Sound quality detection method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of computer processing, and discloses a sound quality detection method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining to-be-detected audio data; feature detection is carried out on the to-be-detected audio data to obtain audio features, the audio features comprise cut-off frequency and / or sonic boom features, the cut-off frequency is obtained based on frequency domain data corresponding to the to-be-detected audio data, and the sonic boom features are obtained based on time domain data corresponding to the to-be-detected audio data; and determining a tone quality detection result of the to-be-detected audio data based on a size relationship between the audio features and corresponding preset thresholds.

Description

Technical Field

[0001] The present disclosure relates to the field of computer processing, and in particular to a sound quality detection method, device, computer equipment and storage medium. Background Art

[0002] In the field of audio processing, it is necessary to test the sound quality so that the audio processing method can be adjusted in time. Among them, sound quality detection is generally performed by training a network model and using the network model to perform sound quality detection. However, the processing method of the network model will cause a large processing delay and is difficult to apply to scenarios that require rapid analysis of sound quality effects. Summary of the invention

[0003] In view of this, the present disclosure provides a sound quality detection method, apparatus, computer equipment and storage medium to solve the problem of sound quality detection.

[0004] In a first aspect, the present disclosure provides a sound quality detection method, the method comprising:

[0005] Get the audio data to be detected;

[0006] Performing feature detection on the audio data to be detected to obtain audio features, the audio features including a cutoff frequency and / or a pop feature, the cutoff frequency being obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature being obtained based on time domain data corresponding to the audio data to be detected;

[0007] Based on the magnitude relationship between the audio feature and the corresponding preset threshold, a sound quality detection result of the audio data to be detected is determined.

[0008] In a second aspect, the present disclosure provides a sound quality detection device, the device comprising:

[0009] An audio data acquisition module, used to acquire audio data to be detected;

[0010] a feature detection module, configured to perform feature detection on the audio data to be detected to obtain audio features, wherein the audio features include a cutoff frequency and / or a pop feature, wherein the cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected;

[0011] The sound quality detection module is used to determine the sound quality detection result of the audio data to be detected based on the size relationship between the audio feature and the corresponding preset threshold.

[0012] In a third aspect, the present disclosure provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the sound quality detection method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0013] In a fourth aspect, the present disclosure provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the sound quality detection method of the first aspect or any corresponding embodiment thereof.

[0014] The sound quality detection method provided by the embodiment of the present disclosure obtains audio features by performing feature detection on the audio data to be detected, and then determines the sound quality detection result based on the size relationship between the audio features and the corresponding preset thresholds, wherein the audio features include a cutoff frequency obtained based on frequency domain data and / or a pop feature obtained based on time domain data. Since the audio features are obtained by digital signal processing rather than by a trained model, the efficiency of sound quality detection of the audio data to be detected can be improved while ensuring the sound quality detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the specific embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 is a flow chart of a sound quality detection method according to an embodiment of the present disclosure;

[0017] Figure 2 is a flow chart of another sound quality detection method according to an embodiment of the present disclosure;

[0018] Figure 3 is a flow chart of another sound quality detection method according to an embodiment of the present disclosure;

[0019] Figure 4 is a flow chart of yet another sound quality detection method according to an embodiment of the present disclosure;

[0020] Figure 5 is a flow chart of another sound quality detection method according to an embodiment of the present disclosure;

[0021] Figure 6 is a structural block diagram of a sound quality detection device according to an embodiment of the present disclosure;

[0022] Figure 7 It is a schematic diagram of the hardware structure of the computer device of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0024] In the related art, when the sound quality of audio data is detected, the overall sound quality effect of the audio data is generally detected. If further analysis is needed to find out the cause of the sound quality problem in the audio data, a trained network model needs to be used to analyze it and obtain the analysis result. Since model training and model operation both take a lot of time, the efficiency of sound quality detection is low when the analysis is performed using a trained network model.

[0025] Based on this, the sound quality detection method provided in the embodiment of the present disclosure analyzes the audio data to be detected by means of digital signal processing to improve the detection efficiency. And in the detection process, the sound quality detection result is determined based on the audio features. The audio features include cutoff frequency and / or popping sound. Through the detection of these features, the audio detection method can be applied to detection scenarios with special scene requirements, thereby improving the applicable scenarios of the sound quality detection method.

[0026] According to an embodiment of the present disclosure, an embodiment of a sound quality detection method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] In this embodiment, a sound quality detection method is provided, which can be used for the above-mentioned computer equipment, such as a mobile terminal, a computer, etc. Figure 1 is a flow chart of a sound quality detection method according to an embodiment of the present disclosure, such as Figure 1 As shown, the process includes the following steps:

[0028] Step S101, obtaining audio data to be detected.

[0029] The audio data to be detected can be real-time audio data or offline audio data. For example, in a live broadcast scenario, sound quality detection is performed on the audio data in the real-time media data to determine the sound quality effect in the current live broadcast scenario. Alternatively, in a scenario where an audio processing strategy is selected, different audio processing strategies are used to process the same audio data separately to obtain audio data to be detected corresponding to the audio processing strategy, and then sound quality detection is performed on each audio data to be detected to determine the sound quality detection result corresponding to the audio processing strategy, thereby using the sound quality detection result to select a better audio processing strategy from multiple audio processing strategies for subsequent audio processing.

[0030] Of course, the method for obtaining the audio to be detected is not limited to the above method, and other methods can also be used to obtain it, which is not limited here.

[0031] Step S102: perform feature detection on the audio data to be detected to obtain audio features.

[0032] The audio features include a cutoff frequency and / or a pop feature, the cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected.

[0033] Feature detection is obtained by performing digital signal processing on the audio data to be detected, wherein the audio features include cutoff frequency and / or popping sound features. If the audio data exceeding a certain frequency is truncated, the frequency is called the cutoff frequency; if the peak value of the audio exceeds a peak value, the played audio will have a popping sound similar to a sizzling sound, resulting in poor playback effect.

[0034] Through the above description of the cutoff frequency, the cutoff frequency is obtained based on the frequency domain data corresponding to the audio data to be detected. Since the cutoff frequency is used to detect whether the audio data exceeding a certain frequency is truncated, when analyzing the frequency domain data, it can be determined by detecting the waveform of the frequency domain data. For the description of the popping sound, it can be determined by detecting the peak point in the time domain data.

[0035] In some optional implementations, audio features may also include loudness, audio energy, etc., and are not limited to the cutoff frequency and / or popping features mentioned above. Which specific audio features need to be included depends on the requirements of the specific sound quality detection scenario, and no limitation is made here.

[0036] Multiple audio feature calculation modules are integrated in the audio detection method, and each audio feature calculation module can be turned on and off as needed. For example, five audio feature calculation modules, namely loudness, audio energy, cutoff frequency, popping sound and phase, are integrated. If the audio features required in the current audio scene are loudness, cutoff frequency and phase, then the functions of the audio feature calculation modules corresponding to the loudness, cutoff frequency and phase are turned on, and the functions of the audio feature calculation modules corresponding to the remaining audio features are turned off. The turning on and off of the functions of each audio feature calculation module can be set according to actual needs, and no limitation is made here.

