Voice volume dynamic control method for VOIP voice stream

Abstract

The invention discloses a voice volume dynamic control method for a VOIP voice stream, and belongs to the field of VoIP. The method specifically comprises the following steps: firstly, optimally designing a voice volume dynamic control module S by utilizing an existing DRC algorithm; for a calling VoIP phone, adding the optimized module S into an audio effector chain of the phone, and meanwhile, adding parameters contained in the module S into an audio configuration file; then, according to scene requirements, a user selects a mode for adding in advance in a parameter list, and after silence of small voice through a noise gate, processing of a compressor and an expander according to the selected mode, volume compensation and an amplitude limiter are carried out in sequence, the processing sequence of the audio stream is set and saved; when a calling VoIP phone begins to call, effector chain processing is completed in an audio stream by relying on a Mediastremeer2 framework, then the processed audio stream is transmitted to a module S for audio processing, a sampling rate is adapted to process a voice signal, and dynamic control of voice volume is accurately realized. According to the invention, the adjustability of dynamic adjustment is higher, and the adjustment is more refined.

Description

Technical Field

[0001] The invention belongs to the field of VoIP (standing for "Voice over IP" or "Voice over Internet Protocol"), and in particular is a method for dynamically controlling the voice volume of a VOIP voice stream. Background Art

[0002] In modern communication technology, voice transmission has become an indispensable part of daily life; VOIP (Voice over Internet Protocol) technology transmits voice signals through the Internet protocol, providing users with real-time and convenient voice communication services.

[0003] However, VOIP voice streams are often affected by multiple factors during transmission, such as network bandwidth, delay or packet loss, which leads to reduced voice quality and poor user experience. Among them, the instability of voice volume is a particularly prominent problem.

[0004] The instability of voice volume mainly comes from the change of signal strength. The loudness of users speaking in different emotions is inconsistent. In some scenarios, the upper and lower limits of users' speaking loudness are very different, resulting in clipping distortion of loud volume when low volume human voice can be heard, or low volume cannot be heard clearly when high volume human voice can be heard. In order to solve these problems, it is particularly important to develop a dynamic control module that can autonomously control the algorithm logic.

[0005] The module that dynamically controls the volume of VOIP voice streams is based on the processing of digital signals.

[0006] Digital signal processing is a technology that extracts information, improves signal quality or implements specific functions by performing mathematical operations and transformations on digital signals. The self-developed DRC (Dynamic Range Compression) algorithm is a digital signal processing technology designed specifically for VOIP voice streams. Compared with traditional fixed gain control or noise suppression algorithms, the DRC algorithm has higher algorithm flexibility and adaptability. It can adjust the logical relationship of each algorithm more concisely and dynamically according to the real-time characteristics of the voice signal, thereby accurately realizing dynamic control of the voice volume. Summary of the invention

[0007] In view of the problems existing in the prior art, the present invention provides a method for dynamic voice volume control of VOIP voice stream, which integrates the four major voice dynamic algorithms of compressor, expander, noise gate and limiter. According to the actual scene requirements, users can independently adjust the order of compressor, expander, noise gate, limiter and the parameters of each algorithm logically; at the same time, a module for optimizing voice coherence is added to optimize the connectivity of real-time voice. This method simplifies the debugging workload of voice debuggers in the VoIP field and improves work efficiency; it is also easy to use, and the algorithm logic can be adjusted by the configuration file, so simple debugging work can be achieved without the reserve of professional knowledge.

[0008] The method for dynamically controlling the voice volume of a VOIP voice stream comprises the following specific steps:

[0009] Step 1: Utilize the existing DRC algorithm to optimize the design of the voice volume dynamic control module S;

[0010] The voice volume dynamic control module S includes a compressor, a limiter, an expander and a noise gate.

[0011] Among them, the controllable parameters of the compressor include: threshold, compression ratio, attack time and release time; the modes are divided into: single effector mode, dual effector series mode, dual effector parallel mode, triple effector series mode and triple effector series-parallel mixed mode.

