Sound leakage detection system and method for wired headphones

Abstract

The present invention proposes a sound leakage detection system and method for wired headphones. The present invention includes: a leakage parameter acquisition module, a leakage parameter processing module, a leakage data storage module and a leakage detection module. The leakage parameter acquisition module detects the first leakage data leaked from the outside into the cavity, and tests to obtain the first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and collects and transmits it to the leakage parameter processing module at the same time; the leakage parameter processing module modally processes the collected first leakage data and the first pressure data to obtain first detection feature data, and all data are transmitted to the leakage data storage module for storage. The leakage detection module also includes a sound leakage detection model based on a machine learning algorithm, and obtains the corresponding first sound leakage level, and generates a sound leakage detection model using the first detection feature data and the first sound leakage level. The model processes the modalized leakage parameter data to obtain the sound leakage level to detect the sound leakage of the earphone, and distinguishes the sound leakage state of the earphone, thereby providing a basis for improving the sound quality of the earphone and improving the user experience.

Description

Technical Field

[0001] The invention belongs to the technical field of audio processing, and in particular relates to a sound leakage detection system and method for wired headphones. Background Art

[0002] Noise-cancelling headphones are headphones that use active or passive noise reduction to reduce noise. Among them, active noise-cancelling headphones have noise-cancelling circuits that counteract external noise. Most of them use a larger head-mounted design, which can use earplugs and headphone shells to block external noise for the first round of sound insulation, and also to provide sufficient space for the installation of active noise-cancelling circuits and power supplies.

[0003] With the development of headphone technology and intelligence, active noise reduction has become the first choice for manufacturers because it can provide a quiet and comfortable listening experience in many scenarios. Active noise reduction headphones need to first detect the current sound leakage status, and then perform noise reduction according to the corresponding calculation strategy. Therefore, accurate detection of the sound leakage status of the headphone is an important prerequisite for noise reduction of the headphone.

[0004] At present, more and more users are accustomed to wearing headphones for learning, talking, leisure and entertainment or work in various occasions. In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems in the above related technologies: At present, as users are more and more accustomed to wearing headphones in their lives, when users are immersed in a work or thing, in public places, users are usually unaware that the sound leaked from the headphones affects other people's behavior.

[0005] In the above situation, some methods for detecting sound leakage of headphones have been disclosed, but the detection of sound leakage is mostly for Bluetooth wireless headphones. In addition, complex computing resources and complex audio processing methods are used to perform multiple processing on audio signals in the detection of sound leakage, resulting in the complexity and high computational complexity of leakage detection.

[0006] Therefore, there is an urgent need for a method that utilizes an improved automatic modalization model to perform differentiated multi-dimensional acquisition processing on the sound leakage level, so as to obtain the sound leakage detection level of wired headphones, thereby providing a basis for subsequently improving the headphone reception sound quality and improving the user experience. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides a sound leakage detection system and method for wired headphones.

[0008] In a first aspect of the present invention, a method for detecting sound leakage of a wired headset is provided, wherein the method comprises:

[0009] S1. Obtaining the first leakage data leaking from the outside into the cavity, and testing and obtaining the first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and obtaining the corresponding first sound leakage level;

[0010] S2. Modally processing the first leakage data and the first pressure data collected to obtain first detection feature data;

[0011] S3. Constructing a sound leakage detection model using the first detection feature data and the first sound leakage level;

[0012] S4. Obtaining the second leakage data leaking from the outside into the cavity, and testing to obtain the second pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and modally processing the second leakage data and the second pressure data to obtain the second detection feature data;

[0013] S5. The sound leakage detection model processes the second detection feature data to obtain a second sound leakage level; and uses the second sound leakage level to detect the sound leakage status of the wired headphones.

[0014] Furthermore, the cavity is the gap between the non-contacting part of the tuning hole and the ear canal and the part from the external auditory canal to the eardrum when the user wears the wired headphone.

[0015] Furthermore, the sound leakage detection model is constructed using a GNN neural network model or an improved machine learning method.

[0016] Furthermore, the GNN neural network model adopts an S-GNN neural network model that adaptively improves the activation function.

[0017] Furthermore, the improved machine learning method is constructed using a support vector machine model based on pressure transformation.