[0037] Step S103: determining a sound quality detection result of the audio data to be detected based on a magnitude relationship between the audio feature and a corresponding preset threshold.

[0038] The audio feature corresponds to the preset threshold value one by one, and the corresponding relationship can be maintained by using a threshold relationship table. After the audio feature is detected in the above step S102, the preset threshold value corresponding to the audio feature is determined by querying the threshold relationship table, and then the size relationship is compared to determine the sound quality detection result of the audio data to be detected.

[0039] The preset threshold corresponding to each audio feature can be set based on experience or obtained by analyzing a large amount of audio data and its sound quality effects. The specific value is not limited to it. The specific value of the preset threshold can also be modified according to actual needs.

[0040] The sound quality detection method provided in this embodiment obtains audio features by performing feature detection on the audio data to be detected, and then determines the sound quality detection result based on the size relationship between the audio features and the corresponding preset thresholds, wherein the audio features include a cutoff frequency obtained based on frequency domain data and / or a pop feature obtained based on time domain data. Since the audio features are obtained by digital signal processing rather than by a trained model, the efficiency of sound quality detection of the audio data to be detected can be improved while ensuring the sound quality detection effect.

[0041] In this embodiment, a sound quality detection method is provided, which can be used for the above-mentioned computer equipment, such as a mobile terminal, a computer, etc. Figure 2 is a flow chart of a sound quality detection method according to an embodiment of the present disclosure, such as Figure 2 As shown, the process includes the following steps:

[0042] Step S201, obtain the audio data to be detected. Figure 1 The detailed description of step S101 in the illustrated embodiment is not repeated here.

[0043] Step S202: perform feature detection on the audio data to be detected to obtain audio features.

[0044] The audio features include a cutoff frequency and / or a pop feature, the cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected.

[0045] If the audio feature includes a cutoff frequency, based on this, the above step S202 includes:

[0046] Step S2021, performing frequency domain conversion on the audio data to be detected to obtain frequency domain data corresponding to the audio data to be detected.

[0047] The method used for frequency domain conversion includes but is not limited to Fourier transform or fast Fourier transform, etc. The audio data to be detected is converted into the frequency domain through the corresponding transformation method to obtain corresponding frequency domain data.

[0048] Step S2022, query the frequency domain data to determine the frequency domain signal segments whose energy drops below the target threshold value among a preset number of sampling points and the indexes corresponding to the frequency domain signal segments.

[0049] The frequency domain data includes the energy corresponding to each frequency. When sampling the frequency domain data, the frequency and energy of each sampling point are recorded, and each sampling point is represented by a unique index. After obtaining the frequency domain data, a query can be performed every preset number of sampling points to compare the energy difference with the target threshold. For example, the preset number is 5 to 20, and the target threshold is 4dB to 8dB. If the preset number is 10, every 10 sampling points, query whether there is a frequency domain signal segment whose energy drops below the target threshold in these 10 sampling points. If so, record the index corresponding to the frequency domain signal segment. Specifically, among the 10 sampling points, compare the energy between adjacent sampling points in turn. If the energy between adjacent sampling points decreases in turn, and the reduction reaches the target threshold, the signal segment formed by these sampling points is called a frequency domain signal segment, and the index of the frequency domain signal segment is the index of the two endpoints of the frequency domain signal segment.

[0050] In some optional implementations, the frequency domain signal segment may also be referred to as a slope according to its waveform, and the two ends of the frequency domain signal segment are the slope vertex and the slope bottom. Accordingly, the index of the frequency domain signal segment includes the index of the vertex and the index of the slope bottom.

[0051] Step S2023: Screen the frequency domain signal segments based on the energy of the frequency domain signal segments to determine candidate frequency domain signal segments.

[0052] Since the index of each frequency domain signal segment is obtained in the above step S2022, as described above, the corresponding frequency and energy can be determined by using the index, so the energy of the frequency domain signal segment can be obtained in this step. The frequency domain signal segment is screened based on the energy of the frequency domain signal segment, and the screening method includes but is not limited to screening out the frequency domain signal segment whose slope is not steep enough or the frequency domain signal segment whose energy at the bottom of the slope is too large, etc. The specific screening method is set according to actual needs and is not limited to it here.

[0053] In some optional implementations, the frequency domain signal segment may also be referred to as a slope. In this implementation, it is necessary to filter out the frequency domain signal segment whose slope is not steep enough. Specifically, step S2023 includes:

[0054] Step a1, smoothing the frequency domain signal segment to obtain a processed frequency domain signal segment and a first endpoint index of the processed frequency domain signal segment, wherein the first endpoint index is greater than the second endpoint index of the processed frequency domain signal segment.

[0055] Step a2: obtaining a preset adjustment index.

[0056] Step a3: Calculate the sum of the first endpoint index and the preset adjustment index to obtain the adjustment index.

[0057] Step a4: determining the first energy and the second energy corresponding to the first endpoint index and the adjustment index respectively based on the processed frequency domain signal segment.

[0058] Step a5: If the difference between the first energy and the second energy is less than the first threshold, filter the corresponding frequency domain signal segment to determine the candidate frequency domain signal segment.

[0059] The smoothing processing method includes using a low-pass filter, etc., by setting the filtering parameters of the low-pass filter, filtering the frequency domain signal, and obtaining the frequency domain signal segment after smoothing. Since the index of the frequency domain signal segment is recorded in the above steps, the index of the processed frequency domain signal segment can be obtained accordingly. As mentioned above, the index corresponds to the frequency and energy, so the corresponding first endpoint frequency can be obtained using the index of the processed frequency domain signal segment.

[0060] The frequency domain signal segment can also be called a slope, the first endpoint index is the slope bottom index of the frequency domain signal segment, and the second endpoint index is the slope top index of the frequency domain signal segment. The preset adjustment index is set according to actual needs, for example, it is set to 6. Of course, the specific value of the preset adjustment index is set according to actual needs, and there is no limitation on it here.

[0061] If the frequency domain signal segment is represented by S, the processed frequency domain signal segment obtained after smoothing is represented by S0, and the preset adjustment index is 6, then the first energy corresponding to the first endpoint index is represented by S0[bottom], the second energy is represented by S0[bottom+6], and the first threshold is represented by T1. If the frequency domain signal segment satisfies the following formula, it needs to be filtered out: S0[bottom]-S0[bottom+6] <T1。

[0062] The value range of the first threshold may be 3dB to 5dB, and the specific value is set according to actual needs.

[0063] The frequency domain signal segments satisfying the above formula are filtered out to determine the candidate frequency domain signal segments. It should be noted that the above is only one way to filter the frequency domain signal segments, and in actual application scenarios, other filtering methods may also be included, which are specifically set according to actual needs.

[0064] By setting the preset adjustment index, the first endpoint index and the adjustment index are obtained, and the frequency domain signal segments are screened using the first energy and the second energy corresponding to the first endpoint index and the adjustment index respectively, so as to screen out the frequency domain signal segments that are not steep enough.

[0065] In some optional implementations, it is necessary to filter out frequency domain signal segments with too much energy at the bottom of the slope. Specifically, step S2023 includes:

[0066] Step b1, obtaining the total energy corresponding to the frequency domain data.

