Communication method and system

Abstract

According to the communication method and system provided by the invention, the shock test signal and the common voice signal can be accurately distinguished through a dual discrimination mechanism of performing time domain statistical parameter analysis and frequency domain entropy calculation on the to-be-detected signal. Due to the fact that accurate recognition of the signal types is achieved, targeted energy suppression processing can be only carried out on the recognized shock test signals, common voice signals are prevented from being mistakenly processed, and the problem that the voice signals are damaged due to the fact that an extremely low suppression threshold value needs to be set in a traditional DRC method is solved. Besides, the method of combining frame-by-frame time domain characteristic analysis with frequency domain characteristic analysis is adopted, and the characteristics that the shock test signal is closer to Gaussian distribution in a short time and the frequency domain entropy value is low are utilized, so that the identification accuracy of the shock test signal is improved.

Description

Technical Field

[0001] The present invention belongs to the field of communication technology, and particularly relates to a communication method and system. Background Art

[0002] Shock (impact or vibration) certification is one type of acoustic performance certification performed by carriers. Using the CMW500 (a device that emits radio frequency signals and establishes a call with a mobile phone), the phone's receiver plays a single-frequency tone signal. The phone is then placed on a simulated human head in a specific position (where the simulated ear receives the maximum signal under the same conditions), simulating a handheld call. The simulated ear receives the signal and transmits it to the ACQUA (a software system for voice and audio quality measurement and analysis) acoustic test system to calculate the sound pressure level, thereby determining whether the phone's call performance meets the shock standard.

[0003] DRC (dynamic range control) can generally set thresholds for the required upper and lower limits respectively. The upper threshold means that when the signal amplitude exceeds the set threshold, the signal will be suppressed, also known as large signal suppression; the lower threshold means that when the signal amplitude is lower than the set threshold, the signal will be suppressed, also known as small signal suppression.

[0004] However, traditional signal suppression methods use DRC, which requires a very low threshold for large signal suppression to effectively process shock test signals. This low threshold setting severely damages normal voice signals, resulting in significant distortion. Subjectively, the sound becomes noticeably quieter and less rich, while objectively, the maximum amplitude decreases from 0dB to approximately -6dB. This processing method fails to meet carrier certification requirements while ensuring voice call quality, impacting the user experience. Summary of the Invention

[0005] The present invention provides a communication method and system, which can maintain the integrity of a voice signal while effectively suppressing a shock test signal.

[0006] The present invention provides a communication method, comprising:

[0007] Performing time domain characteristic analysis on the signal to be detected frame by frame, and calculating time domain statistical parameters of the signal to be detected;

[0008] When the time domain statistical parameter conforms to the Gaussian distribution characteristics, performing frequency domain characteristic analysis on the signal to be detected, and calculating the frequency domain entropy value of the signal to be detected;

[0009] When the frequency domain entropy value is lower than a first set threshold, the signal to be detected is determined to be a shock test signal, and the energy of the shock test signal is weakened.

[0010] Furthermore, when the frequency domain entropy value is lower than a first set threshold, determining that the signal to be detected is a shock test signal, and weakening the energy of the shock test signal includes:

[0011] When the number of the shock test signals is greater than a second set threshold,

[0012] Converting the shock test signal into a frequency domain signal;

[0013] defining a frequency band of the shock test signal based on a spectrum mask;

[0014] performing a spectrum subtraction operation within the target frequency band marked by the spectrum mask;

[0015] The processed shock test signal is restored to the time domain.

[0016] Furthermore, the shock test signal is converted into a frequency domain signal based on a Fourier algorithm, and the processed shock test signal is restored to the time domain.

[0017] Furthermore, when the frequency domain entropy value is lower than a first set threshold, determining that the signal to be detected is a shock test signal, and weakening the energy of the shock test signal further includes:

[0018] When the number of the shock test signals is less than the second set threshold,

[0019] generating an anti-phase signal having the same waveform amplitude and opposite phase as the shock test signal according to the center frequency of the spectrum and the Q value of the shock test signal;

[0020] The inverted signal and the shock test signal are superimposed.

