Automatic regulation and control method for audiometer audio oscillator

Abstract

The invention discloses an automatic regulation and control method for an audiometer audio oscillator, which comprises the following steps of: 1, carrying out real-time monitoring processing on a modulated carrier audio signal broadcasted by the audio oscillator, acquiring amplitude-modulated audio and carrier information of the audio oscillator, and calculating an amplitude-modulated value of the audio oscillator according to the audio and carrier information of the audio oscillator; and 2, presetting a target audio signal and a gain parameter of the target audio signal, and extracting a spectrum feature of the audio signal to be adjusted from the spectrum data according to the spectrum feature of the target audio signal. The method has the advantages that the audio oscillator is monitored in real time to obtain audio and carrier information, then the amplitude adjustment value of the audio oscillator is calculated, then the frequency point recognition model and the audio gain prediction model are obtained by extracting the spectrum features of the audio signal to be adjusted through training, and then the audio oscillator is automatically adjusted and controlled. Therefore, the audio quality of the audio oscillator is improved, and the audio oscillator is convenient for people to use.

Description

Technical Field

[0001] The present invention relates to the technical field of audiometer audio oscillators, and more specifically, to an automatic regulation method for an audiometer audio oscillator. Background Art

[0002] An audio oscillator is an electronic device used to generate audio signals. It can generate continuous audio waves and is commonly used in audio applications such as amplifiers, speakers, headphones, etc. The working principle of an audio oscillator is to utilize the oscillation of electronic components to generate audio signals. An audio oscillator usually consists of an oscillator circuit and an audio transducer. The oscillator circuit is the core part that generates audio signals, which can be an LC oscillator or an RC oscillator. The audio transducer then converts the signal generated by the oscillator into an audio signal suitable for driving a speaker.

[0003] Existing audio oscillators cannot be automatically regulated during use, resulting in a decline in the quality of audio signals and audibility, and unable to provide a stable sound.

[0004] Regarding the problems in the related art, no effective solution has been proposed yet. Summary of the Invention

[0005] The purpose of the present invention is to provide an automatic regulation method for an audiometer audio oscillator to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: An automatic regulation method for an audiometer audio oscillator includes the following steps:

[0007] Step 1: Perform real-time monitoring and processing on the modulated carrier audio signal broadcast by the audio oscillator, obtain the audio and carrier information of the modulation degree of the audio oscillator, and calculate the modulation degree value of the audio oscillator based on the audio and carrier information of the audio oscillator;

[0008] Step 2: Preset the target audio signal and the gain parameter of the target audio signal, and extract the spectral characteristics of the audio signal to be adjusted from the spectral data according to the spectral characteristics of the target audio signal;

[0009] Step 3: Construct a mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of audio signal to be adjusted, as well as a mapping relationship between the spectral characteristics of the target audio signal and the audio gain, and perform training to obtain a frequency point recognition model and an audio gain prediction model;

[0010] Step 4: Input the extracted spectral characteristics of the audio signal to be adjusted into the frequency point recognition model to identify and form a set of frequency points of the audio signal to be adjusted for sounding;

[0011] Step Five: Perform audio amplitude PID adjustment on the audio signal to be broadcast according to the frequency point recognition model and the audio gain prediction model, and deliver the audio signal after audio amplitude PID adjustment to the audio oscillator for broadcasting.

[0012] Further, the modulation depth value of the audio oscillator specifically refers to: the peak modulation depth value of the audio oscillator and the average modulation depth value within a preset time period.

[0013] Further, the steps of constructing the mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of the audio signal to be adjusted, and the mapping relationship between the spectral characteristics of the target audio signal and the audio gain further include the following steps:

[0014] Train the mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of the audio signal to be adjusted through the support vector machine algorithm to obtain a frequency point recognition model, obtain an audio gain prediction model according to the mapping relationship between the spectral characteristics of the target audio signal and the audio gain, and judge the frequency point set and the predicted audio gain.