[0067] Step b2: for each frequency domain signal segment, obtain the total energy of the first endpoint in the frequency domain signal segment and the second threshold corresponding to the frequency domain signal segment.

[0068] Step b3, calculating the ratio of the total energy of the first endpoint to the total energy corresponding to the frequency domain data.

[0069] Step b4: if the ratio is greater than the second threshold, filter the corresponding frequency domain signal segment to determine the candidate frequency domain signal segment.

[0070] Since the frequency domain data represents the corresponding relationship between frequency and energy, the total energy corresponding to the frequency domain data is obtained by calculating the sum of the energies corresponding to all sampling points. As mentioned above, the frequency domain signal segment can also be called a slope, the first endpoint is the bottom point of the slope, and the second endpoint is the top point of the slope. The total energy of the first endpoint is the sum of the energies of all sampling points starting from the bottom point of the slope in each frequency domain signal segment. For example, the total energy corresponding to the audio data is expressed as sum(S), and the total energy of the first endpoint bottom of the i-th frequency domain signal segment is expressed as sum(S[bottom i :]). The second threshold corresponding to the i-th frequency domain signal segment is expressed as T 2i, if the ratio of the total energy of the first endpoint to the total energy corresponding to the frequency domain data is greater than the second threshold, the i-th frequency domain signal segment needs to be filtered out. That is, if the frequency domain signal segment satisfies the following formula, it needs to be filtered out: sum(S[bottom i :]): sum(S)>T 2i .

[0071] Among them, the second threshold T 2i It corresponds to the frequency domain signal segment one by one, and different frequency domain signal segments correspond to different second thresholds. The second threshold is obtained based on the frequency range of the frequency domain signal segment. For example, a relationship table of second thresholds corresponding to different frequency ranges is maintained. By determining the frequency range of the frequency domain signal segment, the relationship table is queried to determine the second threshold corresponding to the frequency domain signal segment. For example, the frequency is negatively correlated with the second threshold, that is, the larger the frequency obtained based on the frequency range of the frequency domain signal segment, the smaller the second threshold; the smaller the frequency obtained based on the frequency range of the frequency domain signal segment, the larger the second threshold.

[0072] The frequency domain signal segments are screened by the ratio of the total energy of the first endpoint to the total energy corresponding to the frequency domain data, so as to screen out the frequency domain signal segments with too large energy of the first endpoint.

[0073] In some optional implementations, the step of obtaining the second threshold corresponding to the frequency domain signal segment in step b2 includes:

[0074] Step b21, determining the frequency range of the frequency domain signal segment using the index corresponding to the frequency domain signal segment.

[0075] Step b22: determine the corresponding second threshold based on the frequency range.

[0076] Each frequency domain signal segment has a corresponding index for the top and bottom of the slope, and the frequency range of the frequency domain signal segment can be determined by the index. If the top of a frequency domain signal segment is top = 101 and the bottom of the slope is bottom = 110, then the frequency corresponding to the top of the slope is F_top = 4300Hz and the frequency corresponding to the bottom of the slope is F_bottom = 4600Hz. The corresponding relationship between the frequency range and the second threshold value includes:

[0077] If the frequency range is [0, 5000 Hz], the second threshold is 0.64;

[0078] If the frequency range is [8000Hz, 10000Hz], the second threshold is 0.6;

[0079] If the frequency range is [16000 Hz, 20000 Hz], the second threshold is 0.05.

[0080] The frequency range [4300Hz, 4600Hz] is used to query the above correspondence and determine that the second threshold corresponding to the frequency domain signal segment is 0.64. When querying the correspondence, the frequency corresponding to the top of the slope can be used, the frequency corresponding to the bottom of the slope can be used, or the average of the frequencies of the top of the slope and the bottom of the slope can be used, etc. Of course, the specific frequency used in the query is processed according to actual needs and is not limited here.

[0081] The frequency range of the frequency domain signal segment is determined by the index corresponding to the frequency domain signal segment, thereby determining the corresponding second threshold, that is, the second threshold corresponds to each frequency domain signal segment, which can ensure the reliability of the second threshold and thus ensure the accuracy of screening.

[0082] In some optional implementations, it is necessary to filter out frequency domain signal segments that include frequency domain signal sub-segments and have a large number of frequency domain signal sub-segments in the frequency domain signal segment. Specifically, step S2023 includes:

[0083] Step c1: if the energy difference between the current frequency domain signal segment and the next frequency domain signal segment exceeds a third threshold, count the number of frequency domain signal sub-segments of the current frequency domain signal segment, and the third threshold is less than the target threshold.

[0084] Step c2: if the number of the frequency domain signal sub-segments is greater than a fourth threshold, filter the current frequency domain signal segment to determine an alternative frequency domain signal segment.

[0085] For each frequency domain signal segment, compare it with the energy of the next adjacent frequency domain signal segment. For example, compare the energy of the slope bottom point of the current frequency domain signal segment with the slope top point energy of the next frequency domain signal segment. If the energy difference exceeds the third threshold, it is necessary to further count the number of frequency domain signal sub-segments in the current frequency domain signal segment. For example, query the energy of the sampling points in the current frequency domain signal segment. If there is a point where the energy change trend changes, the frequency domain signal segment is divided into frequency domain signal sub-segments at this point.

[0086] Among them, the third threshold is less than the first threshold. If the target threshold is expressed as n0, the third threshold can be n0-1, or n0-2, etc. If the number of frequency domain signal sub-segments is greater than the fourth threshold, the current frequency domain signal segment needs to be filtered out to determine the alternative frequency domain signal segment. The value range of the fourth threshold can be 1 to 3, or a number within another value range can be used. No limitation is made to its specific value here.

[0087] There may be multiple smaller frequency domain signal sub-segments in the frequency domain signal segment. The frequency domain signal segment with smaller frequency domain signal sub-segments is screened out by energy difference and the number of frequency domain signal sub-segments, thereby further ensuring that a lower cutoff frequency can be detected.

[0088] Step S2024: determine the cutoff frequency based on the index corresponding to the candidate frequency domain signal segment.

[0089] After determining the candidate frequency domain signal segments, the candidate frequency domain signal segments may be screened again to obtain the target frequency domain signal segments, and then the cutoff frequency may be determined using the index corresponding to the target frequency domain signal.

[0090] In some optional implementations, step S2024 includes:

[0091] Step d1, based on the indexes corresponding to the candidate frequency domain signal segments, the candidate frequency domain signal segments are screened to obtain the target frequency domain signal segments.

[0092] Step d2, determining the cutoff frequency based on the index corresponding to the target frequency domain signal segment.

[0093] When screening the alternative frequency domain signal segments, the alternative frequency domain signal segment corresponding to the largest index can be selected as the target frequency domain signal segment, or the mean of the slope vertex index and the slope bottom point index of each alternative frequency domain signal segment can be calculated, and then the alternative frequency domain signal segment corresponding to the largest mean value can be used as the target frequency domain signal segment.

[0094] After obtaining the target frequency domain signal segment, the index may be converted into a cutoff frequency, for example, by combining parameters such as a sampling rate; or the cutoff frequency may be calculated based on the frequency corresponding to the index, and so on.

