Mute detection method and device, audio equipment and computer readable storage medium

Abstract

The invention discloses a mute detection method and device, audio equipment and a computer readable storage medium, and the method is applied to the audio equipment, and comprises the steps: carrying out the framing processing of an audio input signal, and calculating the energy characteristic quantity of each analysis frame in a preset time window; intercepting a plurality of time periods in a preset time window, and determining a background noise estimation value corresponding to each time period based on the energy characteristic quantity of the plurality of analysis frames in each time period; obtaining statistical characteristic parameters according to the plurality of background noise estimated values; dynamically generating a mute detection threshold according to the background noise estimation value and the statistical characteristic parameters; and in a preset time condition, comparing the energy characteristic quantity corresponding to the real-time analysis frame with a mute detection threshold value, and judging whether the audio channel to be detected is in a mute state or not. Therefore, the silence detection threshold adaptively reflects the stability or fluctuation of the noise level, and the accuracy and anti-jitter capability of silence detection in the non-stationary noise environment are effectively improved in combination with the preset time condition.

Description

Technical Field

[0001] This application relates to the technical field of headphones, and more specifically, to a method, apparatus, audio device, and computer-readable storage medium for detecting silence. Background Technology

[0002] In audio devices such as headphones, speakers, and audio broadcast transmitters, silence detection (SD) is typically required to achieve functions such as automatic mute, power saving, automatic pause playback, and audio source switching. Related technologies determine a mute state by acquiring the energy of the audio signal and identifying a state where the energy consistently falls below a certain fixed threshold.

[0003] However, due to the different background noise levels in the audio front-end circuits of audio devices, the different port noise characteristics of different terminals connected to the audio devices, and the dynamic changes in the environmental noise and usage scenarios of the audio devices, the accuracy of detecting whether the audio device is in a silent state by using a fixed threshold is not ideal.

[0004] Inaccurate mute detection will directly affect the user experience of audio devices' automatic mute, power saving control, automatic pause playback, automatic audio source switching, and broadcast idle detection functions, resulting in unnecessary power consumption and incorrect state switching. Summary of the Invention

[0005] In view of the above problems, this application proposes a silence detection method, device, audio equipment and computer-readable storage medium, which can effectively improve the accuracy of audio equipment status detection, thereby reducing unnecessary power consumption and erroneous state switching.

[0006] In a first aspect, embodiments of this application provide a silence detection method applied to an audio device, comprising: performing frame-by-frame processing on an audio input signal from an audio channel to be detected, and calculating the energy characteristic of each analysis frame within a preset time window; extracting multiple time periods within the preset time window, and determining the background noise estimate corresponding to each time period based on the energy characteristic of multiple analysis frames within each time period; performing statistical analysis on the background noise of the audio channel to be detected based on the multiple background noise estimates, and obtaining statistical characteristic parameters used to characterize the stability or fluctuation range of the noise level; dynamically generating a silence detection threshold based on the background noise estimate and the statistical characteristic parameters; and comparing the energy characteristic of the real-time analysis frame with the silence detection threshold within a preset time condition to determine whether the audio channel to be detected is in a silent state.

[0007] Secondly, embodiments of this application also provide a silence detection device, which is applied to an audio device and includes: a calculation module for performing frame-by-frame processing on the audio input signal from the audio channel to be detected and calculating the energy feature quantity of each analysis frame within a preset time window; a determination module for extracting multiple time periods within the preset time window and determining the background noise estimate corresponding to each time period based on the energy feature quantities of multiple analysis frames within each time period; an estimation module for performing statistical analysis on the background noise of the audio channel to be detected based on the multiple background noise estimates to obtain statistical feature parameters for characterizing the stability or fluctuation range of the noise level; a generation module for dynamically generating a silence detection threshold based on the background noise estimate and the statistical feature parameters; and a judgment module for comparing the energy feature quantity corresponding to the real-time analysis frame with the silence detection threshold within a preset time condition to determine whether the audio channel to be detected is in a silent state.

[0008] Thirdly, embodiments of this application also provide an audio device, including a processor, a memory, and one or more application programs; the one or more application programs are stored in the memory and configured to be executed by the processor to implement the above-described silence detection method.

[0009] Fourthly, embodiments of this application also provide a computer-readable storage medium storing program code, wherein the above-described silence detection method is executed when the program code is run by a processor.