[0015] Further, the specific steps of the frequency point recognition model and the audio gain prediction model performing audio amplitude adjustment on the audio signal to be broadcast include:

[0016] Deliver the frequency point recognition model and the audio gain prediction model to the input end of the audiometer volume controller, and use the audiometer volume controller to perform audio amplitude adjustment on the received audio signal to be broadcast. The input end of the audiometer volume controller is specifically divided into a volume reduction input end and a volume increase input end;

[0017] If the modulation depth of the audio oscillator is in an under-modulation state, deliver the under-frequency point recognition model and the audio gain prediction model to the volume increase input end, and increase the audio amplitude of the audio signal according to the amplitude corresponding to the audio gain prediction model until the peak modulation depth value is greater than the peak amplitude standard value, or the average modulation depth value is greater than the average amplitude standard value, and then stop delivering.

[0018] Further, after the spectral characteristics of the audio signal to be adjusted extracted are input into the frequency point recognition model and the frequency point set in the sound signal is recognized, it further includes:

[0019] Input the spectral characteristics of the target audio signal into the audio gain prediction model to predict the current actual vocal gain;

[0020] Extract the spectral characteristics of the audio signal to be adjusted and the spectral characteristics of the target audio signal in the audio signal, input the spectral characteristics of the audio signal to be adjusted into the frequency point recognition model to recognize the frequency point set, and input the spectral characteristics of the target vocal signal into the audio gain prediction model to obtain the current actual audio gain.

[0021] Further, the audio amplitude PID adjustment specifically includes:

[0022] Adjust the response speed of the audio output parameter U(t) through the proportional adjustment constant Kp, and adjust the audio fast response ability through kp. The audio adjustment parameter of the proportional link = kp * err(t), where Kp is the audio proportional adjustment constant and err(t) is the current audio error value;

[0023] Control the parameter adjustment of the actuator through the integral link, and accumulate the error by adding the previous audio output value U(t - 1) and kp * err(t);

[0024] Control the parameter adjustment of the actuator through the differential link. Adjust the change rate of the current value pv of the feedback through the differential audio adjustment constant Kd * (pv / sp) or Kd * (sp / pv). sp is the set value. When pv < sp, the audio adjustment parameter of the differential link = kd * (pv / sp) * (err(t) - err(t - 1)); when sp < pv, the audio adjustment parameter of the differential link = kd * (sp / pv) * (err(t) - err(t - 1)), where err(t - 1) is the previous error value;

[0025] Readjust the parameters of the actuator through the feedback of the sensor for the adjusted audio adjustment parameters of the proportional link, integral link, and differential link, and finally reach the set parameter value of the sensor to obtain a stable dynamic balance system.

[0026] Compared with the prior art, the present invention has the following beneficial effects: By real-time monitoring of the audio oscillator to obtain audio and carrier information, and then calculating the modulation amplitude value of the audio oscillator, and then training the frequency point recognition model and audio gain prediction model by extracting the spectral characteristics of the audio signal to be adjusted, the audio oscillator is automatically regulated, improving the audio stability of the audio oscillator, and thus improving its audio quality for easy use by personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0028] Figure 1 is the flowchart of an automatic regulation method for an audiometer audio oscillator according to an embodiment of the present invention Figure 1 ;

[0029] Figure 2 is the flow of an automatic regulation method for an audiometer audio oscillator according to an embodiment of the present invention Figure 2 ;

[0030] Figure 3 is the flow of an automatic regulation method for an audiometer audio oscillator according to an embodiment of the present invention Figure 3 。 Detailed implementation manners

[0031] Next, with reference to the accompanying drawings and specific implementation manners, the invention will be further described:

[0032] Embodiment:

[0033] Please refer to Figure 1 , an automatic regulation method for an audiometer audio oscillator according to an embodiment of the present invention includes the following steps:

[0034] Step 1: Perform real-time monitoring and processing on the modulated carrier audio signal broadcast by the audio oscillator, obtain the audio and carrier information of the modulation depth of the audio oscillator, and calculate the modulation depth value of the audio oscillator according to the audio and carrier information of the audio oscillator;