[0012] The dual-effects series mode is to process two compressors or a compressor and an expander in sequence, which has relatively stable audio dynamics; the triple-effects series mode is to process two compressors and an expander in sequence, and then mix the output results by setting the ratio, which can provide targeted control for excessive or insufficient signals and has a stronger dynamic control effect; the three-effects series-parallel mixed mode is to mix the output results of the dual effects by setting the ratio, and then connect an effector in series, which can stably control the audio dynamics.

[0013] The controllable parameters of the expander are: threshold and ratio.

[0014] The expander can be placed after the compressor, before the compressor, or in parallel with the compressor.

[0015] Both the limiter and the gate have only one parameter: threshold. The threshold setting of the gate controls the threshold of small signal muting, which is processed first in the workflow; the threshold setting of the limiter controls the maximum level of the output, which is processed last.

[0016] Step 2: For the calling VoIP phone, add the optimized module S to the audio effect chain of the phone, and add the parameters contained in the module S to the audio configuration file;

[0017] Parameter configuration includes: compressor switch, expander switch, noise gate switch and limiter switch;

[0018] Step 3: According to the scenario requirements, the user selects the mode in the parameter list in advance, adds it, sets the processing order of the audio stream, and saves it;

[0019] Specific modes include:

[0020] Mode 1 dual effector series mode, refers to the dual series structure of dual compressors or the dual series structure of compressors + expanders. The dual compressor is usually a combination of a high-threshold and high-compression ratio compressor + a low-threshold and low-compression ratio compressor, or a normal combination of a compressor and an expander.

[0021] Mode 2 three-effects series mode refers to the three-effects series structure of dual compressors + expanders, usually a combination of high-threshold high-compression ratio compressor + low-threshold low-compression ratio compressor + expander;

[0022] Mode 3 three-effector series-parallel mixed mode refers to the series + parallel structure of the effects, usually a combination of a parallel circuit group of effects a and b connected in series after effect c;

[0023] Mode 4 is a mixed mode of three effects in parallel and series, which refers to the parallel + series structure of the effects; usually it is a combination of the parallel circuit group of effects a and b, and then the series effect c.

[0024] Step 4: The calling VoIP phone starts the call. After the audio stream is processed by the effector chain based on the Mediastreamer2 framework, it is passed to module S for audio processing. The sampling rate is adapted to process the voice signal, and the dynamic control of the voice volume is accurately realized.

[0025] The specific process is:

[0026] The voice signal of the VOIP call is read through the virtual sound card. First, the voice signal is processed by the noise gate, and the RMS loudness value of each frame of input data is analyzed. The loudness value is compared with the set threshold. If it is lower than the set threshold, the audio data is cleared to make it silent.

[0027] Then, the user switches the processing order and linking method of the compressor and expander for the selected mode according to the actual usage scenario and usage requirements;

[0028] A new RMS loudness value is recalculated using each frame of the input signal. If the selected mode includes a compressor, it is determined whether the RMS value is higher than the threshold. If so, the compressor is executed to compress the loudness value exceeding the threshold. The compression amount depends on two parameters: the difference between the current frame loudness value and the threshold and the compression ratio. If the selected mode includes an expander, the expander is executed to expand the loudness value below the threshold. The expansion amount depends on two parameters: the current frame loudness value and the expansion ratio.

[0029] Then, the audio loudness is improved through the volume compensation module;

[0030] Finally, add a limiter to recalculate the new RMS loudness value after gain. By setting the maximum RMS loudness, the maximum loudness value of each frame of data is limited to avoid output audio distortion.

[0031] The advantages of the present invention are:

[0032] 1) The present invention provides a method for dynamically controlling the voice volume of a VOIP voice stream, and proposes an improved DRC algorithm with a more complex structure and greater operability, which has higher adjustability and more refined adjustment for dynamic adjustment.