[0010] The technical solution provided in this application is applied to audio equipment. The method includes: performing frame-by-frame processing on the audio input signal from the audio channel to be detected, and calculating the energy feature quantity of each analysis frame within a preset time window; extracting multiple time periods within the preset time window, and determining the background noise estimate corresponding to each time period based on the energy feature quantities of multiple analysis frames within each time period; performing statistical analysis on the background noise of the audio channel to be detected based on the multiple background noise estimates to obtain statistical feature parameters used to characterize the stability or fluctuation range of the noise level; dynamically generating a silence detection threshold based on the background noise estimate and the statistical feature parameters; and comparing the energy feature quantity corresponding to the real-time analysis frame with the silence detection threshold within a preset time condition to determine whether the audio channel to be detected is in a silent state. Therefore, by estimating the background noise separately within multiple time periods and dynamically generating the silence detection threshold based on the statistical distribution of these estimates, the silence detection threshold can adaptively reflect the stability or fluctuation level of the noise level; simultaneously, by combining the preset time condition for state judgment, the accuracy and anti-jitter capability of silence detection in non-stationary noise environments are effectively improved. This solves the problem that existing technologies are prone to misjudgment or missed judgment when the noise level is unstable, and improves the environmental adaptability and reliability of silent detection. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments and drawings obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0012] Figure 1 A flowchart illustrating a noise detection method provided in an embodiment of this application is shown.

[0013] Figure 2 A flowchart illustrating another noise detection method provided in an embodiment of this application is shown.

[0014] Figure 3 A schematic diagram of a noise detection device provided in an embodiment of this application is shown.

[0015] Figure 4 A schematic diagram of the structure of an audio device provided in an embodiment of this application is shown.

[0016] Figure 5 This illustration shows a schematic diagram of the structure of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0017] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0018] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0019] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0020] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0021] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] References to "one embodiment" or "some embodiments" in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0023] In audio devices such as headphones, speakers, and audio broadcast transmitters, in order to achieve functions such as automatic mute, power saving control, automatic pause playback, and audio source switching, it is usually necessary to perform silence detection (SD) on the input audio signal.

[0024] In related technologies, silence detection typically employs an energy detection algorithm based on a fixed threshold. This algorithm calculates the energy or amplitude of the audio signal within a preset time window, and determines the state as silent when the energy remains below a certain fixed threshold.

[0025] However, since the audio front-end circuits of audio devices (such as audio power amplifier circuits, analog signal front-end processing circuits, auxiliary audio input interfaces, audio connection cables, electrical connectors, etc.) have different background noise levels, the silence judgment based on a certain fixed threshold is inaccurate.

[0026] For example, if the fixed threshold value is set too low, noise may be mistakenly identified as valid audio on panels with high background noise. Conversely, if the fixed threshold value is set too high, low-volume programs may be mistakenly identified as muted on panels with low background noise.

[0027] Furthermore, when audio devices are connected to different terminals via their Aux interfaces, the noise floor, power supply ripple, radio frequency interference, and grounding environment of their output signals vary significantly. The same set of fixed thresholds often exhibits inconsistent performance across different terminals, sometimes requiring individual adjustments for each terminal, resulting in extremely high maintenance costs.

[0028] Furthermore, during the use of audio devices, ambient noise, electromagnetic interference, and power supply conditions dynamically change over time, causing the noise floor of the input signal to be unstable. A silence detection scheme using a fixed threshold cannot track these changes, resulting in significant drift and jitter in the silence detection results, leading to inaccurate silence assessments.

[0029] Inaccurate mute detection directly affects the user experience of audio devices' automatic mute, power saving control, automatic pause playback, automatic audio source switching, and broadcast idle detection functions, resulting in unnecessary power consumption and incorrect state switching.

[0030] To address the aforementioned issues, this application provides a silence detection method, apparatus, audio device, and computer-readable storage medium. The method, applied to an audio device, includes: performing frame-by-frame processing on an audio input signal from an audio channel to be detected, and calculating the energy characteristic of each analysis frame within a preset time window; extracting multiple time periods within the preset time window, and determining the estimated background noise value corresponding to each time period based on the energy characteristic values ​​of multiple analysis frames within each time period; performing statistical analysis on the background noise of the audio channel to be detected based on the multiple background noise estimates to obtain statistical characteristic parameters characterizing the stability or fluctuation range of the noise level; dynamically generating a silence detection threshold based on the background noise estimates and the statistical characteristic parameters; and comparing the energy characteristic of the real-time analysis frame with the silence detection threshold within a preset time condition to determine whether the audio channel to be detected is in a silent state.

[0031] Therefore, by estimating background noise over multiple time periods and dynamically generating a silence detection threshold based on the statistical distribution of these estimates, the silence detection threshold can adaptively reflect the stability or fluctuation of the noise level. Simultaneously, by combining preset time conditions for state judgment, the accuracy and anti-vibration capability of silence detection in non-stationary noise environments are effectively improved. This solves the problem of misjudgment or missed judgment in existing technologies when noise levels are unstable, and improves the environmental adaptability and reliability of silence detection.