[0095] For frequency domain signals, there is a corresponding relationship between their index and frequency. Using the index to screen the candidate frequency domain signal segments can reduce the amount of data processing. On this basis, the cutoff frequency is obtained based on the index, which improves the reliability of the obtained cutoff frequency.

[0096] In some optional embodiments, step d2 includes:

[0097] Step d21, obtaining the sampling rate corresponding to the frequency domain data and the parameters of the frequency domain conversion.

[0098] Step d22, determining the cutoff frequency based on the index corresponding to the target frequency domain signal segment, the sampling rate, and the frequency domain conversion parameters.

[0099] The sampling rate corresponding to the frequency domain data is obtained by recording the sampling parameters before sampling. The parameters of the frequency domain conversion are related to the frequency domain conversion method. If the frequency domain conversion adopts Fourier transform, the parameters of the frequency domain conversion include the length of Fourier transform. Based on this, the calculation methods of the cutoff frequency include:

[0100]

[0101] Among them, f cutoffis the cutoff frequency; the fft length is the parameter of frequency domain conversion, that is, the length of Fourier transform.

[0102] The index is converted into a cutoff frequency through the sampling rate and the parameters of the frequency domain conversion, thereby improving the reliability of the cutoff frequency.

[0103] Step S203: Determine the sound quality detection result of the audio data to be detected based on the size relationship between the audio feature and the corresponding preset threshold. Figure 1 The detailed description of step S103 of the illustrated embodiment is not repeated here.

[0104] The sound quality detection method provided in this embodiment determines the frequency domain signal segments and their corresponding indexes at the target energy drop threshold in a preset number of sampling points by querying the frequency domain data, and further screens the candidate frequency domain segments on this basis to determine the cutoff frequency. Multiple screenings can be closer to the listening sense, and when audio with a low cutoff frequency and audio with a high cutoff frequency are mixed, a lower cutoff frequency can be detected.

[0105] In this embodiment, a sound quality detection method is provided, which can be used for the above-mentioned computer equipment, such as a mobile terminal, a computer, etc. Figure 3 is a flow chart of a sound quality detection method according to an embodiment of the present disclosure, such as Figure 3 As shown, the process includes the following steps:

[0106] Step S301, obtain the audio data to be detected. Figure 1 The detailed description of step S101 in the illustrated embodiment is not repeated here.

[0107] Step S302: perform feature detection on the audio data to be detected to obtain audio features.

[0108] The audio features include a cutoff frequency and / or a pop feature, the cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected.

[0109] If the audio feature includes a pop feature, based on this, step S302 includes:

[0110] Step S3021, obtaining a preset frame length and a preset coefficient.

[0111] The preset frame length is the length of data processed each time, and the preset coefficient is a coefficient used to adjust the peak value of the time domain signal segment. The specific values ​​of the preset frame length and the preset coefficient are set according to actual needs and are not limited here.

[0112] Step S3022, segmenting the time domain data based on a preset frame length to obtain time domain signal segments and peak values ​​of the time domain signal segments.

[0113] The time domain data is segmented by a preset frame length, that is, data of corresponding length is extracted from the time domain data using the preset frame length each time to obtain each time domain signal segment. The amplitudes of the sampling points in each time domain signal segment are compared to obtain the peak value of the time domain signal segment.

[0114] Step S3023, sampling the time domain signal segment, and comparing the amplitude differences between adjacent sampling points.

[0115] The sampling rate for sampling the time domain signal segment is set according to actual needs. For each time domain signal segment, the amplitude difference between adjacent sampling points is compared and recorded.

[0116] Step S3024: if there is a preset number of continuous sampling points whose amplitude differences are smaller than a preset value, it is determined that there is a pop in the time domain signal segment to obtain the number of sampling points corresponding to the pop.

[0117] The preset value is the product of the peak value of the time domain signal segment and a preset coefficient.

[0118] Each time domain signal segment has a corresponding peak value, and the product of the peak value and the preset coefficient is calculated to obtain a preset value. The preset value corresponds to the time domain signal segment one by one, and different time domain signal segments correspond to different preset values. For each time domain signal segment, the amplitude difference between adjacent sampling points is recorded. If there is a continuous preset number of sampling points whose amplitude difference is less than the preset value, it is determined that there is a pop in the time domain signal segment. If the preset number is step_n, if the amplitude difference of the continuous step_n sampling points is less than the preset value, it is considered a pop.

[0119] For each time domain signal segment, since the above comparison is performed based on the amplitude of the sampling point, accordingly, after determining that there is a pop in the time domain signal segment, the number of sampling points corresponding to the pop can also be obtained. That is, the number of continuous sampling points whose amplitude difference is less than the preset value can be considered as the number of sampling points corresponding to the pop.

[0120] Step S3025, determining the pop feature based on the number of sampling points corresponding to the pop and the number of sampling points corresponding to the peak.

[0121] Among them, the popping characteristics include the popping rate and the peak ratio.

[0122] For each time domain signal segment, the number of sampling points corresponding to the pop and the number of sampling points corresponding to the peak can be obtained. The number of sampling points corresponding to the peak is obtained by counting the sampling points whose amplitude is equal to the peak.

[0123] After obtaining the number of sampling points corresponding to the pop and the number of sampling points corresponding to the peak, the corresponding pop rate and peak ratio can be calculated using the above two numbers.

[0124] In some optional implementations, the above step S3025 includes:

[0125] Step e1, obtaining the total number of sampling points of time domain data.

[0126] Step e2, determining the pop rate by using the ratio of the number of sampling points corresponding to the pop to the total number of sampling points.

[0127] Step e3, using the ratio of the number of sampling points corresponding to the peak value to the total number of sampling points to determine the peak value proportion.

[0128] The total number of sampling points can be obtained by statistics or by calculating the length of the time domain data and the sampling rate. The pop rate is the ratio of the number of sampling points corresponding to the pop to the total number of sampling points, and the peak ratio is the ratio of the number of sampling points corresponding to the peak to the total number of sampling points.

[0129] Due to the setting of the preset coefficient, there may be omissions when using the preset value for pop detection. On this basis, the calculation of the peak ratio is introduced to further improve the accuracy of the pop feature.

[0130] Step S303: Determine the sound quality detection result of the audio data to be detected based on the magnitude relationship between the audio feature and the corresponding preset threshold. Figure 1 The detailed description of step S103 of the illustrated embodiment is not repeated here.

[0131] The sound quality detection method provided in this embodiment determines whether there is a pop in the time domain signal segment through the relationship between the amplitude of the sampling point of the time domain signal segment and the preset value, that is, provides a method for determining the pop, so that it can be applied to the scenario where the sound quality detection requirement is to have pop.

[0132] In some optional implementations, the audio feature also includes a phase feature. Based on this, the step of performing feature detection on the audio data to be detected in the above step to obtain the audio feature includes:

[0133] Step f1, obtaining the energy value corresponding to the audio data to be detected.

[0134] Step f2, obtaining left channel audio data and right channel audio data in the audio data to be detected.

[0135] Step f3: superimpose the left channel audio data and the right channel audio data to obtain a downmix signal and an energy value of the downmix signal.