[0035] Step 2: Preset the target audio signal and the gain parameter of the target audio signal, and extract the spectral characteristics of the audio signal to be adjusted from the spectral data according to the spectral characteristics of the target audio signal;

[0036] Step 3: Construct a mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of audio signal to be adjusted, and a mapping relationship between the spectral characteristics of the target audio signal and the audio gain, and perform training to obtain a frequency point recognition model and an audio gain prediction model;

[0037] Step 4: Input the extracted spectral characteristics of the audio signal to be adjusted into the frequency point recognition model to identify and form a set of frequency points of the audio signal to be adjusted for sounding;

[0038] Step 5: Perform audio amplitude PID regulation on the audio signal to be broadcast according to the frequency point recognition model and the audio gain prediction model, and transmit the audio signal after audio amplitude PID regulation to the audio oscillator for broadcasting.

[0039] In a further embodiment, the modulation depth value of the audio oscillator specifically refers to: the peak modulation depth value of the audio oscillator and the average modulation depth value within a preset time period.

[0040] Real-time monitoring is mainly carried out by sampling and detecting the modulated carrier audio signal broadcast by the audio oscillator, so as to obtain the audio and carrier information of the audio oscillator. Finally, the modulation degree value of the audio oscillator is calculated according to the audio and carrier information of the audio oscillator. The modulation degree value is the peak modulation degree value of the audio oscillator and the average modulation degree value within a preset time period, so as to clarify the adjustment amplitude and prevent the adjustment amplitude from being too large.

[0041] In a further embodiment, the steps of constructing the mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of the audio signal to be adjusted, and the mapping relationship between the spectral characteristics of the target audio signal and the audio gain further include the following steps:

[0042] The mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of the audio signal to be adjusted is trained by the support vector machine algorithm to obtain a frequency point recognition model. According to the mapping relationship between the spectral characteristics of the target audio signal and the audio gain, an audio gain prediction model is obtained, and the frequency point set and the predicted audio gain are judged.

[0043] The mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of the audio signal to be adjusted is mainly trained by the vector machine algorithm to obtain a frequency point recognition model. Then, the mapping relationship between the spectral characteristics of the target audio signal and the audio gain is trained by the vector machine algorithm to obtain an audio gain prediction model. Subsequently, the frequency point recognition model and the audio gain prediction model are judged to improve the audio adjustment accuracy.

[0044] Such as Figure 3 As shown, in a further embodiment, the specific steps of the frequency point recognition model and the audio gain prediction model for adjusting the audio amplitude of the audio signal to be broadcast include:

[0045] The frequency point recognition model and the audio gain prediction model are sent to the input end of the audiometer volume controller, and the audiometer volume controller is used to adjust the audio amplitude of the received audio signal to be broadcast. The input end of the audiometer volume controller is specifically divided into a volume reduction input end and a volume increase input end;

[0046] If the modulation degree of the audio oscillator is in the under-modulation state, the under-frequency point recognition model and the audio gain prediction model are sent to the volume increase input end, and the audio amplitude of the audio signal is increased according to the amplitude corresponding to the audio gain prediction model until the peak modulation degree value is greater than the peak amplitude standard value, or the average modulation degree value is greater than the average amplitude standard value, and then the transmission is stopped.