[0032] This invention provides a method for detecting silence. The execution subject of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, or an audio device. In other words, the silence detection method can be executed by software or hardware installed on a terminal device or server device, etc. The software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0033] Specifically, please refer to Figure 1 , Figure 1This illustration shows a flowchart of a silence detection method provided in an embodiment of this application, which can be applied to audio devices. Figure 1 As shown, the method may include steps 110 to 150.

[0034] In step 110, the audio input signal from the audio channel to be detected is processed by frame segmentation, and the energy feature of each analysis frame is calculated within a preset time window.

[0035] Audio devices include, but are not limited to: Bluetooth headsets, smart speakers, wireless audio receivers, wired speakers, USB audio interfaces, LEAudio broadcast transmitters, conference system terminals, hearing aids, or in-vehicle audio head units. In other words, audio devices broadly refer to any electronic device with an audio input channel that needs to determine in real time whether a valid audio signal is present, especially suitable for audio product scenarios that are sensitive to power consumption, have inherent noise, or are subject to variable noise characteristics in their operating environment.

[0036] The audio channel to be detected can be a logical or physical audio path in an audio device that requires mute status determination. The audio channel to be detected can receive audio data streams from external audio sources or internal processing modules, and serve as the processing object of the mute detection method of this invention, i.e., the audio input signal.

[0037] In some implementations, the audio channel to be detected includes, but is not limited to: analog audio input channels (e.g., AUX interface), digital audio input channels (e.g., USBAudio), intermediate audio buses in the device's internal DSP processing chain, or a single audio stream channel in LEAudio broadcasting.

[0038] The audio input signal can be a raw or processed digital audio sample sequence acquired from the audio channel to be detected, and is usually represented as a discrete-time signal. ( (For sampling point index). The audio input signal is acquired by the audio device from the audio channel to be detected at a fixed period (e.g., every 5~20 milliseconds) to ensure the real-time performance and timing consistency of silence detection.

[0039] A preset time window is used as a sliding statistical window for background noise estimation in subsequent steps. For example, the audio input signal is first divided into frames of a fixed frame length (e.g., 10 milliseconds), and the energy feature of each frame is calculated independently. Then, a time window of a preset length (e.g., containing N=300 frames, about 2 seconds) is maintained to cache the most recent energy feature values ​​so that subsequent steps can determine the background noise estimate based on these energy features.

[0040] In other words, the preset time window is essentially a historical data window on which background noise modeling is based. Its length is much longer than the duration of a single frame, with a typical value of hundreds of milliseconds to several seconds.

[0041] After acquiring the audio input signal, the audio device divides the continuous audio input signal into multiple analysis frames of fixed length according to time. Each frame contains a preset number of sampling points (for example, at a sampling rate of 16kHz, each frame can contain 160 sampling points, corresponding to a frame length of 10ms).

[0042] For each analysis frame within a preset time window, its energy feature is calculated to characterize the signal strength of that frame. Optionally, the sum of the squares of the amplitudes of all sampling points within each frame is defined as the energy feature corresponding to each analysis frame. Optionally, the energy feature is the root mean square (RMS) value of each analysis frame, obtained by averaging the squares of the amplitudes of all sampling points within that frame and then taking the square root. Optionally, the energy feature is the logarithmic energy, obtained by taking the logarithm of the frame energy, used to convert the signal energy into a logarithmic domain representation, facilitating matching with human auditory perception characteristics and threshold setting.

[0043] However, since the original frame energy (or RMS) can fluctuate drastically even under pure noise due to randomness, window function effects, quantization errors, etc., the energy feature can be further smoothed by a first-order IIR filter or a moving average filter to obtain a smoothed energy feature value E (E_smooth). This effectively suppresses the energy jump caused by instantaneous signal fluctuations or short-term interference, and significantly improves the stability of background noise estimation and the accuracy of silence detection.

[0044] Therefore, by segmenting the audio input signal into frames and calculating the energy characteristics of each frame within a preset time window, the continuous time-domain signal is transformed into a quantifiable and comparable discrete energy sequence, providing a stable, real-time, and time-resolution feature basis for subsequent background noise modeling and silence state determination.

[0045] In step 120, multiple time periods are extracted within a preset time window, and the background noise estimate corresponding to each time period is determined based on the energy characteristics of multiple analysis frames within each time period.

[0046] Within a relatively long preset time window, it is divided into several shorter sub-time periods. A background noise estimate is calculated independently for each time period to obtain background noise estimates for multiple time periods. In subsequent steps, the distribution of these multiple background noise estimates is used to determine whether the noise is stable.

[0047] Background noise estimates can be obtained by low-energy filtering and statistical aggregation based on the energy characteristics of analysis frames within a certain time period, and are used to characterize the background noise level of the environment within that time period.

