corresponding to the first influencing characteristic information), the third type of influencing characteristic information two (corresponding to the second influencing characteristic information) and the third type of influencing characteristic information three (corresponding to the third influencing characteristic information), the third type of evaluation result one (corresponding to the first evaluation result), the third type of evaluation result two (corresponding to the second evaluation result) and the third type of evaluation result three (corresponding to the third evaluation result), if the third type of evaluation result one, the third type of evaluation result two and the third type of evaluation result three are all qualified as W1, then the final test result of the audio is determined to be qualified; if two of the results are qualified as W1 and one is unqualified as W2, or the three results are different and not all qualified, then the final test result of the audio is determined to be unqualified.

[0105] The specific step S5 also includes the following sub-steps:

[0106] When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are all different, determine whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result. If so, the final audio test result is determined to be qualified, and the audio qualified test result is output; if not, and the judgment result is that there are two identical results, then re-test and output secondary test information.

[0107] According to the secondary test information, a first new evaluation result, a second new evaluation result, a third new evaluation result and a fourth new evaluation result are statistically obtained.

[0108] Determine whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result. If so, determine that the final audio test result is qualified and output the qualified audio test result. If not, determine whether there are still two identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result. If so, determine that the final audio test result is unqualified and output the unqualified audio test result.

[0109] Specifically, when the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are all different (that is, the first influencing characteristic information is the test distance h1, temperature h2, humidity h3, atmospheric pressure h4, the second influencing characteristic information is the test distance x1, temperature x2, humidity x3, atmospheric pressure x4, the third influencing characteristic information is the test distance y1, temperature y2, humidity y3, atmospheric pressure y4 and the fourth influencing characteristic information is the test distance t1, temperature t2, humidity t3, atmospheric pressure t4, all of which are different), determine whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result (for example, the first evaluation result, the second evaluation result and the third evaluation result are all W1 qualified), then determine that the final audio test result is qualified; if two of the results are W1 qualified and one is W2 unqualified , or if the three results are different and not all qualified, it may be due to different environmental impact feature information, and a direct judgment cannot be made. It is necessary to continue the judgment and analysis to reduce the probability of misjudgment and improve the accuracy and reliability of the final result of the audio quality test, that is, re-test (repeat the analysis steps of the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result), and finally obtain the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result (if it is judged that the first new evaluation result, the second new evaluation result and the third new evaluation result are all W1 qualified), then the final audio test result is judged to be qualified. If the judgment is still the same as the above result: two of the results are W1 qualified and one is W2 unqualified, or the three results are different and not all qualified, then the final audio test result is judged to be unqualified.

[0110] A sound-based quality testing system, by applying the above-mentioned sound-based quality testing method, includes a test location point selection unit, a prediction model training unit, a signal loss prediction unit, a test quality assessment unit 1, a test quality assessment unit 2, a test quality assessment unit 3 and a test quality assessment unit 4, with reference to Figure 2, obtain the audio emission position information through the test position point selection unit, preset the circular range test area, select the first original test position point, the second original test position point, the third original test position point and the fourth original test position point from the circular range test area, and obtain the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data according to the statistics of the four position points; obtain the historical audio emission signals, the historical environmental impact feature information and the historical received test signals tested in the historical period through the prediction model training unit, and train to obtain the signal loss prediction model; divide the circular range test area into the first area to be tested, the second area to be tested, the third area to be tested and the fourth area to be tested through the signal loss prediction unit, and detect the first impact feature information, the second impact feature information, the third impact feature information and the fourth impact feature information, and predict the corresponding through the signal loss prediction model test. The first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result; through the test quality evaluation unit one, according to the first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result, the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data are evaluated after data correction to obtain the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result; through the test quality evaluation unit 2, the comparison results between the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are analyzed to be the same quantity, and the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result are compared and judged accordingly, and then the audio qualified test result or the audio unqualified test result is output.