[0047] Mainly, the received audio signal to be broadcasted will be adjusted in audio amplitude by using the volume controller of the audiometer. When the modulation depth of the audio oscillator is in the under-modulation state, the under-modulation frequency point recognition model and the audio gain prediction model will be sent to the volume increase input end, and the audio amplitude of the audio signal will be increased according to the amplitude corresponding to the frequency point recognition model and the audio gain prediction model until the peak modulation depth value is greater than the peak amplitude standard value, or the average modulation value is greater than the average amplitude standard value, then stop sending. For example, if the modulation depth of the audio oscillator is in the under-modulation state, the under-modulation frequency point recognition model and the audio gain prediction model will be sent to the volume increase input end to increase the amplitude of the output audio signal. If the pulse width of the under-modulation frequency point recognition model and the audio gain prediction model is 1s, the volume of the audio signal will increase by 1dB per second; if the pulse width of the under-modulation frequency point recognition model and the audio gain prediction model is 2s, the volume of the audio signal will increase by 2dB per second; if the pulse width of the under-modulation frequency point recognition model and the audio gain prediction model is 3s, the volume of the audio signal will increase by 3dB per second, and so on.

[0048] In a further embodiment, after the spectral features of the audio signal to be adjusted extracted are input into the frequency point recognition model and the frequency point set in the sound signal is recognized, it further includes:

[0049] Input the spectral features of the target audio signal into the audio gain prediction model to predict the current actual vocal gain;

[0050] By extracting the spectral features of the audio signal to be adjusted and the spectral features of the target audio signal, inputting the spectral features of the audio signal to be adjusted into the frequency point recognition model to recognize the frequency point set, and inputting the spectral features of the target vocal signal into the audio gain prediction model, the current actual audio gain is obtained, improving the audio gain effect.

[0051] As Figure 2 shown, in a further embodiment, the audio amplitude PID adjustment specifically includes:

[0052] The response speed of the audio output parameter U(t) is adjusted by the proportional adjustment constant Kp, and the audio fast response ability is adjusted by kp. The proportional link audio adjustment parameter = kp * err(t), where Kp is the audio proportional adjustment constant and err(t) is the current audio error value. Specifically, the larger Kp is, the more rapid the system's response to the error. For example, when the actual audio amplitude deviates from the target value, a larger Kp will immediately generate a stronger adjustment signal (such as gain or attenuation), forcing the system to quickly approach the target amplitude. Exemplarily, assume the target audio amplitude is 0 dB and the current actual amplitude is -6 dB: Error: err(t) = 0 dB - (-6 dB) = 6 dB, Proportional adjustment output: If Kp = 0.5, then U(t) = 0.5 × 6 = 3 dB. Effect: The system will immediately increase the gain by 3 dB, making the actual amplitude approach the target value. If there is still a residual error, the PID controller will continue to iterate and adjust.

[0053] The parameter adjustment of the actuator is controlled through the integral link. The error is accumulated by adding the previous audio output value U(t - 1) to kp * err(t). The core function of the integral link is to eliminate the steady-state error, which is specifically manifested as follows: When the proportional adjustment (P link) cannot completely eliminate the error (for example, there is a constant disturbance in the system), the integral term will accumulate the error value over time, gradually enhancing the control signal, and finally forcing the error to approach zero. For example, in audio amplitude control, if environmental noise causes the output amplitude to be continuously low, the integral term will continuously increase the gain until the actual amplitude is consistent with the target amplitude. In addition, the proportional adjustment (P) only responds to the current error. If there is a continuous disturbance in the system, the proportional adjustment may not be able to completely eliminate the error. The integral link gradually enhances the correction amount by continuously accumulating the error value, and finally eliminates the residual error.

[0054] The parameter adjustment of the actuator is controlled through the derivative link. The change rate of the current value pv of the feedback is adjusted by the derivative audio adjustment constant Kd * (pv / sp) or Kd * (sp / pv), where sp is the set value. When pv < sp, the derivative link audio adjustment parameter = kd * (pv / sp) * (err(t) - err(t - 1)); when sp < pv, the derivative link audio adjustment parameter = kd * (sp / pv) * (err(t) - err(t - 1)), where err(t - 1) is the previous error value. Exemplarily, if the audio amplitude suddenly increases, the derivative term detects that the amplitude increase rate is too fast and immediately reduces the gain to suppress the peak. If the amplitude slowly decreases, the derivative term weakens the adjustment strength to avoid overcompensation.