[0021] Furthermore, the method further comprises:

[0022] Calculating the energy spectrum integral value of the shock test signal in real time;

[0023] When the energy spectrum integral values ​​are all lower than a third set threshold value within a preset time window, it is determined that the shock test signal disappears.

[0024] Furthermore, the method further comprises:

[0025] When it is detected that the shock test signal disappears, the signal energy suppression process is stopped and the full-band processing mode is switched.

[0026] Furthermore, the signal energy suppression process is stopped based on the cosine gain attenuation algorithm.

[0027] Furthermore, when the signal to be detected does not conform to the Gaussian distribution characteristics, the signal to be detected is transmitted to the noise reduction algorithm module for processing;

[0028] Alternatively, when the frequency domain entropy value is higher than the first set threshold, it is determined that the signal to be detected is transmitted to a noise reduction algorithm module for processing.

[0029] On the other hand, the present invention further discloses a communication system, comprising:

[0030] A first signal processing module is configured to perform frame-by-frame time domain characteristic analysis on the signal to be detected, and when the time domain statistical parameters of the signal to be detected conform to the Gaussian distribution characteristics, perform frequency domain characteristic analysis on the signal to be detected, and calculate the frequency domain entropy value of the signal to be detected;

[0031] The second signal processing module is used to compare the frequency domain entropy value of the signal to be detected with a first set threshold value, and when the frequency domain entropy value is lower than the first set threshold value, determine that the signal to be detected is a shock test signal and weaken the energy of the shock test signal.

[0032] Furthermore, the second signal processing module includes an adaptive IIR notch filter.

[0033] Compared with the prior art, the present invention has at least the following technical effects:

[0034] The present invention can accurately distinguish shock test signals from ordinary voice signals through a dual discrimination mechanism of performing time domain statistical parameter analysis and frequency domain entropy value calculation on the signal to be detected. Due to the accurate identification of the signal type, the present invention can only perform targeted energy suppression processing on the identified shock test signal, while the ordinary voice signal is prevented from being misprocessed, thus avoiding the problem of voice signal damage caused by the need to set an extremely low suppression threshold in the traditional DRC method. In addition, the present invention improves the recognition accuracy of the shock test signal by adopting a method of combining frame-by-frame time domain characteristic analysis with frequency domain characteristic analysis, taking advantage of the characteristics that the shock test signal is closer to the Gaussian distribution in a shorter time and has a lower frequency domain entropy value. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a waveform diagram of the shock authentication signal in Example 1;

[0036] Figure 2 is a waveform diagram of a normal speech signal in Example 1;

[0037] Figure 3 This is the energy spectrum diagram of the shock test signal in Example 1;

[0038] Figure 4 This is an energy spectrum diagram of a normal speech signal in Example 1;

[0039] Figure 5 This is a simplified flow chart of the communication method in Example 1;

[0040] Figure 6 This is another schematic diagram of the communication method in Example 1. DETAILED DESCRIPTION

[0041] The following description of a communication method and system of the present invention is provided in conjunction with a schematic diagram, which illustrates a preferred embodiment of the present invention. It should be understood that those skilled in the art may modify the present invention described herein while still achieving the advantageous effects of the present invention. Therefore, the following description should be understood as a general guide for those skilled in the art and is not intended to limit the present invention.

[0042] The following paragraphs describe the present invention in more detail by way of example with reference to the accompanying drawings. The advantages and features of the present invention will become more apparent from the following description. It should be noted that the drawings are greatly simplified and not to exact scale, and are provided solely for the purpose of assisting in the description of the embodiments of the present invention.

[0043] Example 1

[0044] The inventors found that the conventional dynamic range compression DRC (dynamic range control) method used in the prior art to process shock test signals has technical contradictions. Figure 1 The waveform of the shock authentication signal shown and Figure 2In the waveform of a normal speech signal shown, the shock certification signal amplitude is only around -12dB, significantly lower than the 0dB threshold for a normal speech signal ("normal speech signal" refers to the core useful signal emitted by the human voice in communication scenarios and carries the actual conversation content). To ensure that the shock test signal triggers DRC processing, the DRC (dynamic range control) large signal suppression threshold must be set very low. Although the shock test signal amplitude itself is not large, due to the human ear's varying sensitivity to different frequencies, the acoustic signals at certain frequencies, when collected by the artificial ear and calculated by the ACQUA (software system for voice and audio quality measurement and analysis) test system, can be very large. Without suppression, the signal would fail to meet carrier certification standards. However, when the DRC large signal processing threshold is set low enough to process the shock test signal, it severely damages the normal speech signal. Many large-amplitude speech signals are over-suppressed, resulting in severe speech distortion. Subjectively, the sound is noticeably quieter and less rich. Objectively, the maximum amplitude of the speech signal decreases from 0dB to around -6dB.