[0033] 2) The present invention provides a method for dynamically controlling the voice volume of a VOIP voice stream, which integrates conventional processing modes according to common operations, does not require manual operation of the effector combination, and directly uses mode switching to obtain a better algorithm effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flow chart of a method for dynamically controlling the voice volume of a VOIP voice stream according to the present invention;

[0035] Figure 2 The present invention accurately realizes the dynamic control flow chart of the voice volume. DETAILED DESCRIPTION

[0036] In order to facilitate those skilled in the art to understand and implement the present invention, the present invention is described in further detail and in depth with reference to the accompanying drawings.

[0037] The method for dynamically controlling the voice volume of the VOIP voice stream is as follows: Figure 1 As shown, the specific steps are as follows:

[0038] Step 1: Utilize the existing DRC algorithm to optimize the design of the voice volume dynamic control module S;

[0039] The voice volume dynamic control module S includes a compressor, a limiter, an expander and a noise gate.

[0040] The main function of the compressor is to attenuate loud volumes that exceed a given threshold, and is often used in recording systems to protect hardware and increase overall loudness; the main function of the limiter is to achieve infinite compression above the set threshold; the main function of the expander is to attenuate volumes below the set threshold; the noise gate is an extreme expander, and its main function is to mute volumes below the set threshold. Since the four small modules have different functions, the compressor is used to reduce dynamics, the expander is used to amplify dynamics, and the limiter and noise gate are two special processes for protecting volume, so the appropriate processing mode plays a crucial role in the processing results.

[0041] In real-time VOIP communication, human voice is usually the main communication content. Under normal circumstances, the dynamic control of human voice usually requires a small dynamic. Therefore, the compressor plays a very important role in the entire voice volume dynamic control module. The workflow of the voice volume dynamic control module of the entire VOIP voice stream is as follows: the noise gate suppresses small signal noise, the compressor and expander control the dynamic, and the limiter controls the output level.

[0042] The controllable parameters of the compressor include threshold, compression ratio, attack time and release time; the controllable parameters of the expander include threshold and ratio. Since the expander is not very useful in actual speech signals, the attack time and release time are set to a fixed 10ms to make the signal transition as smooth as possible.

[0043] Therefore, the following modes are designed in the compressor and expander parts: single effector mode, dual effector series mode, dual effector parallel mode, triple effector series mode and triple effector series-parallel mixed mode.

[0044] When the RMS value calculated for every 10ms of audio data exceeds the threshold corresponding to the compressor, the entire compressor is started, and the compression ratio determines the amount of compression of the exceeding part; the attack time and release time affect the transition effect of the compressor starting and ending compression; when the RMS value calculated for every 10ms of audio data is lower than the threshold corresponding to the compressor, the entire expander is started, and the expansion ratio is the scaling amplitude of the actual signal.

[0045] Single Effects mode is a single compressor or expander processing, performing a regular compression or expansion.

[0046] The dual-effects series mode processes two compressors or one compressor and one expander in sequence. For dual compressors, you can first select a compressor to suppress excessive signals, and then select a compressor to control the overall dynamics; you can also first select a compressor to suppress the overall dynamics, and then select a compressor to control excessive signals. This processing method is more extreme for audio signal processing, with a large degree of compression, and can effectively control the sound that fluctuates. For a single compressor and single expander, you can first select a compressor to suppress excessive signals, and then select an expander to scale excessive signals. This processing method is more common and is a universal compressor and expander hybrid method.

[0047] The three-effects series mode processes two compressors and one expander in sequence, and then mixes the output results by the set ratio. This can complete the overall control dynamics, suppress excessive signals to a certain extent, and scale excessive signals to a certain extent. In actual speech signal processing, it is necessary to select a mode according to various situations such as actual scenes and device configurations to achieve the most perfect effect.

[0048] The three-effects series-parallel hybrid mode is a mixture of the first two modes, that is, a compressor or expander is connected in series under the structure of the original mode two, which can achieve complex, high-precision, and high-fidelity sound dynamic effects.

[0049] The threshold is used to determine whether the compressor or expander is activated, and the ratio is used to control the scaling of signals below the threshold. Unlike traditional expanders, the expander in the voice volume dynamic control module does not completely expand. It can not only reduce the signal below the threshold to suppress small signals, but also amplify the signal below the threshold to enhance unstable small signals. This is also due to the AI ​​noise reduction algorithm in the front of the voice volume dynamic control module, which makes the data signal-to-noise ratio entering the DRC module higher, and there is no need to worry about the problem of amplifying small signal noise.