[0048] Specifically, in some implementations, the step "determining the background noise estimate for each time period based on the energy characteristics of multiple analysis frames within each time period" may include the following steps: (1) Select a target energy feature that is less than a preset value from the energy feature values ​​of multiple analysis frames in each time period; (2) Determine the background noise estimate for each time period based on the mean of the target energy characteristic.

[0049] The preset value can be set based on the default noise level of the audio device (e.g., -50dBFS) or dynamically given by a coarse noise estimate from the previous stage. In one specific implementation, the device maintains a time window of a preset length, such as N frames, counts multiple smoothed energy feature values ​​E within this time window, selects a subset of samples with lower energy, calculates their mean, and obtains the current background noise estimate.

[0050] For example, when performing background noise estimation, the audio device selects the 20% of frames with the lowest energy from the energy characteristics of each time period, or all frames below the preset value, and calculates the mean of the selected samples as the background noise estimate for each time period.

[0051] By retaining only low-energy frames for noise estimation, the pollution of the noise model by high-energy frames containing speech, music, or sudden interference is effectively eliminated, thus obtaining a purer and more stable background noise representation.

[0052] In step 130, based on multiple background noise estimates, the background noise of the audio channel to be detected is statistically analyzed to obtain statistical characteristic parameters that characterize the stability or fluctuation range of the noise level.

[0053] Statistical characteristic parameters include, but are not limited to: the mean of multiple background noise estimates (which reflects the average level of recent background noise), the variance of multiple background noise estimates (which reflects the dispersion of noise energy, i.e., the magnitude of fluctuation), kurtosis, or quantile interval (which is used to characterize the sharpness of the noise distribution or the sensitivity to outliers).

[0054] In one specific implementation, statistical analysis is performed on multiple background noise estimates over a longer time scale to calculate the mean μ and standard deviation σ of the noise estimates, or to calculate the corresponding quantiles, thus forming noise statistical characteristics. By performing statistical analysis on multiple background noise estimates, silence detection is upgraded from a "static threshold" to a "dynamic context awareness," solving the problem of misjudgment in scenarios such as panel background noise fluctuations and external intermittent interference.

[0055] In step 140, a silence detection threshold is dynamically generated based on the estimated background noise value and statistical characteristic parameters.

[0056] The silence detection threshold can be an energy comparison benchmark used to determine whether the current audio frame is in a silent state.

[0057] Specifically, in some implementations, the step of "dynamically generating a silence detection threshold based on the estimated background noise value and statistical characteristic parameters" may include the following steps: (1) Calculate the mean and variance of multiple background noise estimates; (2) The silence detection threshold is generated by using the threshold formula, the mean and variance of multiple background noise estimates.

[0058] In one specific implementation, the expression for the threshold formula can be: in, The threshold for detecting silence. The mean of multiple background noise estimates. This is the scaling factor (which can be set via registers or configuration tables, with a value range of 1.5 to 3.0). The variance of multiple background noise estimates.

[0059] When the background noise is stable ( (Smaller), the silence detection threshold is close to the noise mean, which is beneficial for capturing weak speech; when there are significant fluctuations in noise ( (Larger), the silence detection threshold is automatically raised, effectively avoiding misjudging noise peaks as valid audio signals.

[0060] Therefore, by combining the threshold generation method of mean and variance, the accuracy, stability and robustness of silence detection in complex acoustic environments are significantly improved.

[0061] In addition, in some embodiments, the silence detection method may further include: when the number of background noise estimates used to generate statistical feature parameters is insufficient, generating a silence detection threshold based on the sum of the most recently obtained background noise estimates and a preset offset.

[0062] To ensure that the audio device can still work properly during the initialization phase or in scenarios with insufficient data, when the number of background noise estimates used to calculate statistical feature parameters (such as mean and variance) is less than the preset minimum number of samples (for example, the preset minimum number of samples N_min=10), the audio device automatically switches to the backup threshold generation mode.

[0063] In generation mode, the silence detection threshold is generated based on the sum of the most recently obtained valid background noise estimate and a preset offset, which can be expressed as: in, The threshold for detecting silence. This is the most recently obtained effective background noise estimate. The preset offset can be set via register or configuration table, with a value range corresponding to an energy margin of 3 to 6 dB.

[0064] The device continuously collects energy characteristic samples until the number of background noise estimates reaches the minimum statistical sample size, at which point it automatically switches back to a dynamic threshold calculation mode based on statistical features, such as ( ) This processing logic effectively solves the problem of insufficient statistical samples causing the inability to calculate the silence detection threshold in scenarios such as the initial startup of audio devices, hot-plugging of audio sources, or reactivation after a long period of silence. It ensures the continuity and availability of the silence detection function, while avoiding misjudgment or failure due to the lack of a silence detection threshold.