[0136] Step f4, obtaining a phase feature based on a difference between the energy value of the downmix signal and the energy value corresponding to the audio data to be detected.

[0137] The energy value corresponding to the audio data to be detected may be the energy average of all sampling points in the audio data to be detected, or may be obtained by other methods. For subsequent unified calculation, the energy average of all sampling points is used.

[0138] If there is a phase difference between the left channel data and the right channel data in the audio data to be detected, the energy value will decrease when the two are superimposed. Based on this, a downmix signal is obtained by superimposing the left channel data and the right channel data, and the energy value of the downmix signal is obtained. The difference between the energy value of the downmix signal and the energy value corresponding to the audio data to be detected is calculated to obtain the phase feature.

[0139] When detecting the phase feature, the downmix signal is obtained by superimposing the left and right channel audio data, and the phase feature is determined by the difference between the energy value of the downmix signal and the energy value corresponding to the audio data to be detected, which simplifies the calculation process of the phase feature and improves the efficiency of feature detection.

[0140] In this embodiment, a sound quality detection method is provided, which can be used for the above-mentioned computer equipment, such as a mobile terminal, a computer, etc. Figure 4 is a flow chart of a sound quality detection method according to an embodiment of the present disclosure, such as Figure 4 As shown, the process includes the following steps:

[0141] Step S401, obtain the audio data to be detected. For details, see Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0142] Step S402: perform feature detection on the audio data to be detected to obtain audio features.

[0143] The audio features include a cutoff frequency and / or a pop feature, the cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected.

[0144] For details, please see Figure 2 Step S202 of the embodiment shown or Figure 3 The detailed description of step S203 of the illustrated embodiment is not repeated here.

[0145] Step S403: determining a sound quality detection result of the audio data to be detected based on a magnitude relationship between the audio feature and a corresponding preset threshold.

[0146] Specifically, the above step S403 includes:

[0147] Step S4031, based on the magnitude relationship between the audio feature and the corresponding preset threshold, determine the target audio feature corresponding to the sound quality problem.

[0148] As described above, there is a one-to-one correspondence between the audio feature and the preset threshold, and the correspondence is represented by a threshold relationship table. After the processing of step S402, the audio feature is obtained. The threshold relationship table is queried using the audio feature to obtain the corresponding preset threshold.

[0149] Then, the audio feature is compared with the corresponding preset threshold to determine whether the audio feature will cause sound quality problems. If it will cause sound quality problems, the audio feature is determined as the target audio feature.

[0150] Step S4032: normalize the sound quality features to obtain sound quality score values ​​corresponding to the sound quality features.

[0151] Since different sound quality features are calculated in different ways, the magnitude of the calculated values ​​is different. Therefore, they are normalized to obtain the sound quality score corresponding to each sound quality feature. If the normalized interval is 0 to 1, 1 indicates a serious sound quality problem, and 0 indicates no sound quality problem.

[0152] If the audio features include popping, cutoff frequency, phase characteristics, audio energy and loudness, correspondingly, it represents the popping problem, the problem of too low cutoff frequency, the problem of reverse phase, the problem of too much energy and the problem of too much loudness. Based on this, the normalization processing method corresponding to each audio feature is as follows:

[0153]

[0154]

[0155]

[0156]

[0157]

[0158] Among them, score1 is the score of the popping sound, clipping_rate is the popping sound rate, T 1min and T 1max is the minimum and maximum value corresponding to the sonic boom rate;

[0159] score2 is the score value of the cutoff frequency, cutoff_f is the cutoff frequency, T 2min and T 2max is the minimum and maximum value corresponding to the cutoff frequency;

[0160] score3 is the score value of the phase feature, downmix_rms is the phase feature, T 3min and T 3max is the minimum and maximum value corresponding to the phase feature;

[0161] score4 is the score value of audio energy, rms_db is the audio energy, T 4min and T 4max is the minimum and maximum value corresponding to the audio energy;

[0162] score5 is the score of loudness, loudness is the loudness, T 5min and T 5max The minimum and maximum values ​​corresponding to the loudness.

[0163] Step S4033: Determine the score value of the audio data to be detected based on the size of the sound quality score value.

[0164] The sound quality detection result includes the target audio features and the score value of the audio data to be detected.

[0165] After obtaining each sound quality score value, the average of the sound quality score values ​​can be used as the score value of the audio data to be detected. Since a larger score value indicates a more serious sound quality problem, the maximum value among the sound quality score values ​​can be used as the score value of the audio data to be detected.

[0166] The sound quality detection method provided in this embodiment has different detection methods for different sound quality features, and the magnitudes of the numerical values ​​of the obtained sound quality features are also inconsistent. Therefore, the sound quality features are normalized to obtain the normalized sound quality score values. The unified sound quality score values ​​can be used to determine the target audio features with sound quality problems and the score values ​​of the entire audio data to be detected, thereby realizing sound quality detection of the entire audio data to be detected and locating the sound quality problems.

[0167] In some optional implementations, step S402 includes:

[0168] Step g1: If the sound quality detection is segmented detection, the size of the segmented data is obtained.

[0169] Step g2, segmenting the audio data to be detected based on the size of the segmented data to obtain audio sub-data to be detected.

[0170] Step g3, performing feature detection on the audio sub-data to be detected to obtain audio features.

[0171] Before testing the audio data to be tested, you can set the sound quality test method to be overall test or segmented test. If overall test is performed, the output includes the score value of the entire audio data to be tested and the target audio features; if segmented test is performed, the output includes the sound quality test results of each audio sub-segment to be tested.

[0172] Specifically, if it is segmented detection, it is also necessary to set the size of the segmented data, and use the size of the segmented data to segment the audio data to be detected to obtain the audio sub-data to be detected. Figure 2 or Figure 3 The feature detection is performed in the manner of step S202 or step S302 of the illustrated embodiment to obtain the audio features of each audio sub-data to be detected.

[0173] When the audio data to be detected is long, segmenting it for detection can obtain more accurate sound quality detection results.

[0174] Accordingly, step S403 includes:

[0175] Step h1: for each audio sub-data to be detected, based on the size relationship between the audio feature and the corresponding preset threshold, count the number of audio sub-data to be detected that have sound quality problems.

[0176] Step h2, calculate the ratio of the number of audio sub-data to be detected with sound quality problems to the number of all audio sub-data to be detected, determine the proportion of audio sub-data with sound quality problems, and the sound quality detection result includes the proportion of audio sub-data with sound quality problems.

[0177] For each audio sub-data to be detected, the audio feature is compared with the corresponding preset threshold to determine whether each audio sub-data to be detected has a sound quality problem, so as to count the number of audio sub-data to be detected with sound quality problems. In the above step S402, the audio data to be detected is segmented, and the number of segments, i.e., the number of audio sub-data to be detected, is recorded.

[0178] After the detection determines that there are sound quality problems, a label can be set for the audio sub-data to be detected. After all the audio sub-data to be detected are detected, the number of labels can be counted to determine the number of audio sub-data to be detected with sound quality problems. Then calculate the ratio of this number to the number of all audio sub-data to be detected to obtain the proportion of sound quality problems. The proportion of sound quality problems can be used to more accurately characterize the sound quality detection results.