[0111] 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 sound-based quality testing method, characterized in that: The following steps are involved: Acquire audio emission position information, preset a circular range test area, select a first original test position point, a second original test position point, a third original test position point and a fourth original test position point from the circular range test area, and obtain first received audio feature data, second received audio feature data, third received audio feature data and fourth received audio feature data according to statistics of the four position points; Obtain historical audio signals, historical environmental impact feature information, and historical reception test signals tested in historical periods, and train a signal loss prediction model; The circular range test area is divided into a first area to be tested, a second area to be tested, a third area to be tested, and a fourth area to be tested, and first influencing characteristic information, second influencing characteristic information, third influencing characteristic information, and fourth influencing characteristic information are detected and obtained, and a corresponding first signal loss prediction result, a second signal loss prediction result, a third signal loss prediction result, and a fourth signal loss prediction result are predicted through a signal loss prediction model test; According to the first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result, performing data-corrected audio quality evaluation on the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data to obtain a first evaluation result, a second evaluation result, a third evaluation result and a fourth evaluation result; The comparison results between the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are analyzed to be the same quantity, and the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result are compared and judged accordingly, and then the audio qualified test result or the audio unqualified test result is output.

2. The sound-based quality testing method according to claim 1, characterized in that: The steps of analyzing the same quantity of comparison results among the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information, and outputting an audio qualified test result or an audio unqualified test result after performing corresponding comparison and judgment on the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result are specifically as follows: When the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are the same, it is determined whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, and an audio qualified test result or an audio unqualified test result is output; When two of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are the same, and the other two influencing feature information are the same, select the first category evaluation result one, the first category evaluation result two, the second category evaluation result one and the second category evaluation result two from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, judge whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same, and output an audio qualified test result or an audio unqualified test result; When three of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are the same, selecting a third category evaluation result one, a third category evaluation result two and a third category evaluation result three from the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result, judging whether the third category evaluation result one, the third category evaluation result two and the third category evaluation result three are all the same, and outputting an audio qualified test result or an audio unqualified test result; When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are all different, the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result are statistically obtained, and it is determined whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result, and the audio qualified test result or the audio unqualified test result is output.

3. The sound-based quality testing method according to claim 1, characterized in that: The steps of obtaining audio emission position information, presetting a circular range test area, selecting a first original test position point, a second original test position point, a third original test position point, and a fourth original test position point from the circular range test area, and obtaining first received audio feature data, second received audio feature data, third received audio feature data, and fourth received audio feature data according to statistics of the four position points are specifically as follows: Acquire audio emission position information, and preset a circular range test area according to the audio emission position information, wherein the circular range test area is centered on the audio emission position information; Select four audio test receiving position points from the edge of the circular range test area, and output a first original test position point, a second original test position point, a third original test position point, and a fourth original test position point, wherein the first original test position point, the second original test position point, the third original test position point, and the fourth original test position point are all in an equidistant relationship; According to the distance between the audio emission position information and the first original test position point, audio emission parameter data is preset, and based on the audio emission parameter data, the first received audio feature data, the second received audio feature data, the third received audio feature data and the fourth received audio feature data received by the first original test position point, the second original test position point, the third original test position point and the fourth original test position point are respectively counted.

4. The sound-based quality testing method according to claim 2, characterized in that: The steps of obtaining historical audio signals, historical environmental impact feature information, and historical receiving test signals tested in historical periods and training a signal loss prediction model are as follows: Obtain historical audio emission signals, historical environmental impact feature information, and historical reception test signals tested during historical periods; Extracting a historical valid reception test signal for a valid test from the historical reception test signal; According to the historical valid reception test signal, respectively extract the valid record information from the historical audio emission signal and the historical environmental impact characteristic information, and output the historical valid audio emission signal and the historical valid environmental impact characteristic information; Counting the difference between the historical valid audio transmission signal and the historical valid receiving test signal to obtain the historical preprocessing effective difference; A signal loss prediction model is trained based on the historical preprocessing effective difference and the historical effective environmental impact feature information.