[0065] In step 150, within a preset time condition, the energy feature quantity corresponding to the real-time analysis frame is compared with the silence detection threshold to determine whether the audio channel to be detected is in a silent state.

[0066] The preset time condition does not refer to a fixed moment, but rather to the time accumulation condition based on multiple consecutive analysis frames. Specifically, the system will not immediately determine the silence state based solely on the comparison result of a single frame's energy and the silence detection threshold. Instead, it requires that the comparison result continuously meet the silence condition within several consecutive frames before the silence state can be confirmed.

[0067] For example, if the minimum number of consecutive frames required for mute confirmation is set to M (e.g., M=50, corresponding to approximately 500 milliseconds), then the system will only determine the audio channel to be detected as mute when the energy characteristic values ​​of M consecutive frames are all lower than the current mute detection threshold.

[0068] Furthermore, to further improve the stability of the state output, in some implementations, the step of "comparing the energy feature quantity corresponding to the real-time analysis frame with the silence detection threshold to determine whether the audio channel to be detected is in a silent state" may include the following steps: (1)When the energy eigenvalue of consecutive M frames is less than the silence detection threshold, it is determined that the audio channel to be detected is in the silent state; (2)When the energy eigenvalue of consecutive K frames is greater than or equal to the silence detection threshold, it is determined that the audio channel to be detected is in the non-silent state; where M and K are positive integers, and M≥K≥1.

[0069] When it is detected that the smoothed energy eigenvalue of consecutive M frames is less than the current silence detection threshold, the audio device determines that the audio channel to be detected enters the silent state; when it is detected that the smoothed energy eigenvalue of consecutive M frames is greater than or equal to the current silence detection threshold, the audio device determines that the audio channel to be detected exits the silent state, that is, enters the non-silent state.

[0070] In a specific embodiment, the device compares the real-time smoothed energy eigenvalue E with the silence detection threshold T. When consecutive M frames satisfy E<T, the current channel is determined to be in the silent state; when consecutive K frames satisfy E≥T, the current channel is determined to be in the non-silent state, where M≥K≥1. The hang time or hysteresis mechanism is used to prevent frequent switching of the silent state.

[0071] By setting M to a value greater than K, it can be ensured that silence is only confirmed when there is continuous lack of valid audio, avoiding misjudging short speech gaps as silence, and once a valid audio signal is detected, the audio device can quickly respond and exit the silent state, ensuring that the starting part of the speech is not truncated.

[0072] In some embodiments, the silence detection method further includes: when it is determined that the audio channel to be detected is in the silent state, sending a status notification message, updating the silence status flag bit in the audio control register, or embedding silence indication information in the metadata field of the audio data stream.

[0073] That is to say, the silent state can send a status notification message to the internal tasks or driver modules of the system through the software interface. For example, in an embedded audio platform based on a real-time operating system, the silence detection module can send event messages (such as the SILENCE_DETECTED event) to the audio management task, the power management unit, or the Bluetooth protocol stack to notify them that there is no valid signal in the current audio channel. This message can be used to trigger low-power scheduling, turn off unnecessary peripherals, or adjust link parameters.

[0074] Audio devices can also update the mute status flag in the audio control register. This register, typically located in the audio subsystem of the audio codec, DSP control interface, or SoC, contains one or more dedicated bits (such as SILENCE_FLAG). Hardware logic or firmware can read this flag via polling or interrupts to determine the current mute status. For example, DAC control logic can automatically shut down the analog output stage when this flag is detected as set, preventing noise floor leakage.

[0075] Mute information can also be embedded in the metadata fields of the audio data stream for transmission. In audio protocols that support in-band metadata (e.g., ...), , ,or The mute state can be encapsulated as control information along with the audio sample. Specifically, in... In broadcast application scenarios, this metadata field is specifically manifested as Defined in the message Sign. According to ( According to the provisions of ), "" indicates that the current broadcast stream contains valid audio content, while " "Indicates no valid audio. Therefore, when this invention determines that it is in a silent state, the broadcast source device will actively set this flag to "mute". This enables the receiving end (e.g., Headphones or hearing aids can thus pause audio decoding, reduce radio frequency monitoring, or enter a light sleep mode, thereby achieving end-to-end energy saving.

[0076] In some implementations, once the audio device confirms that the audio channel under test is muted, it can trigger a series of energy-saving control operations related to the audio processing chain to reduce system power consumption and suppress background noise output. These energy-saving control operations include, but are not limited to: reducing the digital signal processor's output (to reduce the gain of subsequent amplification stages on the background noise), shutting down the digital-to-analog converter (to avoid continuously outputting background noise or consuming quiescent current when there is no valid audio), or, in applications supporting low-power audio broadcasting, the silence detection result of this invention can be used to dynamically control the status indicator field in the broadcast data stream to notify the receiver that there is currently no valid audio content, allowing it to enter a low-power monitoring mode or pause audio rendering.