[0055] In other embodiments, assume an adaptive derivative gain is adopted:

[0056] Differential term = Kd * (err(t) - err(t - 1)) / Δt / (1 + |PV / SP|). When PV << SP, PV / SP → 0, the gain approaches Kd, and the differential action is normal, accelerating towards the target. When PV ≈ SP, PV / SP ≈ 1, the gain drops to Kd / 2 to suppress overshoot. When PV >> SP, PV / SP → ∞, the gain approaches 0 to avoid violent oscillations. Specifically, when the audio amplitude changes rapidly, Kd is temporarily increased; in the steady state, Kd is decreased to avoid tremors. The differential link provides the system with "forward-looking" adjustment ability by predicting the change trend of the error.

[0057] The adjusted proportional link audio adjustment parameters, integral link audio adjustment parameters, and differential link audio adjustment parameters are used to readjust the parameters of the actuator through the feedback of the sensor, and finally reach the set parameter value of the sensor, obtaining a stable dynamic balance system. In the audio amplitude control system, the PID controller dynamically adjusts the outputs of the proportional (P), integral (I), and differential (D) links by real-time collecting the feedback data of the sensor (such as the actual audio amplitude value PV) and combining it with the set value SP, and finally drives the actuator (such as a voltage-controlled amplifier, digital gain module, etc.) to achieve precise tracking and stability of the target parameters. Exemplarily, assume the target amplitude SP = -6dB and the current PV = -10dB (error err = 4dB): Immediately output P = Kp × 4dB to quickly increase the gain. If the error persists, the integral term I increases linearly with time, continuously increasing the gain. If the increase in gain causes PV to quickly approach SP, the differential term detects the rate of decrease in the error (d(err) / dt < 0) and reduces the gain adjustment amplitude in advance to prevent overshoot. Finally, the combined force of the three makes the system converge smoothly to -6dB and maintain stability.

[0058] Through the above solution of the present invention, the audio and carrier information are obtained by real-time monitoring of the audio oscillator, and then the modulation depth value of the audio oscillator is calculated. Subsequently, the spectral characteristics of the audio signal to be adjusted are extracted, and the frequency point recognition model and audio gain prediction model are obtained through training, and then the audio oscillator is automatically regulated to improve the audio stability of the audio oscillator, and thus improve its audio quality for easy use by personnel.

[0059] According to the above solution, the following two exemplary implementation scenarios are given:

[0060] Scenario 1: Sudden increase in background noise

[0061] The sensor detects a decrease in amplitude (PV drops from -6dB to -9dB), and the error err = 3dB.

[0062] The proportional term immediately increases the gain by 3 × Kp dB.

[0063] The integral term accumulates over time, continuously increasing the gain until the error is zero.

[0064] If the noise increase rate of the differential term slows down, reduce the gain adjustment range to avoid overshoot.

[0065] Scenario 2: The user suddenly increases the volume.

[0066] The sensor detects that the amplitude exceeds the limit (PV rises from -6 dB to -3 dB), and the error err = -3 dB.

[0067] The proportional term immediately reduces the gain by 3×Kp dB.

[0068] The differential term detects the amplitude soaring rate and further increases the attenuation strength.

[0069] The integral term gradually reduces the accumulated value to prevent over-suppression.

[0070] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic regulation method for an audiometer audio oscillator, characterized in that It includes the following steps: Step 1: Perform real-time monitoring and processing on the modulated carrier audio signal broadcast by the audio oscillator, obtain the audio and carrier information of the modulation degree of the audio oscillator, and calculate the modulation degree value of the audio oscillator according to the audio and carrier information of the audio oscillator; Step 2: Preset the target audio signal and the gain parameter of the target audio signal, and extract the spectral characteristics of the audio signal to be adjusted from the spectral data according to the spectral characteristics of the target audio signal; Step 3: Construct the mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of audio to be adjusted, and the mapping relationship between the spectral characteristics of the target audio signal and the audio gain, and perform training to obtain a frequency point recognition model and an audio gain prediction model; Step 4: Input the extracted spectral characteristics of the audio signal to be adjusted into the frequency point recognition model to identify and form a set of frequency points of the audio signal to be adjusted for sounding; Step 5: Perform audio amplitude PID adjustment on the audio signal to be broadcast according to the frequency point recognition model and the audio gain prediction model, and deliver the audio signal after audio amplitude PID adjustment to the audio oscillator for broadcasting.