[0018] A wired headset sound leakage detection system is also provided, comprising a leakage data acquisition terminal and a leakage status detection and analysis terminal, characterized in that:

[0019] The leakage data acquisition terminal includes a leakage parameter acquisition module and a leakage parameter processing module, and the leakage status detection and analysis terminal includes a leakage data storage module and a leakage detection module, wherein:

[0020] The leakage parameter acquisition module detects first leakage data leaking from the outside into the cavity, tests and obtains first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, obtains second leakage data leaking from the outside into the cavity, tests and obtains second pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and transmits the data to the leakage parameter processing module at the same time;

[0021] The leakage parameter processing module: performs modal processing on the first leakage data and the first pressure data collected to obtain first detection feature data; and performs modal processing on the second leakage data and the second pressure data collected to obtain second detection feature data;

[0022] The leakage data storage module is used to store the first detection feature data and the first sound leakage level corresponding thereto, and the second detection feature data and the second sound leakage level corresponding thereto;

[0023] The leakage detection module: constructs a sound leakage detection model using the first detection feature data stored in the leakage data storage module and the first sound leakage level corresponding thereto; receives the second detection feature data from the leakage data storage module or the leakage parameter processing module, and generates the second sound leakage level using the sound leakage detection model according to the second detection feature data.

[0024] Furthermore, the cavity is the gap between the non-contacting part of the tuning hole and the ear canal and the part from the external auditory canal to the eardrum when the user wears the wired headphone.

[0025] Furthermore, the sound leakage detection model is constructed using a GNN neural network model or an improved machine learning method.

[0026] Furthermore, the GNN neural network model adopts an S-GNN neural network model that adaptively improves the activation function.

[0027] Furthermore, the improved machine learning method is constructed using a support vector machine model based on pressure transformation.

[0028] The present invention uses an improved machine learning model, only considers audio and external pressure parameters, and realizes modal fusion. It uses the improved machine learning model for wired headphones to perform differentiated multi-dimensional level detection of sound leakage, providing a basis for subsequent improvement of headphone sound quality and user experience. The hollow audio reception used in the present invention can better reflect the degree of sound leakage, and the level classification processing method is more suitable for the scenario of wired headphones pressing on the outer ear, and the sound leakage status can be detected accurately and timely.

[0029] More embodiments and improved effects of the present invention will be further introduced in conjunction with the drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of a sound leakage detection method of a wired headphone of the present invention;

[0031] Figure 2It is a schematic diagram of a leakage data acquisition terminal in a sound leakage detection system for wired headphones of the present invention;

[0032] Figure 3 It is a schematic diagram of a leakage state detection and analysis terminal in a sound leakage detection system for wired headphones of the present invention;

[0033] Figure 4 is a certain activation function diagram of the neural network model of the present invention;

[0034] Figure 5 is a schematic diagram of a person wearing a wired headset in the sound leakage detection system of the present invention;

[0035] Figure 6 It is a schematic diagram of the hollow condition of the sound leakage detection system for wearing wired headphones in the present invention; DETAILED DESCRIPTION

[0036] The invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0037] In order to solve the above technical problems, the present invention provides a sound leakage detection system and method for wired headphones.

[0038] In a first aspect of the present invention, a method for detecting sound leakage of a wired headset is provided, wherein the method comprises:

[0039] S1. Obtaining the first leakage data leaking from the outside into the cavity, and testing and obtaining the first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and obtaining the corresponding first sound leakage level;

[0040] In this embodiment, the pressure data is obtained by transmitting the pressure sensor built into the headset. The first leakage data is the difference between the volume in the ear canal and the volume set by the headset when the headset is worn. The second leakage data described later is also the difference between the volume in the ear canal and the volume set by the headset.

[0041] S2. Modally processing the first leakage data and the first pressure data collected to obtain first detection feature data;

[0042] S3. Constructing a sound leakage detection model using the first detection feature data and the first sound leakage level;

[0043] S4. Obtaining the second leakage data leaking from the outside into the cavity, and testing to obtain the second pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and modally processing the second leakage data and the second pressure data to obtain the second detection feature data;

[0044] S5. The sound leakage detection model processes the second detection feature data to obtain a second sound leakage level; and uses the second sound leakage level to detect the sound leakage status of the wired headphones.

[0045] The wired headphones in this embodiment are generally tested by an electroacoustic tester, a frequency sweeper, and a listening instrument.

[0046] Furthermore, the cavity is the gap between the non-contacting part of the tuning hole and the ear canal and the part from the external auditory canal to the eardrum when the user wears the wired headphone.

[0047] Furthermore, the sound leakage detection model is constructed using a GNN neural network model or an improved machine learning method.