[0077] For example, when the audio device confirms that the audio channel to be tested is muted, it will... Carried in broadcast stream Field set to Conversely, when valid audio content is detected, it is set to... Among them, the fields yes The standard metadata defined in the specification is used to send data to receiving devices (such as...). (Headphones, hearing aids, or speakers) transmit information in real time whether the current audio stream contains valid audio content.

[0078] Furthermore, to address scenarios where the environmental noise characteristics change significantly during the operation of audio equipment, in some implementations, the silence detection method further includes the following steps: (1) When the audio device starts up or detects that there is no significant audio content for a long time, it enters the port self-learning process; (2) During the port self-learning process, collect several background noise estimates and initialize or correct the silence detection threshold based on the background noise estimates to adapt to the noise characteristics of the currently connected terminal.

[0079] When the audio device starts up or detects that there is no significant audio content for a long time, the audio device clears the historical noise statistics cache, and re-determines the background noise estimate based on the newly acquired audio input signal, and updates the statistical feature parameters and silence detection threshold accordingly.

[0080] In one specific embodiment, energy characteristics of the current channel are continuously collected over a certain period of time, assuming the content is mainly background noise; the collected data are statistically analyzed to obtain a more robust noise estimate and its statistical characteristics; a silence detection threshold T is initialized based on this noise estimate, thereby making the threshold adaptive to the currently accessed PC / mobile phone. Ports and corresponding panel hardware; during subsequent operation, the threshold is continuously fine-tuned in conjunction with the dynamic update algorithm in Implementation Example 1.

[0081] Through this self-learning mechanism, even if different users connect to different models or brands of PCs or mobile phones, the device can still automatically adapt to the corresponding noise characteristics without the need for manual reset of the mute threshold.

[0082] By resetting the noise model, audio devices can quickly adapt to new acoustic environments, avoiding misjudgments or omissions of silence due to the use of outdated noise estimates, thereby significantly improving long-term operational reliability and user experience.

[0083] In addition, when the audio device is in the startup phase (e.g., during power-on initialization or the initial running period after firmware loading), the audio device enters the working state from a no-signal state, and the initial noise floor is unknown; when the audio device detects the access of an audio source (e.g., successful Bluetooth pairing, AUX cable insertion, or completion of USB audio enumeration), the audio device enters the port self-learning process.

[0084] In one specific embodiment, the present invention can be applied to various electronic devices with audio processing capabilities, such as wireless headphones, smart speakers, hearing aids, Bluetooth broadcast transmitters, etc. The following... (For example, , , , Bluetooth audio SoC ( Taking this platform as an example, this chip is widely used in low-power audio devices that support LE Audio, serving as their core audio processing and wireless communication unit. Its core lies in the "dual-differential judgment combined with adaptive reference energy update" mechanism.

[0085] Please see Figure 2 , Figure 2 This document illustrates a flowchart of another noise detection method provided in an embodiment of this application, as shown below. Figure 2 As shown, the method may include steps 210 to 260.

[0086] In step 210, the energy of the left and right channels is collected in each detection cycle and smoothed to obtain the energy feature of the current frame.

[0087] The energy E_cur of the left and right channels is collected in each detection cycle T (e.g., 1000ms) and smoothed (e.g., median filtering or moving average). The current energy E_cur,L / R represents the energy of the left and right channels in the current detection cycle.

[0088] In step 220, if the energy feature of the current frame is lower than the preset mute candidate threshold, then the mute candidate stage is entered.

[0089] If the energy of both channels is lower than the mute candidate threshold E_sil, then the candidate stage is entered. E_sil represents the mute candidate energy threshold and is regarded as the upper limit of the energy of the "mute candidate". It can be preset to 70.

[0090] In step 230, during the silence candidate stage, the first difference between the energy feature of the current frame and the energy feature of the previous frame, and the second difference between the energy feature of the current frame and the adaptive reference energy feature are calculated.

[0091] The energy feature of the previous frame is E_last. The first difference is D_last = E_cur - E_last; the second difference is D_ref = max(E_cur - E_ref, 0). E_ref is used as an adaptive reference value to characterize the trend of environmental noise floor.

[0092] In step 240, if the first difference is less than the first threshold and the second difference is less than the second threshold, the mute cumulative count is accumulated; otherwise, the mute cumulative count is cleared to zero.