5. The sound-based quality testing method according to claim 4, characterized in that: The circular range test area is divided into a first test area, a second test area, a third test area, and a fourth test area, and first influencing characteristic information, second influencing characteristic information, third influencing characteristic information, and fourth influencing characteristic information are detected and obtained, and the steps of predicting the corresponding first signal loss prediction result, second signal loss prediction result, third signal loss prediction result, and fourth signal loss prediction result through a signal loss prediction model test are specifically as follows: According to the first original test position point, the second original test position point, the third original test position point and the fourth original test position point, the circular range test area is divided into four equal spatial areas, and the first area to be tested, the second area to be tested, the third area to be tested and the fourth area to be tested are output, wherein the first area to be tested includes the first original test position point, the second area to be tested includes the second original test position point, the third area to be tested includes the third original test position point, and the fourth area to be tested includes the fourth original test position point; Detecting environmental impact characteristic information in a first test area to obtain first impact characteristic information, and inputting the first impact characteristic information and audio emission parameter data into a signal loss prediction model for testing to obtain a first signal loss prediction result; Detecting the environmental impact characteristic information in the second test area to obtain second impact characteristic information, and inputting the second impact characteristic information and the audio emission parameter data into the signal loss prediction model for testing to obtain a second signal loss prediction result; Detecting the environmental impact characteristic information in the third test area to obtain third impact characteristic information, and inputting the third impact characteristic information and the audio emission parameter data into the signal loss prediction model for testing to obtain a third signal loss prediction result; The environmental impact characteristic information in the fourth test area is detected to obtain fourth impact characteristic information, and the fourth impact characteristic information and audio emission parameter data are input into the signal loss prediction model for testing to obtain a fourth signal loss prediction result.

6. The sound-based quality testing method according to claim 5, characterized in that: The step of determining whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result, and the fourth evaluation result when the first influencing feature information, the second influencing feature information, the third influencing feature information, and the fourth evaluation result, and outputting an audio qualified test result or an audio unqualified test result is specifically as follows: According to the first signal loss prediction result, the first received audio feature data is corrected to obtain a first received audio correction result; According to the second signal loss prediction result, the second received audio feature data is corrected to obtain a second received audio correction result; According to the third signal loss prediction result, the third received audio feature data is corrected to obtain a third received audio correction result; According to the fourth signal loss prediction result, the fourth received audio feature data is corrected to obtain a fourth received audio correction result; Performing audio quality evaluation on the first received audio correction result, the second received audio correction result, the third received audio correction result, and the fourth received audio correction result, respectively, to obtain a first evaluation result, a second evaluation result, a third evaluation result, and a fourth evaluation result, where the evaluation results include qualified or unqualified; When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are the same, determine whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result. If so, the final audio test result is determined to be qualified and the audio qualified test result is output; if not, the final audio test result is determined to be unqualified and the audio unqualified test result is output. The same result refers to qualified.

7. The sound-based quality testing method according to claim 6, characterized in that: When two of the first influencing feature information, the second influencing feature information, the third influencing feature information, and the fourth influencing feature information are the same, and the other two influencing feature information are the same, the steps of selecting the first category evaluation result one, the first category evaluation result two, the second category evaluation result one, and the second category evaluation result two from the first evaluation result, the second evaluation result, the third evaluation result, and the fourth evaluation result, judging whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same, and outputting the audio qualified test result or the audio unqualified test result are specifically as follows: When two of the first impact characteristic information, the second impact characteristic information, the third impact characteristic information, and the fourth impact characteristic information are the same, the first type of impact characteristic information 1 and the first type of impact characteristic information 2 are output, and when the other two types of impact characteristic information are the same, the second type of impact characteristic information 1 and the second type of impact characteristic information 2 are output; According to the first type of impact feature information 1 and the first type of impact feature information 2, select the first type of assessment result 1 and the first type of assessment result 2 from the first assessment result, the second assessment result, the third assessment result and the fourth assessment result; According to the second-category impact feature information one and the second-category impact feature information two, correspondingly selecting the second-category assessment result one and the second-category assessment result two from the first assessment result, the second assessment result, the third assessment result, and the fourth assessment result; Determine whether the first category evaluation result one and the first category evaluation result two are the same, and whether the second category evaluation result one and the second category evaluation result two are the same. If the judgment results are both yes, then the final audio test result is determined to be qualified, and the audio qualified test result is output. If one of the judgment results is no, then the final audio test result is determined to be unqualified, and the audio unqualified test result is output.