[0048] Furthermore, the GNN neural network model adopts an S-GNN neural network model that adaptively improves the activation function, and its activation function σ(X) is:

[0049]

[0050] Where X is the input first detection feature data or second detection feature data, P 1 is the maximum preset pressure value of the contact surface between the earphone and the outer ear when the earphone is worn during model training, P 2 This is the optimal preset pressure value for the contact surface between the earphone and the outer ear when the earphone is worn, which is used during model training.

[0051] In this embodiment, since the training method of the GNN neural network model is a commonly used model for processing audio data in this field, the training steps and methods of other processes are not repeated here.

[0052] Furthermore, the improved machine learning method is constructed using a support vector machine model based on pressure transformation, and its specific function expression is:

[0053]

[0054] Where ω, b are the normal vector and intercept of the hyperplane, respectively, P 1 is the maximum preset pressure value of the contact surface between the earphone and the outer ear when the earphone is worn during model training. is the optimal preset pressure value of the contact surface between the earphone and the outer ear when the i-th user wears the earphone, is the sum of the optimal preset pressure values ​​of the contact surface between the earphone and the outer ear when the earphone is worn by the i-th user, X i is the input first detection feature data or the second detection feature data.

[0055] In this embodiment, if the value calculated by the support vector machine model based on pressure transformation is greater than 1, it means that the second sound leakage level is high. At this time, the wired headphones will issue a first-level warning, which is the highest-level warning. If the value calculated by the support vector machine model based on pressure transformation is equal to 1, it means that the second sound leakage level is low. At this time, the wired headphones will not issue a warning. If the value calculated by the support vector machine model based on pressure transformation is less than 1, it means that the second sound leakage level is medium, and a second-level warning will be issued.

[0056] A wired headset sound leakage detection system is also provided, comprising a leakage data acquisition terminal and a leakage status detection and analysis terminal, characterized in that:

[0057] The leakage data acquisition terminal includes a leakage parameter acquisition module and a leakage parameter processing module, and the leakage status detection and analysis terminal includes a leakage data storage module and a leakage detection module, wherein:

[0058] The leakage parameter acquisition module detects first leakage data leaking from the outside into the cavity, tests and obtains first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, obtains second leakage data leaking from the outside into the cavity, tests and obtains second pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and transmits the data to the leakage parameter processing module at the same time;

[0059] The leakage parameter processing module: performs modal processing on the first leakage data and the first pressure data collected to obtain first detection feature data; and performs modal processing on the second leakage data and the second pressure data collected to obtain second detection feature data;

[0060] The leakage data storage module is used to store the first detection feature data and the first sound leakage level corresponding thereto, and the second detection feature data and the second sound leakage level corresponding thereto;

[0061] The leakage detection module: constructs a sound leakage detection model using the first detection feature data stored in the leakage data storage module and the first sound leakage level corresponding thereto; receives the second detection feature data from the leakage data storage module or the leakage parameter processing module, and generates the second sound leakage level using the sound leakage detection model according to the second detection feature data.

[0062] Furthermore, the cavity is the gap between the non-contacting part of the tuning hole and the ear canal and the part from the external auditory canal to the eardrum when the user wears the wired headphone.

[0063] Furthermore, the sound leakage detection model is constructed using a GNN neural network model or an improved machine learning method.

[0064] Furthermore, the GNN neural network model adopts an S-GNN neural network model that adaptively improves the activation function, and its activation function σ(X) is:

[0065]

[0066] Where X is the input first detection feature data or second detection feature data, P 1 is the maximum preset pressure value of the contact surface between the earphone and the outer ear when the earphone is worn during model training, P 2 This is the optimal preset pressure value for the contact surface between the earphone and the outer ear when the earphone is worn, which is used during model training.

[0067] Furthermore, the improved machine learning method is constructed using a support vector machine model based on pressure transformation, and its specific function expression is:

[0068]

[0069] Where ω, b are the normal vector and intercept of the hyperplane, respectively, P 1 is the maximum preset pressure value of the contact surface between the earphone and the outer ear when the earphone is worn during model training. is the optimal preset pressure value of the contact surface between the earphone and the outer ear when the i-th user wears the earphone, is the sum of the optimal preset pressure values ​​of the contact surface between the earphone and the outer ear when the earphone is worn by the i-th user, X i is the input first detection feature data or the second detection feature data.