[0093] The first threshold refers to the first-order difference threshold, which is the energy difference threshold from the previous frame, denoted as LV1: the threshold for comparing the energy difference between "the current and the last time". The second threshold refers to the second-order difference threshold, which is the energy difference threshold from the reference energy, denoted as LV2: the threshold for comparing the energy difference between "the current and the reference". If it is calculated that D_last < LV1 and D_ref < LV2, then the cumulative silence count C_sil is accumulated; otherwise, it is cleared.

[0094] In step 250, if the cumulative silence count reaches the preset cumulative threshold, it is determined that the audio channel enters the silent state, and the adaptive reference energy feature quantity is updated to the current frame energy feature quantity.

[0095] The silence accumulation count is the periodic count that continuously satisfies the silence condition, denoted as C_sil. The preset cumulative threshold is the silence accumulation count threshold, denoted as C_max. If C_sil ≥ C_max, it is determined to be in the silent state, and E_ref is updated to the current energy. C_max can be preset to 60s.

[0096] In step 260, in the silent state, if the current frame energy feature quantity is not lower than the silence candidate threshold, or the first difference is not less than the first threshold, or the second difference is not less than the second threshold, it is determined that the audio channel resumes to the non-silent state.

[0097] In the silent state, if E_cur ≥ E_sil or D_last ≥ LV1 or D_ref ≥ LV2, the non-silent state is immediately resumed.

[0098] In some embodiments, the step "if the cumulative silence count reaches the preset cumulative threshold, it is determined that the audio channel enters the silent state, and the adaptive reference energy feature quantity is updated to the current frame energy feature quantity" may further include: updating the adaptive reference energy feature quantity according to the current frame energy feature quantity and its median filtering result.

[0099] If E_ref > E_cur, then E_ref = E_cur, indicating following the decrease; if the median filtering result E_med - E_ref ≥ LV2, then E_ref = E_med, indicating resisting the sudden noise lift.

[0100] In In the audio device constructed based on It is responsible for receiving audio data streams from Bluetooth RF decoding or local analog / digital inputs and running a silence detection algorithm on its internal DSP, thereby providing the entire device with real-time, adaptive silence status judgment capabilities. The judgment result is further used to control the device's output behavior (such as turning off the amplifier) ​​and protocol interactions (such as setting...). Through power management strategies, system-level goals such as noise reduction, energy saving, and user experience optimization are ultimately achieved.

[0101] therefore, This is not merely an example of a general-purpose processor, but rather a main control SoC for a typical audio device, fully embodying all the functions of this invention, from signal acquisition and feature calculation to status output and linkage control.

[0102] Please see Figure 3 , Figure 3 This illustration shows a schematic diagram of a silence detection device provided in an embodiment of this application, applied to an audio device. The silence detection device 300 includes: a calculation module 310, a determination module 320, an estimation module 330, a generation module 340, and a judgment module 350. Specifically: The calculation module 310 is used to perform frame-by-frame processing on the audio input signal from the audio channel to be detected, and to calculate the energy feature of each analysis frame within a preset time window. The determination module 320 is used to extract multiple time periods within a preset time window and determine the background noise estimate corresponding to each time period based on the energy characteristics of multiple analysis frames within each time period. The estimation module 330 is used to perform statistical analysis on the background noise of the audio channel to be detected based on multiple background noise estimates, and obtain statistical characteristic parameters to characterize the stability or fluctuation range of the noise level. The generation module 340 is used to dynamically generate a silence detection threshold based on the background noise estimate and statistical characteristic parameters. The judgment module 350 is used to compare the energy feature quantity corresponding to the real-time analysis frame with the silence detection threshold within a preset time condition to determine whether the audio channel to be detected is in a silent state.

[0103] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described device and module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0104] In the several embodiments provided in this application, the coupling or direct coupling or communication connection between the modules shown or discussed may be an indirect coupling or communication connection through some interface, device or module, and may be electrical, mechanical or other forms.

[0105] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0106] Please see Figure 4 , Figure 4 The diagram shows a structural schematic of an audio device provided in an embodiment of this application. The audio device 400 in this application may include one or more of the following components: a processor 410, a memory 420, and one or more application programs. The one or more application programs may be stored in the memory 420 and configured to be executed by one or more processors 410. The one or more programs are configured to perform the silence detection method as described in the foregoing method embodiments.

[0107] Processor 410 may include one or more processing cores. Processor 410 connects to various parts within the audio device 400 using various interfaces and lines, and performs various functions and processes data of the audio device 400 by running or executing instructions, programs, code sets, or instruction sets stored in memory 420, and by calling data stored in memory 420. Optionally, processor 410 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). Processor 410 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem may also not be integrated into processor 410, but may be implemented separately using a communication chip.