[0045] By analyzing the characteristic differences between the two types of signals, the inventors found that the above technical contradictions can be solved by using signal recognition. Figure 3 and Figure 4 , Figure 3 is the energy spectrum of the shock test signal, from Figure 4 It can be seen that the energy of the shock test signal is concentrated in a short bandwidth of about 100 Hz in the frequency domain and is relatively stable. Figure 4 The energy of the ordinary speech signal shown in the figure fluctuates greatly in the frequency domain and is widely distributed. From a statistical perspective, the speech signal is closer to a non-Gaussian distribution due to its burstiness and sparsity, while the shock test signal is closer to a Gaussian distribution in a shorter period of time. In addition, the spectral entropy (complexity) of the speech signal is significantly higher than that of the shock test signal.

[0046] Based on the above-mentioned characteristic differences, the inventors have designed a communication method, please refer to Figure 5-Figure 6 , the method comprising:

[0047] S1 is to be detected signal frame by frame time domain characteristics analysis, calculate the time domain statistical parameters of the signal to be detected;

[0048] S2. When the time domain statistical parameters conform to the Gaussian distribution characteristics, the frequency domain characteristics of the signal to be detected are analyzed to calculate the frequency domain entropy value of the signal to be detected;

[0049] S3. When the frequency domain entropy value is lower than a first set threshold, determine that the signal to be detected is a shock test signal, and weaken the energy of the shock test signal.

[0050] This embodiment can accurately distinguish shock test signals from ordinary voice signals through a dual discrimination mechanism of time domain statistical parameter analysis and frequency domain entropy value calculation of the signal to be detected. Due to the accurate identification of the signal type, the present invention can only perform targeted energy suppression processing on the identified shock test signal, while the ordinary voice signal is prevented from being misprocessed, thus avoiding the problem of voice signal damage caused by the need to set an extremely low suppression threshold in the traditional DRC method. In addition, the present invention improves the recognition accuracy of the shock test signal by adopting a method of combining frame-by-frame time domain characteristic analysis with frequency domain characteristic analysis, taking advantage of the characteristics that the shock test signal is closer to the Gaussian distribution in a shorter time and has a lower frequency domain entropy value.

[0051] In step S1, the signal processing system first performs frame processing on the input signal, and the length of each frame can be set to 10-30ms, for example, 10ms, 13ms, 20ms, 26ms and 30ms.

[0052] Furthermore, the time domain statistical parameters include indicators such as mean, variance and skewness, etc. Of course, those skilled in the art can also select different time domain statistical parameters according to actual conditions, and no specific limitation is made here.

[0053] In step S2, when the time domain statistical parameters conform to the Gaussian distribution characteristics, it means that when the statistical parameters of multiple consecutive frames conform to the Gaussian distribution, it indicates that the signal may contain a shock test signal.

[0054] Furthermore, in step S3, when the frequency domain entropy value of the shock test signal is lower than a first set threshold, determining that the signal to be detected is a shock test signal, and weakening the energy of the shock test signal includes:

[0055] When the number of the shock test signals is greater than a second set threshold:

[0056] S31. The shock test signal is converted into a frequency domain signal;

[0057] S33 defines the frequency band of the shock test signal based on the spectrum mask;

[0058] S34 performs a spectrum subtraction operation within the target frequency band marked by the spectrum mask;

[0059] S35. Restoring the processed shock test signal to the time domain.