[0070] The present invention uses an improved machine learning model, only considers audio and external pressure parameters, and realizes modal fusion. The improved machine learning model is used for wired headphones to perform differentiated multi-dimensional level detection of sound leakage levels, providing a basis for subsequent improvement of headphone sound quality and user experience. The hollow audio reception used in the present invention can better reflect the degree of sound leakage, and the level classification processing method is more suitable for the scenario of wired headphones pressing on the outer ear. The detection of sound leakage status is accurate and timely. The present invention trains the model by utilizing the fusion characteristics of the pressure value of the contact between the limited headphone and the outer ear and the difference between the measured volume in the ear canal and the volume set by the headphone. Data is easy to obtain, and sound leakage detection is more accurate.

[0071] Of course, it can be understood that each embodiment of the present invention can achieve one of the effects alone, and a combination of multiple embodiments of the present invention can achieve all of the above effects, but it is not required that every embodiment of the present invention achieve all of the above advantages and effects, because each embodiment of the present invention can constitute a separate technical solution and make one or more contributions to the prior art.

[0072] The present invention does not particularly specify the structure of some modules, which shall be subject to the contents recorded in the prior art. The prior art mentioned in the above background technology section and the specific embodiment section of the present invention can be used as part of the present invention to understand the meaning of some technical features or parameters. The scope of protection of the present invention shall be subject to the contents actually recorded in the claims.

Claims

1. A method for detecting sound leakage of wired headphones, characterized in that: The method comprises: S1. Obtaining the first leakage data leaking from the outside into the cavity, and testing and obtaining the first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and obtaining the corresponding first sound leakage level; S2. Modally processing the first leakage data and the first pressure data collected to obtain first detection feature data; S3. Constructing a sound leakage detection model using the first detection feature data and the first sound leakage level; S4. Obtaining the second leakage data leaking from the outside into the cavity, and testing to obtain the second pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and modally processing the second leakage data and the second pressure data to obtain the second detection feature data; S5. The sound leakage detection model processes the second detection feature data to obtain a second sound leakage level; using the second sound leakage level to detect the sound leakage state of the wired headphones; The cavity is the gap between the external auditory canal and the eardrum where the tuning hole and the ear canal are not in contact when the user wears the wired headphone.

2. A method for detecting sound leakage of wired headphones as claimed in claim 1, characterized in that: The sound leakage detection model is constructed using a GNN neural network model or an improved machine learning method.

3. A method for detecting sound leakage of wired headphones as claimed in claim 2, characterized in that: The GNN neural network model adopts the S-GNN neural network model that adaptively improves the activation function.

4. A method for detecting sound leakage of wired headphones as claimed in claim 2, characterized in that: The improved machine learning method is constructed using a support vector machine model based on pressure transformation.

5. A wired headphone sound leakage detection system, comprising a leakage data acquisition terminal and a leakage status detection and analysis terminal, characterized in that: The leakage data acquisition terminal includes a leakage parameter acquisition module and a leakage parameter processing module, and the leakage status detection and analysis terminal includes a leakage data storage module and a leakage detection module, wherein: The leakage parameter acquisition module detects first leakage data leaking from the outside into the cavity, tests and obtains first pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, obtains second leakage data leaking from the outside into the cavity, tests and obtains second pressure data of the contact surface between the earphone and the outer ear when the earphone is worn, and transmits the data to the leakage parameter processing module at the same time; The leakage parameter processing module: performs modal processing on the first leakage data and the first pressure data collected to obtain first detection feature data; and performs modal processing on the second leakage data and the second pressure data collected to obtain second detection feature data; The leakage data storage module is used to store the first detection feature data and the first sound leakage level corresponding thereto, and the second detection feature data and the second sound leakage level corresponding thereto; The leakage detection module: constructs a sound leakage detection model using the first detection feature data stored in the leakage data storage module and the first sound leakage level corresponding thereto; receives the second detection feature data from the leakage data storage module or the leakage parameter processing module, and generates the second sound leakage level using the sound leakage detection model according to the second detection feature data; The cavity is the gap between the external auditory canal and the eardrum where the tuning hole and the ear canal are not in contact when the user wears the wired headphone.

6. A wired headphone sound leakage detection system as claimed in claim 5, characterized in that: The sound leakage detection model is constructed using a GNN neural network model or an improved machine learning method.

7. A wired headphone sound leakage detection system as claimed in claim 6, characterized in that: The GNN neural network model adopts the S-GNN neural network model that adaptively improves the activation function.

8. The sound leakage detection system for wired headphones as claimed in claim 6, characterized in that: The improved machine learning method is constructed using a support vector machine model based on pressure transformation.

Images