[0179] In some other optional implementations, step S402 includes:

[0180] Step i1: If the sound quality detection is segmented detection, obtain the detection time period.

[0181] Step i2: extracting audio data from the audio data to be detected based on the detection time period.

[0182] Step i3, performing feature detection on the extracted audio data to obtain audio features.

[0183] When testing the sound quality, you can test a certain time period. Specifically, get the test time period, and use the test time period to extract the audio data of the corresponding time period from the audio data to be tested to obtain the audio data. Then perform feature detection on the extracted audio data to obtain the audio features. For the feature detection method, please refer to the above. Figure 2 The description of step S202 of the illustrated embodiment, or Figure 3 The description of step S302 of the illustrated embodiment will not be repeated here.

[0184] The audio data is extracted from the audio data to be detected during the detection time period and detected, so that the sound quality effect of the current audio can be known in real time during the audio quality detection process.

[0185] In some optional implementations, the audio feature includes audio energy. In the above steps, feature detection is performed on the audio data to be detected to obtain the audio feature, including:

[0186] Step j1, divide the audio data to be detected into frames every n sampling points, and calculate the sum of the energy of each frame sampling point:

[0187] E=S[0] 2 +S[1] 2 +...+S[n] 2

[0188] Step j2, calculate the historical energy average: E_avg = Sum(E) / count(E), the background noise average is calculated in the frame that meets the energy threshold (e.g., the frame with energy greater than -90dbfs), and the sum of the sampling point values ​​sample_sum is calculated

[0189] Step j3, cache the maximum peak value: Peak = max(Peak, abs(s)), and the minimum peak value: mininum = min(Peak, abs(s))

[0190] Step j4, converting dB values ​​and calculating related indicators, as follows:

[0191] total_left=10*lg(Peak)

[0192] Peak=20*lg(Peak)

[0193] minimum=20*lg(minimum)

[0194] left_right_diff=total_left-total_right

[0195] dc_offset = sumple_sum / total sample value

[0196] The audio features include left_right_diff and DC offset dc_offset.

[0197] In some optional implementations, the audio feature includes a frequency spectrum, and the above steps of performing feature detection on the audio data to be detected to obtain the audio feature include:

[0198] Step k1, calculating the frequency domain signal corresponding to the time domain signal, for example, by using fast Fourier transform.

[0199] Step k2, calculate the spectrum centroid and cutoff frequency confidence, etc.

[0200] Audio features include spectral centroid and cutoff frequency confidence, etc. The spectral centroid is calculated using the following formula:

[0201]

[0202] Among them, centriod is the spectral centroid, f k is the frequency corresponding to spectrum k, s k is the spectrum value at k, b1 and b2 are the index values ​​of the spectrum, and the spectrum centroid is calculated in the range [b1, b2].

[0203] In some optional implementations, each audio feature includes multiple indicators for characterizing the sound quality. For example, if the audio feature includes a spectrum, its corresponding indicators include DC offset, peak energy, etc. In order to clearly characterize the audio features, the corresponding indicators and the corresponding thresholds, the following table is used for representation:

[0204]

[0205]

[0206] It should be noted that the above table is only an example and does not limit the protection scope of the present disclosure. It can be set according to actual needs.

[0207] As a specific application example of the embodiment of the present disclosure, Figure 5As shown, a variety of audio feature calculation modules are integrated in the sound quality detection algorithm, including an audio energy calculation module, a spectrum calculation module, a phase calculation module, a pop calculation preparation module and a loudness calculation module. Each feature calculation module can be turned on and off according to actual needs. For example, if only the audio energy calculation module is needed in the current application scenario, the switch of the audio energy calculation module is turned on in the algorithm calculation preparation module, and the switches of the remaining calculation modules are turned off. The algorithm calculation preparation module is also used to set whether the current detection needs to be segmented, and if segmented, the segment size is set, etc. After the output of the above-mentioned corresponding calculation modules, the overall sound quality score of the audio to be detected and the proportion of segments with poor sound quality, etc. are calculated on this basis. Further, the overall sound quality results and target audio features are output.

[0208] As a specific application embodiment of the disclosed embodiment, the method is applied in a live broadcast scenario to detect the sound quality of real-time audio. At the beginning of the live broadcast, segmented detection is turned on, and a sound quality detection is performed every 10 seconds to obtain the score value of each sound quality detection and the target audio features corresponding to poor sound quality. At the same time, the score value and target audio features of each audio detection are displayed so that the user can know the sound quality effect in the current live broadcast scenario.

[0209] In this embodiment, a sound quality detection device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0210] This embodiment provides a sound quality detection device, such as Figure 6 As shown, including:

[0211] The audio data acquisition module 601 is used to acquire the audio data to be detected.

[0212] The feature detection module 602 is used to perform feature detection on the audio data to be detected to obtain audio features, wherein the audio features include a cutoff frequency and / or a pop feature. The cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected.

[0213] The sound quality detection module 603 is used to determine the sound quality detection result of the audio data to be detected based on the size relationship between the audio feature and the corresponding preset threshold.

[0214] In some optional implementations, if the audio feature includes the cutoff frequency, the feature detection module 602 includes:

[0215] The frequency domain conversion unit is used to perform frequency domain conversion on the audio data to be detected to obtain frequency domain data corresponding to the audio data to be detected.

[0216] The query unit is used to query the frequency domain data to determine the frequency domain signal segments whose energy drops below the target threshold value among a preset number of sampling points and the indexes corresponding to the frequency domain signal segments.

[0217] The screening unit is used to screen the frequency domain signal segments based on the energy of the frequency domain signal segments to determine the candidate frequency domain signal segments.

[0218] The cutoff frequency determining unit is used to determine the cutoff frequency based on the index corresponding to the candidate frequency domain signal segment.

[0219] In some optional embodiments, the screening unit comprises:

[0220] The smoothing subunit is used to smooth the frequency domain signal segment to obtain the processed frequency domain signal segment and the first endpoint index of the processed frequency domain signal segment, wherein the first endpoint index is greater than the second endpoint index of the processed frequency domain signal segment.

[0221] The preset adjustment frequency obtaining subunit is used to obtain the preset adjustment index.

[0222] The sum calculation subunit is used to calculate the sum of the first endpoint index and the preset adjustment index to obtain the adjustment index.

[0223] The energy determination subunit is used to determine the first energy and the second energy corresponding to the first endpoint index and the adjustment index respectively based on the processed frequency domain signal segment.

[0224] The first filtering subunit is used to filter the corresponding frequency domain signal segment to determine an alternative frequency domain signal segment if the difference between the first energy and the second energy is less than a first threshold.

[0225] In some optional embodiments, the screening unit comprises:

[0226] The total energy acquisition subunit is used to obtain the total energy corresponding to the frequency domain data.

[0227] The threshold acquisition subunit is used to acquire, for each frequency domain signal segment, the total energy of the first endpoint in the frequency domain signal segment and the second threshold corresponding to the frequency domain signal segment.

[0228] The ratio calculation subunit is used to calculate the ratio of the total energy of the first endpoint to the total energy corresponding to the frequency domain data.