[0050] Both the limiter and the gate have only one parameter: threshold. The threshold setting of the gate controls the threshold of small signal muting, which is processed first in the workflow; the threshold setting of the limiter controls the maximum level of the output, which is processed last.

[0051] Step 2: For the calling VoIP phone, add the optimized module S to the audio effect chain of the phone, and add the parameters contained in the module S to the audio configuration file;

[0052] Parameter configuration includes: compressor switch, expander switch, noise gate switch and limiter switch;

[0053] Step 3: According to the scenario requirements, the user adds the mode to the parameter list in advance, sets the processing order of the audio stream and saves it;

[0054] Specific modes include:

[0055] Mode 1 dual effector series mode, refers to the dual series structure of dual compressors or the dual series structure of compressors + expanders. The dual compressor is usually a combination of a high-threshold and high-compression ratio compressor + a low-threshold and low-compression ratio compressor, or a normal combination of a compressor and an expander.

[0056] Mode 2 three-effects series mode refers to the three-effects series structure of dual compressors + expanders, usually a combination of high-threshold high-compression ratio compressor + low-threshold low-compression ratio compressor + expander;

[0057] Mode 3 three-effector series-parallel mixed mode refers to the series + parallel structure of the effects, usually a combination of a parallel circuit group of effects a and b connected in series after effect c;

[0058] Mode 4 is a mixed mode of three effects in parallel and series, which refers to the parallel + series structure of the effects; usually it is a combination of the parallel circuit group of effects a and b, and then the series effect c.

[0059] Step 4: The calling VoIP phone starts the call. After the audio stream is processed by the effector chain based on the Mediastreamer2 framework, it is passed to module S for audio processing. The sampling rate is adapted to process the voice signal, and the dynamic control of the voice volume is accurately realized.

[0060] like Figure 2 As shown, the specific process is:

[0061] Step 401, read the voice signal of the VOIP call through the virtual sound card end, first, determine whether the noise gate switch is turned on, if so, the voice signal enters the noise gate module, analyzes the RMS loudness value of each frame of input data, compares the loudness value with the set threshold, clears the audio data below the set threshold, and mutes it, and then enters step 402; otherwise, directly enters step 402;

[0062] The function of the noise gate is to mute very small signals in a signal with a large signal-to-noise ratio, achieving a certain degree of denoising (noise reduction) effect.

[0063] The calculation steps of RMS loudness value are as follows: normalization, square accumulation (summing up the squares of all sample values, that is, the audio data on the computer), calculating the mean of the sum of squares, and taking the square root to get the RMS value; finally, decibel conversion; RMS is finally converted to db value form. For the convenience of parameter setting, the units of basic parameters are all in db.

[0064] Step 402: determine whether the compressor expander hybrid switch is turned on. If so, the user adjusts the mode and the position and parameters of the effector according to the actual usage scenario and usage requirements, performs the compressor and expander processing according to the selected mode, and enters step 403; otherwise, enters step 404;

[0065] The function of a compressor is to reduce fluctuations (control dynamics), while the function of an expander is to increase fluctuations (increase dynamics). The combination of these two functions achieves the best dynamic control effect.

[0066] The specific processing is:

[0067] Using each frame of the input signal, a new RMS loudness value is recalculated to determine whether the value is higher than the threshold. If so, the compressor is executed to compress the loudness value that exceeds the threshold. The compression amount depends on two parameters: the difference between the current frame loudness value and the threshold and the compression ratio.

[0068] The principle of the compressor is: input an audio, and set a threshold value such as -10db. Calculate the RMS value of 10ms data once for each frame (such as 10ms), and convert the RMS value into a db value; if the current db value exceeds the db value of the threshold, then compress. Otherwise, do not compress. Compression amount = (current db-threshold)*(1-1 / compression ratio).