[0060] Specifically, after the shock test signal is detected, the signal can be framed and windowed first, for example, using a Hamming window to reduce spectral leakage. In step S31, a fast Fourier transform is performed on each frame of the signal to convert the time domain waveform into a frequency domain complex representation containing real and imaginary parts. In the frequency domain processing stages of S33 and S34, the target frequency band can be selectively attenuated according to a preset spectrum mask, for example, a spectrum subtraction operation is performed in the range of 1kHz to 3kHz. In step S35, after the frequency domain processing is completed, the modified frequency domain data is reconstructed into a time domain signal through an inverse fast Fourier transform, and finally a smooth transition between frames is achieved through the overlap-addition method to avoid phase jumps in the processed signal.

[0061] Through the above technical solution, this application can effectively suppress shock test signals while maintaining the integrity of normal speech signals. Due to the frequency selectivity of frequency domain processing, interfering components can be accurately eliminated without reducing the overall signal amplitude, avoiding the speech dynamic range compression and signal distortion problems caused by traditional DRC methods. The processed speech signal maintains the natural and smooth acoustic characteristics in subjective listening.

[0062] Furthermore, in step S3, when the frequency domain entropy value of the shock test signal is lower than a first set threshold, determining that the signal to be detected is a shock test signal, and weakening the energy of the shock test signal further includes:

[0063] When the number of the shock test signals is less than a second set threshold:

[0064] S32. Generates an inverted signal having the same amplitude and opposite phase as the shock test signal waveform according to the center frequency and Q value of the shock test signal spectrum;

[0065] S33. Superimpose the inverted signal and the shock test signal.

[0066] It is understandable that in step S32, the center frequency of the spectrum refers to the frequency point where the energy of the signal is most concentrated in the frequency domain distribution, which can be obtained by calculating the weighted average of the signal spectrum, for example, by using the energy center of gravity method or the peak search algorithm. The Q value refers to the quality factor of the filter, which is used to characterize the proportional relationship between the filter bandwidth and the center frequency, and can be achieved by setting the pole position of the filter transfer function. The anti-phase signal refers to a signal with the same amplitude as the original signal and a phase difference of 180°, which can be generated by performing an inversion operation on the original signal through a digital signal processing algorithm.

[0067] In step S33, superposition involves adding the inverted signal to the original signal in the time domain. This can be accomplished using an adder or point-by-point addition in digital signal processing. For example, when a single pulse-type shock test signal is detected, the system automatically calculates its center frequency to be 1.8kHz, sets the Q value to 10, generates an inverted signal with the corresponding parameters, and then precisely aligns and superimposes it in the time domain to achieve signal suppression within a specific frequency band.

[0068] In this embodiment, when the number of shock test signals is small, a precisely matched inverted signal is generated to effectively suppress a localized frequency band, avoiding the voice distortion caused by over-suppression in traditional methods. This solution maintains the original dynamic characteristics and frequency domain integrity of the voice signal while ensuring the effectiveness of shock test signal processing, resolving the conflict between degraded voice quality and certification requirements in existing technologies.

[0069] In a specific embodiment, the second set threshold is generally set to a smaller integer range, preferably 2-4. When the number of impact signals is within this range, the system has sufficient computing resources to perform accurate inverse superposition suppression on each signal, thereby ensuring the integrity of the speech signal to the greatest extent.

[0070] Furthermore, the method of this embodiment also includes:

[0071] The energy spectrum integral value of the shock test signal is calculated in real time. When the energy spectrum integral value is lower than a third set threshold value within a preset time window, it is determined that the shock test signal disappears.

[0072] The energy spectrum integral value is the sum of signal energy within the target frequency band, obtained through integration. This value is used to quantitatively assess whether the signal strength has attenuated to an undetectable level. The preset time window is a fixed duration for continuous signal monitoring. This can be achieved using a sliding window mechanism combined with timestamps to ensure the stability of signal extinction determination.

[0073] Specifically, the system continuously acquires time-domain signals within the target frequency band, converts each frame into a frequency-domain energy distribution, and then accumulates the energy within the specific frequency band. If the integral value obtained over multiple consecutive calculation cycles does not exceed a predefined threshold, it determines that there is no valid shock test signal within the frequency band. At this point, the current suppression algorithm is immediately terminated, and the system switches to normal speech processing mode.