[0229] The second filtering subunit is used to filter the corresponding frequency domain signal segment to determine an alternative frequency domain signal segment if the ratio is greater than a second threshold.

[0230] In some optional implementations, the threshold acquisition subunit includes:

[0231] The frequency range determination subunit is used to determine the frequency range of the frequency domain signal segment using the index corresponding to the frequency domain signal segment.

[0232] The threshold determination subunit is used to determine the corresponding second threshold based on the frequency range.

[0233] In some optional embodiments, the screening unit comprises:

[0234] The statistical subunit is used to count the number of frequency domain signal subsegments of the current frequency domain signal segment if the energy difference between the current frequency domain signal segment and the next frequency domain signal segment exceeds a third threshold, and the third threshold is less than a target threshold.

[0235] The third filtering subunit is configured to filter the current frequency domain signal segment to determine an alternative frequency domain signal segment if the number of the frequency domain signal sub-segments is greater than a fourth threshold.

[0236] In some optional implementations, the cutoff frequency determining unit includes:

[0237] The screening subunit is used to screen the candidate frequency domain signal segments based on the indexes corresponding to the candidate frequency domain signal segments to obtain the target frequency domain signal segments.

[0238] The first cutoff frequency determination subunit is used to determine the cutoff frequency based on the index corresponding to the target frequency domain signal segment.

[0239] In some optional implementations, the first cutoff frequency determination subunit includes:

[0240] The parameter acquisition subunit is used to obtain the sampling rate corresponding to the frequency domain data and the parameters of the frequency domain conversion.

[0241] The first cutoff frequency determination subunit is used to determine the cutoff frequency based on the index corresponding to the target frequency domain signal segment, the sampling rate, and the frequency domain conversion parameters.

[0242] In some optional implementations, if the audio feature includes the pop feature, the feature detection module 602 includes:

[0243] The coefficient acquisition unit is used to acquire a preset frame length and a preset coefficient.

[0244] The segmentation unit is used to segment the time domain data based on a preset frame length to obtain time domain signal segments and peak values ​​of the time domain signal segments.

[0245] The sampling unit is used to sample the time domain signal segment and compare the amplitude difference between adjacent sampling points.

[0246] The popping sound determination unit is used to determine that there is a popping sound in the time domain signal segment if the amplitude difference between a continuous preset number of sampling points is less than a preset value, so as to obtain the number of sampling points corresponding to the popping sound, and the preset value is the product of the peak value of the time domain signal segment and the preset coefficient.

[0247] The pop feature determination unit is used to determine the pop feature based on the number of sampling points corresponding to the pop and the number of sampling points corresponding to the peak, where the pop feature includes the pop rate and the peak ratio.

[0248] In some optional embodiments, the pop feature determination unit is configured to include:

[0249] The total number of sampling points obtaining subunit is used to obtain the total number of sampling points of time domain data.

[0250] The pop rate determination subunit is used to determine the pop rate by using the ratio of the number of sampling points corresponding to the pop to the total number of sampling points.

[0251] The peak value proportion determination subunit is used to determine the peak value proportion by using the ratio of the number of sampling points corresponding to the peak value to the total number of sampling points.

[0252] In some optional implementations, the audio feature further includes a phase feature, and the feature detection module 602 includes:

[0253] The energy value acquisition unit is used to acquire the energy value corresponding to the audio data to be detected.

[0254] The channel data acquisition unit is used to acquire left channel audio data and right channel audio data in the audio data to be detected.

[0255] The superposition unit is used to superimpose the left channel audio data and the right channel audio data to obtain a downmix signal and an energy value of the downmix signal.

[0256] The difference unit is used to obtain the phase feature based on the difference between the energy value of the downmix signal and the energy value corresponding to the audio data to be detected.

[0257] In some optional implementations, the sound quality detection module 603 includes:

[0258] The target audio feature determination unit is used to determine the target audio feature corresponding to the sound quality problem based on the size relationship between the audio feature and the corresponding preset threshold.

[0259] The normalization processing unit is used to perform normalization processing on the sound quality features to obtain the sound quality score values ​​corresponding to the sound quality features.

[0260] The score value determination unit is used to determine the score value of the audio data to be detected based on the size of the sound quality score value. The sound quality detection result includes the target audio feature and the score value of the audio data to be detected.

[0261] In some optional implementations, the feature detection module 602 includes:

[0262] The segmented data size acquisition unit is used to acquire the size of the segmented data if the sound quality detection is segmented detection.

[0263] The segmentation unit is used to segment the audio data to be detected based on the size of the segmented data to obtain the audio sub-data to be detected.

[0264] The first detection unit is used to perform feature detection on the audio sub-data to be detected to obtain audio features.

[0265] In some optional implementations, the sound quality detection module 603 includes:

[0266] The statistical unit is used to count the number of audio sub-data to be detected that have sound quality problems based on the size relationship between the audio feature and the corresponding preset threshold for each audio sub-data to be detected.

[0267] The ratio calculation unit is used to calculate the ratio of the number of audio sub-data to be detected with sound quality problems to the number of all audio sub-data to be detected, and determine the proportion of audio sub-data with sound quality problems. The sound quality detection result includes the proportion of audio sub-data with sound quality problems.

[0268] In some optional implementations, the feature detection module 602 includes:

[0269] The detection time period acquisition unit is used to acquire the detection time period if the sound quality detection is segmented detection.

[0270] The extracting unit is used to extract audio data from the audio data to be detected based on the detection time period.

[0271] The second detection unit is used to perform feature detection on the extracted audio data to obtain audio features.

[0272] The sound quality detection device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0273] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0274] The present disclosure also provides a computer device having the above Figure 6 The sound quality detection device shown.

[0275] See also Figure 7 , Figure 7 is a schematic diagram of a computer device provided by an optional embodiment of the present disclosure, such as Figure 7 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 7 A processor 10 is taken as an example.

[0276] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0277] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0278] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0279] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0280] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 7 The example of connecting through bus is taken in the following.

[0281] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0282] The embodiments of the present disclosure also provide a computer-readable storage medium. The above-mentioned method according to the embodiments of the present disclosure can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium and downloaded through a network, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0283] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0284] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0285] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0286] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0287] Although the embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A sound quality detection method, characterized in that: The method comprises: Get the audio data to be detected; Performing feature detection on the audio data to be detected to obtain audio features, the audio features including a cutoff frequency and / or a pop feature, the cutoff frequency being obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature being obtained based on time domain data corresponding to the audio data to be detected; Based on the magnitude relationship between the audio feature and the corresponding preset threshold, a sound quality detection result of the audio data to be detected is determined.

2. The method according to claim 1, characterized in that If the audio feature includes the cutoff frequency, performing feature detection on the audio data to be detected to obtain the audio feature includes: Performing frequency domain conversion on the audio data to be detected to obtain frequency domain data corresponding to the audio data to be detected; Query the frequency domain data to determine the frequency domain signal segment whose energy drops to the target threshold value among a preset number of sampling points and the index corresponding to the frequency domain signal segment; Based on the energy of the frequency domain signal segments, the frequency domain signal segments are screened to determine candidate frequency domain signal segments; The cutoff frequency is determined based on the index corresponding to the candidate frequency domain signal segment.