[0069] Otherwise, the expander process is performed to expand the loudness value of the part below the threshold, and the expansion amount depends on two parameters: the current frame loudness value and the expansion ratio.

[0070] The principle of the expander is the same as that of the compressor, but in reverse. There is no compression ratio. There is only a threshold. If the current db value exceeds the threshold db, no expansion is performed. Otherwise, expansion is performed; the amount of expansion = current audio data value * expander ratio.

[0071] Switches the order and chaining of compressors and expanders for the selected mode;

[0072] The details are as follows:

[0073] Mode 1 dual effector series mode, refers to the dual series structure of dual compressors or the dual series structure of compressors + expanders. The dual compressor is usually a combination of a high-threshold and high-compression ratio compressor + a low-threshold and low-compression ratio compressor, or a normal combination of a compressor and an expander.

[0074] Mode 2 three-effects series mode refers to the three-effects series structure of dual compressors + expanders, usually a combination of high-threshold high-compression ratio compressor + low-threshold low-compression ratio compressor + expander;

[0075] Mode 3 three-effector series-parallel mixed mode refers to the series + parallel structure of the effects, usually a combination of a parallel circuit group of effects a and b connected in series after effect c;

[0076] Mode 4 is a mixed mode of three effects in parallel and series, which refers to the parallel + series structure of the effects; usually it is a combination of the parallel circuit group of effects a and b, and then the series effect c.

[0077] Step 403, the audio loudness is increased through a volume compensation module;

[0078] Step 404, determine whether the limiter switch is turned on, if so, use the limiter module to recalculate the new RMS loudness value after gain, limit the maximum loudness value of each frame of data by the set maximum RMS loudness, and output the signal to avoid the problem of output audio distortion. Otherwise, directly output the signal.

[0079] The volume in audio ranges from -∞ to 0, with 0db being the maximum loudness. If the audio loudness exceeds 0, the sound will be abnormal. The limiter will control the loudness of all signals to the set parameters. For example, if -0.5db is set, all signals exceeding -0.5db will be directly limited to -0.5db, which can prevent the audio from exceeding 0db and causing distortion.

[0080] The present invention integrates four speech dynamic algorithms of compressor, limiter, expander and noise gate, regenerates speech volume dynamic control module S, and can link any several algorithms in complex speech scenes to maximize the effect of speech volume dynamic control module S. Users can adjust the logic mechanism independently, and adjust the logical order of compressor, expander, noise gate and limiter according to the actual scene requirements. Not only can they choose the traditional series processing method, that is, each effector is processed separately and in order; they can also choose the parallel processing method, that is, two effectors are processed at the same time and the output data is mixed in a set ratio. In addition, the parameters in the compressor, expander, noise gate and limiter can be adjusted independently, and a module S for optimizing speech coherence is added at the same time, so that the connectivity of real-time speech is optimized. This module S simplifies the debugging workload of voice debugging personnel in the VoIP field and improves work efficiency; the module is easy to use, and the algorithm logic can be adjusted by the configuration file, and simple debugging work can be achieved without the reserve of professional knowledge.

[0081] Example:

[0082] Step 1: For the VoIP phone to be used for a call, add the voice volume dynamic control module S to the audio effect chain, and add the parameters contained in the module to the audio configuration file corresponding to the device;

[0083] Parameter configuration includes: compressor switch, expander switch, noise gate switch, and limiter switch. The compressor includes compressor threshold, compression ratio, attack time, and release time; the expander includes threshold and ratio; the noise gate and limiter include threshold, and the parameters can be modified through the parameter configuration file provided by the phone.

[0084] Step 2: If the user wants to modify the order of audio stream processing, he can add a mode in the parameter list; before executing the voice volume dynamic control module S, he must ensure that it is set according to his actual scenario requirements.

[0085] Mode 1 refers to a dual series structure, mode 2 refers to a triple series structure, mode 3 refers to a series + parallel structure, and mode 4 refers to a parallel + series structure.