[0074] Through the above technical solution, the present application can stop the signal processing operation in time while effectively suppressing the shock test signal machine, avoid the normal voice signal from being continuously suppressed, solve the voice distortion problem caused by the fixed threshold setting of the traditional method, and improve the acoustic quality perception in the call scenario.

[0075] In one specific embodiment, the third threshold is an adaptive value dynamically calculated based on the current ambient background noise level, with a preset protection margin added. When the ambient background noise level is high, the dynamically calculated third threshold can be adjusted higher accordingly. When the ambient background noise level is low, the third threshold can be adjusted lower accordingly. This allows the system to more sensitively detect when the signal has attenuated to an imperceptible level, thereby promptly terminating the suppression operation.

[0076] Furthermore, the method further comprises:

[0077] When it is detected that the shock test signal disappears, the signal energy suppression process is stopped and the full-band processing mode is switched to.

[0078] Specifically, when the monitoring value continuously output by the energy spectrum integrator remains below the third set threshold within the sliding time window, the signal extinction judgment mechanism is triggered. At this time, the spectrum subtraction operation or the inverted signal superposition processing is terminated, and the target frequency band marked by the frequency domain mask is unlocked. The signal processing process automatically switches to the full-band equalization mode, and by restoring the passband range of the broadband filter, it achieves lossless transmission of the full-band signal while maintaining the ability to monitor residual interference. This mode switching process is achieved through a smooth transition through the gain control module to avoid auditory discomfort caused by sudden signal changes.

[0079] Through the above technical solution, this embodiment can accurately identify the disappearance of the shock test signal and promptly restore full-band signal processing, eliminating the voice quality degradation caused by continuous signal suppression in traditional methods, and maintaining the integrity and clarity of the call signal while meeting the certification test pass rate.

[0080] In a specific embodiment, the signal energy suppression process is stopped based on the cosine gain decay algorithm.

[0081] The cosine gain reduction algorithm is a signal processing method that uses a cosine function to smoothly adjust the gain coefficient. Specifically, this can be achieved by using a piecewise cosine curve to nonlinearly reduce the gain value. During the signal suppression stop phase, the algorithm gradually transitions the gain coefficient from a suppressed state to a normal state, avoiding sudden changes in signal amplitude.

[0082] Specifically, when the shock test signal disappears, the gain coefficient of the signal suppression module is initialized to the value corresponding to the suppression state. Subsequently, within a preset time window, the cosine gain reduction algorithm is triggered, dynamically calculating the gain adjustment for each frame of the signal based on the cosine function curve.

[0083] Furthermore, when the signal to be detected does not conform to the Gaussian distribution characteristics, the signal to be detected is transmitted to the noise reduction algorithm module for processing; or, when the frequency domain entropy value is higher than the first set threshold, it is determined that the signal to be detected is transmitted to the noise reduction algorithm module for processing.

[0084] It is understandable that when the time domain statistical parameters of the signal to be detected deviate from the Gaussian distribution characteristics, it indicates that it may contain non-impact test signal components, such as conventional speech or environmental noise. At this time, it is directly transferred to the noise reduction algorithm module for general noise reduction processing. When the frequency domain entropy value is higher than the first set threshold, it indicates that the signal frequency domain energy distribution is relatively dispersed, for example, it contains multi-band speech or broadband noise. At this time, it is also transferred to the noise reduction algorithm module for processing. By dynamically judging the signal characteristics, the suppression operation is only started when the presence of an impact test signal is confirmed. In other cases, the conventional noise reduction process is used.

[0085] Compared to existing technologies, traditional methods require setting the dynamic range compression threshold too low to cover the impact test signal, resulting in excessive suppression of the amplitude of normal speech signals. This solution distinguishes signal types and activates targeted suppression only when necessary, avoiding unnecessary suppression of normal speech signals.

[0086] Example 2

[0087] This embodiment discloses a communication system to implement the communication method disclosed in the first embodiment. The system includes:

[0088] A first signal processing module is configured to perform frame-by-frame time domain characteristic analysis on the signal to be detected, and when the time domain statistical parameters of the signal to be detected conform to the Gaussian distribution characteristics, perform frequency domain characteristic analysis on the signal to be detected, and calculate the frequency domain entropy value of the signal to be detected;

[0089] The second signal processing module is used to compare the frequency domain entropy value of the signal to be detected with a first set threshold value, and when the frequency domain entropy value is lower than the first set threshold value, determine that the signal to be detected is a shock test signal and weaken the energy of the shock test signal.