3. The method according to claim 2, characterized in that The screening of the frequency domain signal segments based on the energy of the frequency domain signal segments to determine candidate frequency domain signal segments includes: Performing smoothing processing on the frequency domain signal segment to obtain a processed frequency domain signal segment and a first endpoint index of the processed frequency domain signal segment, wherein the first endpoint index is greater than a second endpoint index of the processed frequency domain signal segment; Get preset adjustment index; Calculating the sum of the first endpoint index and the preset adjustment index to obtain an adjustment index; Based on the processed frequency domain signal segment, determine a first energy and a second energy corresponding to the first endpoint index and the adjustment index respectively; If the difference between the first energy and the second energy is less than a first threshold, the corresponding frequency domain signal segment is filtered to determine the candidate frequency domain signal segment.

4. The method according to claim 2, characterized in that: The screening of the frequency domain signal segments based on the energy of each of the frequency domain signal segments to determine candidate frequency domain signal segments includes: Obtaining total energy corresponding to the frequency domain data; For each of the frequency domain signal segments, obtaining a total energy of a first endpoint in the frequency domain signal segment and a second threshold corresponding to the frequency domain signal segment; Calculating a ratio of the total energy of the first endpoint to the total energy corresponding to the frequency domain data; If the ratio is greater than the second threshold, the corresponding frequency domain signal segment is filtered to determine the candidate frequency domain signal segment.

5. The method according to claim 4, characterized in that Acquiring a second threshold corresponding to the frequency domain signal segment includes: Determine the frequency range of the frequency domain signal segment using the index corresponding to the frequency domain signal segment; The corresponding second threshold is determined based on the frequency range.

6. The method according to claim 3 or 4, characterized in that: The screening of the frequency domain signal segments based on the energy of each of the frequency domain signal segments to determine candidate frequency domain signal segments includes: If the energy difference between the current frequency domain signal segment and the next frequency domain signal segment exceeds a third threshold, counting the number of frequency domain signal subsegments of the current frequency domain signal segment, and the third threshold is less than the target threshold; If the number of the frequency domain signal sub-segments is greater than a fourth threshold, the current frequency domain signal segment is filtered to determine the candidate frequency domain signal segment.

7. The method according to claim 2, characterized in that: The determining the cutoff frequency based on the index corresponding to the candidate frequency domain signal segment includes: Based on the index corresponding to the candidate frequency domain signal segment, the candidate frequency domain signal segment is screened to obtain a target frequency domain signal segment; The cutoff frequency is determined based on the index corresponding to the target frequency domain signal segment.

8. The method according to claim 7, characterized in that The determining the cutoff frequency based on the index corresponding to the target frequency domain signal segment includes: Obtaining a sampling rate corresponding to the frequency domain data and a parameter of the frequency domain conversion; The cutoff frequency is determined based on the index corresponding to the target frequency domain signal segment, the sampling rate, and the parameters of the frequency domain conversion.

9. The method according to claim 1, characterized in that: If the audio feature includes the pop feature, the step of performing feature detection on the audio data to be detected to obtain the audio feature includes: Get the preset frame length and preset coefficient; Segmenting the time domain data based on the preset frame length to obtain time domain signal segments and peak values ​​of the time domain signal segments; Sampling the time domain signal segment and comparing amplitude differences between adjacent sampling points; If there is a continuous preset number of sampling points whose amplitude differences are less than a preset value, determine that there is a pop in the time domain signal segment to obtain the number of sampling points corresponding to the pop, and the preset value is the product of the peak value of the time domain signal segment and the preset coefficient; The pop feature is determined based on the number of sampling points corresponding to the pop and the number of sampling points corresponding to the peak, where the pop feature includes a pop rate and a peak ratio.

10. The method according to claim 9, characterized in that The determining the pop feature based on the number of sampling points corresponding to the pop and the number of sampling points corresponding to the peak, includes: Obtaining the total number of sampling points of the time domain data; Determine the pop rate by using the ratio of the number of sampling points corresponding to the pop to the total number of sampling points; The peak ratio is determined by using the ratio of the number of sampling points corresponding to the peak value to the total number of sampling points.

11. The method according to claim 1, characterized in that: The audio feature also includes a phase feature, and the feature detection is performed on the audio data to be detected to obtain the audio feature, including: Obtaining an energy value corresponding to the audio data to be detected; Acquire left channel audio data and right channel audio data from the audio data to be detected; Superimposing the left channel audio data and the right channel audio data to obtain a downmix signal and an energy value of the downmix signal; The phase feature is obtained based on a difference between an energy value of the downmix signal and an energy value corresponding to the audio data to be detected.

12. The method according to claim 1, characterized in that The determining, based on the magnitude relationship between the audio feature and the corresponding preset threshold, the sound quality detection result of the audio data to be detected includes: Based on the magnitude relationship between the audio feature and the corresponding preset threshold, determining the target audio feature corresponding to the sound quality problem; Normalizing the sound quality feature to obtain a sound quality score value corresponding to the sound quality feature; Based on the size of the sound quality score value, the score value of the audio data to be detected is determined, and the sound quality detection result includes the target audio feature and the score value of the audio data to be detected.

13. The method according to claim 1, characterized in that The performing feature detection on the audio data to be detected to obtain audio features includes: If the sound quality detection is segmented detection, obtain the size of the segmented data; Segmenting the audio data to be detected based on the size of the segmented data to obtain audio sub-data to be detected; Perform feature detection on the audio sub-data to be detected to obtain the audio feature.

14. The method according to claim 13, characterized in that The determining, based on the magnitude relationship between the audio feature and the corresponding preset threshold, the sound quality detection result of the audio data to be detected includes: For each of the audio sub-data to be detected, based on the magnitude relationship between the audio feature and the corresponding preset threshold, counting the number of the audio sub-data to be detected that have sound quality problems; The ratio of the number of audio sub-data to be detected with sound quality problems to the number of all audio sub-data to be detected is calculated to determine the proportion of audio sub-data with sound quality problems, and the sound quality detection result includes the proportion of audio sub-data with sound quality problems.

15. The method according to claim 1, characterized in that The performing feature detection on the audio data to be detected to obtain audio features includes: If the sound quality detection is segmented detection, obtain the detection time period; extracting audio data from the audio data to be detected based on the detection time period; Perform feature detection on the extracted audio data to obtain the audio feature.

16. A sound quality detection device, characterized in that: The device comprises: An audio data acquisition module, used to acquire audio data to be detected; a feature detection module, configured to perform feature detection on the audio data to be detected to obtain audio features, wherein the audio features include a cutoff frequency and / or a pop feature, wherein the cutoff frequency is obtained based on frequency domain data corresponding to the audio data to be detected, and the pop feature is obtained based on time domain data corresponding to the audio data to be detected; The sound quality detection module is used to determine the sound quality detection result of the audio data to be detected based on the size relationship between the audio feature and the corresponding preset threshold.

17. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the sound quality detection method according to any one of claims 1 to 15 by executing the computer instructions.

18. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the sound quality detection method according to any one of claims 1 to 15.

Images