[0108] The memory 420 may include random access memory (RAM) or read-only memory (ROM). The memory 420 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 420 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function, instructions for implementing the various method embodiments described below, etc. The data storage area may also store data created by the audio device 400 during use.

[0109] Please see Figure 5 , Figure 5 The diagram shows a computer-readable storage medium 500 provided in an embodiment of this application. The computer-readable storage medium 500 stores program code, which can be called by a processor to execute the silence detection method described in the above method embodiment.

[0110] The computer-readable storage medium 500 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium 500 includes a non-transitory computer-readable storage medium. The computer-readable storage medium 500 has storage space for program code 510 that performs any of the method steps described above. This program code can be read from or written to one or more computer program devices. The program code 510 may, for example, be compressed in a suitable form.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for detecting silence, characterized in that, Applied to an audio device, the method includes: The audio input signal from the audio channel to be detected is processed by frame segmentation, and the energy feature of each analysis frame is calculated within a preset time window; Multiple time periods are extracted within a preset time window, and the background noise estimate for each time period is determined based on the energy characteristics of multiple analysis frames within each time period. Based on multiple background noise estimates, statistical analysis is performed on the background noise of the audio channel to be detected to obtain statistical characteristic parameters that characterize the stability or fluctuation range of the noise level. Based on the estimated background noise value and statistical characteristic parameters, a silent detection threshold is dynamically generated. Within a preset time condition, the energy feature quantity corresponding to the real-time analysis frame is compared with the silence detection threshold to determine whether the audio channel to be detected is in a silent state.

2. The silence detection method according to claim 1, characterized in that, The step of determining the estimated background noise value for each time period based on the energy feature values ​​of multiple analysis frames within each time period includes: From the energy feature values ​​of multiple analysis frames within each time period, select a target energy feature value that is less than a preset value. Based on the mean of the target energy characteristic, the estimated value of the background noise corresponding to each time period is determined.

3. The silence detection method according to claim 1, characterized in that, The step of dynamically generating a silence detection threshold based on the estimated background noise value and statistical characteristic parameters includes: Calculate the mean and variance of the multiple background noise estimates; The silence detection threshold is generated using a threshold formula and the mean and variance of the multiple background noise estimates. The expression for the threshold formula is: in, The silence detection threshold is... The mean of the multiple background noise estimates. This is the proportionality coefficient. The variance of the multiple background noise estimates is given.

4. The silence detection method according to claim 1, characterized in that, The method further includes: When the number of background noise estimates used to generate the statistical feature parameters is insufficient, the silence detection threshold is generated based on the sum of the most recently obtained background noise estimates and a preset offset.

5. The silence detection method according to claim 1, characterized in that, The step of comparing the energy feature quantity corresponding to the real-time analysis frame with the silence detection threshold to determine whether the audio channel to be detected is in a silent state includes: When the energy characteristic values ​​of M consecutive frames are all less than the silence detection threshold, the audio channel to be detected is determined to be in a silent state. When the energy characteristic values ​​of K consecutive frames are all greater than or equal to the silence detection threshold, the audio channel to be detected is determined to be in a non-silent state. Where M and K are positive integers, and M K.

6. The silence detection method according to claim 1, characterized in that, The method further includes: When the audio device starts up or detects that there is no significant audio content for a long time, it enters the port self-learning process. During the port self-learning process, several background noise estimates are collected, and the silence detection threshold is initialized or corrected based on the background noise estimates to adapt to the noise characteristics of the currently connected terminal.

7. The silence detection method according to claim 1, characterized in that, The method further includes: When it is determined that the audio channel to be detected is in a mute state, a status notification message is sent, the mute status flag in the audio control register is updated, or mute indication information is embedded in the metadata field of the audio data stream.

8. A noise detection device, characterized in that, Applied to an audio device, the device includes: The calculation module is used to perform frame-by-frame processing on the audio input signal from the audio channel to be detected, and to calculate the energy feature of each analysis frame within a preset time window. The determination module is used to extract multiple time periods within a preset time window and determine the background noise estimate for each time period based on the energy characteristics of multiple analysis frames within each time period. The estimation module is used to perform statistical analysis on the background noise of the audio channel to be detected based on multiple background noise estimates, and obtain statistical characteristic parameters to characterize the stability or fluctuation range of the noise level. The generation module is used to dynamically generate a silence detection threshold based on the background noise estimate and statistical characteristic parameters. The judgment module is used to compare the energy feature quantity corresponding to the real-time analysis frame with the silence detection threshold within a preset time condition to determine whether the audio channel to be detected is in a silent state.

9. An audio device, characterized in that, include: One or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications being configured to perform the silence detection method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code, which can be called by a processor to execute the silence detection method as described in any one of claims 1-7.

Images