[0090] It is understandable that the technical effects that can be achieved by the communication method disclosed in the first embodiment can also be achieved by the communication system disclosed in the second embodiment, and will not be described in detail here.

[0091] In one specific implementation, the second signal processing module includes an adaptive IIR (Infinite Impulse Response) notch filter.

[0092] An adaptive IIR notch filter is an infinite impulse response filter that dynamically adjusts its center frequency and bandwidth based on the input signal's characteristics. It's implemented using a second-order transfer function structure, with its transfer function coefficients updated in real time based on the target frequency band. By automatically tracking the center frequency of the shock test signal, the filter creates a narrowband suppression region in the frequency domain, accurately eliminating interference at specific frequencies.

[0093] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A communication method, characterized in that: The method comprises: Performing time domain characteristic analysis on the signal to be detected frame by frame, and calculating time domain statistical parameters of the signal to be detected; When the time domain statistical parameter conforms to the Gaussian distribution characteristics, performing frequency domain characteristic analysis on the signal to be detected, and calculating the frequency domain entropy value of the signal to be detected; When the frequency domain entropy value is lower than a first set threshold, the signal to be detected is determined to be a shock test signal, and the energy of the shock test signal is weakened.

2. The communication method according to claim 1, wherein: When the frequency domain entropy value is lower than a first set threshold, determining that the signal to be detected is a shock test signal, and weakening the energy of the shock test signal includes: When the number of the shock test signals is greater than a second set threshold, Converting the shock test signal into a frequency domain signal; defining a frequency band of the shock test signal based on a spectrum mask; performing a spectrum subtraction operation within the target frequency band marked by the spectrum mask; The processed shock test signal is restored to the time domain.

3. The communication method according to claim 2, wherein: The shock test signal is converted into a frequency domain signal based on a Fourier algorithm, and the processed shock test signal is restored to the time domain.

4. The communication method according to claim 1, wherein: When the frequency domain entropy value is lower than a first set threshold, determining that the signal to be detected is a shock test signal, and weakening the energy of the shock test signal further includes: When the number of the shock test signals is less than the second set threshold, generating an anti-phase signal having the same waveform amplitude and opposite phase as the shock test signal according to the center frequency of the spectrum and the Q value of the shock test signal; The inverted signal and the shock test signal are superimposed.

5. The communication method according to claim 1, wherein: The method further comprises: Calculating the energy spectrum integral value of the shock test signal in real time; When the energy spectrum integral values ​​are all lower than a third set threshold value within a preset time window, it is determined that the shock test signal disappears. The communication method according to claim 5 , wherein: The method further comprises: When it is detected that the shock test signal disappears, the signal energy suppression process is stopped and the full-band processing mode is switched.

7. The communication method according to claim 6, wherein: Stop signal energy suppression based on the cosine gain attenuation algorithm.

8. The communication method according to claim 1, wherein: When the signal to be detected does not conform to the Gaussian distribution characteristics, the signal to be detected is transmitted to the noise reduction algorithm module for processing; Alternatively, when the frequency domain entropy value is higher than the first set threshold, it is determined that the signal to be detected is transmitted to a noise reduction algorithm module for processing.

9. A communication system, characterized in that: The system comprises: A first signal processing module is configured to perform frame-by-frame time domain characteristic analysis on the signal to be detected, and when the time domain statistical parameters of the signal to be detected conform to the Gaussian distribution characteristics, perform frequency domain characteristic analysis on the signal to be detected, and calculate the frequency domain entropy value of the signal to be detected; The second signal processing module is used to compare the frequency domain entropy value of the signal to be detected with a first set threshold value, and when the frequency domain entropy value is lower than the first set threshold value, determine that the signal to be detected is a shock test signal and weaken the energy of the shock test signal.

10. The communication system according to claim 9, wherein The second signal processing module includes an adaptive IIR notch filter.

Images