2. The automatic regulation method of an audiometer audio oscillator according to claim 1, wherein The modulation degree value of the audio oscillator specifically refers to: the peak modulation degree value of the audio oscillator and the average modulation degree value within a preset time period.

3. A method for automatically regulating an audiometer audio oscillator according to claim 1, characterized in that, The construction of the mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of audio to be adjusted, and the mapping relationship between the spectral characteristics of the target audio signal and the audio gain further includes the following steps: Train the mapping relationship between the spectral characteristics of the audio signal to be adjusted and the type of audio to be adjusted through the support vector machine algorithm to obtain a frequency point recognition model, obtain an audio gain prediction model according to the mapping relationship between the spectral characteristics of the target audio signal and the audio gain, and judge the set of frequency points and the predicted audio gain.

4. The automatic regulation method of an audiometer audio oscillator according to claim 1, characterized in that, The frequency point recognition model and the audio gain prediction model perform audio amplitude adjustment on the audio signal to be broadcast specifically including: Deliver the frequency point recognition model and the audio gain prediction model to the input end of the audiometer volume controller, and use the audiometer volume controller to perform audio amplitude adjustment on the received audio signal to be broadcast. The input end of the audiometer volume controller is specifically divided into a volume reduction input end and a volume increase input end; If the modulation degree of the audio oscillator is in an under-modulation state, deliver the under-frequency point recognition model and the audio gain prediction model to the volume increase input end, and increase the audio amplitude of the audio signal according to the amplitude corresponding to the audio gain prediction model until the peak modulation degree value is greater than the peak amplitude standard value, or the average modulation degree value is greater than the average amplitude standard value, then stop delivering.

5. A method for automatically regulating an audiometer audio oscillator according to claim 1, characterized in that, After inputting the extracted spectral characteristics of the audio signal to be adjusted into the frequency point recognition model and identifying the set of frequency points in the sound signal, it further includes: Input the spectral characteristics of the target audio signal into the audio gain prediction model to predict the current actual sounding gain; By extracting the spectral features of the audio signal to be adjusted and the spectral features of the target audio signal, inputting the spectral features of the audio signal to be adjusted into the frequency point recognition model to identify the frequency point set, and inputting the spectral features of the target sound signal into the audio gain prediction model to obtain the current actual audio gain.

6. A method for automatically regulating an audiometer audio oscillator according to claim 1, characterized in that The specific audio amplitude PID adjustment includes: Adjust the response speed of the audio output parameter U(t) through the proportional adjustment constant Kp, and adjust the audio fast response ability through kp. The audio adjustment parameter of the proportional link = kp * err(t), where Kp is the audio proportional adjustment constant and err(t) is the current audio error value; Control the parameter adjustment of the actuator through the integral link, and accumulate the error by adding the previous audio output value U(t - 1) and kp * err(t); Control the parameter adjustment of the actuator through the derivative link, and adjust the change rate of the current value pv of the feedback through the derivative audio adjustment constant Kd * (pv / sp) or Kd * (sp / pv). sp is the set value. When pv < sp, the audio adjustment parameter of the derivative link = kd * (pv / sp) * (err(t) - err(t - 1)); when sp < pv, the audio adjustment parameter of the derivative link = kd * (sp / pv) * (err(t) - err(t - 1)), where err(t - 1) is the previous error value; Readjust the parameters of the actuator through the feedback of the sensor for the adjusted audio adjustment parameters of the proportional link, integral link, and derivative link, and finally reach the set parameter value of the sensor to obtain a stable dynamic balance system.

Images