8. The sound-based quality testing method according to claim 6, characterized in that: When three of the first influencing feature information, the second influencing feature information, the third influencing feature information, and the fourth influencing feature information are the same, selecting a third category evaluation result one, a third category evaluation result two, and a third category evaluation result three from the first evaluation result, the second evaluation result, the third evaluation result, and the fourth evaluation result, determining whether the third category evaluation result one, the third category evaluation result two, and the third category evaluation result three are all the same, and outputting an audio qualified test result or an audio unqualified test result, specifically comprises: When three of the first impact characteristic information, the second impact characteristic information, the third impact characteristic information, and the fourth impact characteristic information are the same, the impact characteristic information of the three same impact characteristic information is screened out to obtain third-category impact characteristic information one, third-category impact characteristic information two, and third-category impact characteristic information three; According to the third-category impact characteristic information one, the third-category impact characteristic information two, and the third-category impact characteristic information three, correspondingly selecting the third-category assessment result one, the third-category assessment result two, and the third-category assessment result three from the first assessment result, the second assessment result, the third assessment result, and the fourth assessment result; Determine whether the third category evaluation result one, the third category evaluation result two and the third category evaluation result three are all the same. If so, determine that the final audio test result is qualified and output the audio qualified test result. If not, determine that the final audio test result is unqualified and output the audio unqualified test result.

9. The sound-based quality testing method according to claim 6, characterized in that: When the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information are all different, the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result are obtained by statistics, and it is determined whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result, and the step of outputting an audio qualified test result or an audio unqualified test result is specifically as follows: When the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information are all different, it is determined whether there are at least three identical results among the first evaluation result, the second evaluation result, the third evaluation result and the fourth evaluation result. If so, the final audio test result is determined to be qualified, and the audio qualified test result is output; if not, and the judgment result is that there are two identical results, a new test is performed and secondary test information is output; According to the secondary test information, a first new evaluation result, a second new evaluation result, a third new evaluation result and a fourth new evaluation result are obtained by statistics; Determine whether there are at least three identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result. If so, determine that the final audio test result is qualified and output the qualified audio test result. If not, determine whether there are still two identical results among the first new evaluation result, the second new evaluation result, the third new evaluation result and the fourth new evaluation result. If so, determine that the final audio test result is unqualified and output the unqualified audio test result.

10. A sound-based quality testing system, characterized in that: The system is used to implement the sound-based quality testing method according to any one of claims 1 to 9, comprising: A test position point selection unit is used to obtain audio emission position information, preset a circular range test area, select a first original test position point, a second original test position point, a third original test position point and a fourth original test position point from the circular range test area, and obtain first received audio feature data, second received audio feature data, third received audio feature data and fourth received audio feature data according to statistics of the four position points; A prediction model training unit, used to obtain historical audio transmission signals, historical environmental impact feature information and historical reception test signals tested in historical periods, and train a signal loss prediction model; A signal loss prediction unit is used to divide the circular range test area into a first test area, a second test area, a third test area and a fourth test area, and detect the first influencing characteristic information, the second influencing characteristic information, the third influencing characteristic information and the fourth influencing characteristic information, and predict the corresponding first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result and the fourth signal loss prediction result through a signal loss prediction model test; A test quality evaluation unit 1 is used to perform a data-corrected audio quality evaluation on the first received audio feature data, the second received audio feature data, the third received audio feature data, and the fourth received audio feature data according to the first signal loss prediction result, the second signal loss prediction result, the third signal loss prediction result, and the fourth signal loss prediction result to obtain a first evaluation result, a second evaluation result, a third evaluation result, and a fourth evaluation result; Test quality assessment unit two is used to analyze the comparison results of the first influencing feature information, the second influencing feature information, the third influencing feature information and the fourth influencing feature information to find that the results are the same, and to output an audio qualified test result or an audio unqualified test result after making corresponding comparisons and judgments on the first assessment result, the second assessment result, the third assessment result and the fourth assessment result.

Images