[0086] Step 3: Start a call and perform audio processing in the audio stream through the voice volume dynamic control module;

[0087] Specific processing includes:

[0088] 1) VOIP calls are processed by the effector chain based on the Mediastreamer2 framework. The voice signal is read through the virtual sound card, and after a series of effector processing, it is transmitted to the voice volume dynamic control module S, and the voice signal processing begins after the sampling rate is adapted;

[0089] Mediastreamer2 is an open source framework for real-time streaming media processing, which is highly flexible and extensible. It provides a series of audio and video processing functions, including encoding and decoding, filtering, capturing and playing. It supports self-expansion based on the original effector module, and external parameters are imported into the module by calling Method to configure the module parameters.

[0090] 2) First, perform noise gate processing on the voice signal, analyze the RMS loudness value of each frame of input data, and compare the loudness value with the set threshold. If it is lower than the set threshold, the audio data is cleared to make it silent;

[0091] 3) Perform compressor and expander processing according to the selected mode. This part requires the user to adjust the mode and the position and parameters of the effector according to the actual usage scenario and usage requirements.

[0092] If the compressor is executed, a new RMS loudness value is recalculated for each frame of signal input. If the value is higher than the threshold, the compressor processing is started to compress the loudness value that exceeds the threshold.

[0093] The compressor process is divided into three states: start state, execution state and release state.

[0094] The start state is that the RMS value of the previous frame data analysis does not meet the conditions, but the RMS value of the current frame data analysis meets the conditions, then the audio transition processing will be completed at this frame rate using any function such as a linear function, exponential function, power function, etc.

[0095] The execution status means that the RMS value of the previous frame data analysis meets the conditions, the RMS value of the current frame data analysis still meets the conditions, and the compressor performs the compression task normally.

[0096] The release state means that the RMS value of the previous frame data analysis meets the conditions, but the RMS value of the current frame data analysis does not meet the conditions, then the audio transition processing will be completed at this frame rate using any of the above functions.

[0097] If the compressor is executed, the new RMS loudness value is recalculated for each frame signal input. If the value is lower than the threshold, the expander processing is started to expand the loudness value below the threshold. The expansion amount depends on two parameters: the current frame loudness value and the expansion ratio.

[0098] Expander processing is also divided into three states: start state, execution state and release state, and audio transition processing is the same as the compressor.

[0099] Switches the order and chaining of the compressor and expander processing for the selected mode, as follows:

[0100] Mode 1 dual series structure can usually set up dual compressors or single compressor + single expander;

[0101] The mode2 three-series structure is generally a combination of a high-threshold, high-compression ratio compressor + a low-threshold, low-compression ratio compressor + an expander. After controlling the dynamics, the level of the small signal can still be slightly adjusted to maintain a stable state.

[0102] Mode 3 is a relatively complex mode. If you need to amplify a small signal first, you can select the expander as the first effector, and then process it separately through a high-threshold and high-compression ratio compressor and a low-threshold and low-compression ratio compressor. One is used to suppress sudden large voice signals, and the other is used to control the overall dynamics. Through proportional mixing, it can prevent the voice dynamics from being controlled too small after passing through the double compressors.

[0103] The principles of mode4 and mode3 are basically the same. They can expand small signals and suppress large signals. After mixing in a set ratio, the overall dynamics are controlled by the compressor.

[0104] Among them, mode3 and mode4 are not limited to the above combinations. The combination of compressor and expander can achieve the best effect according to the actual scenario.

[0105] 4) After executing the above modules, the audio loudness is improved through the volume compensation module.

[0106] 5) At the end of the entire module, add a limiter to recalculate the new RMS loudness value after gain, and limit the maximum loudness value of each frame of data by setting the maximum RMS loudness to avoid output audio distortion.

Claims

1. A method for dynamically controlling the voice volume of a VOIP voice stream, characterized in that: The specific steps are as follows: Step 1: Utilize the existing DRC algorithm to optimize the design of the voice volume dynamic control module S; The voice volume dynamic control module S includes a compressor, a limiter, an expander and a noise gate; Step 2: For the calling VoIP phone, add the optimized module S to the audio effect chain of the phone, and add the parameters contained in the module S to the audio configuration file; Step 3: According to the scenario requirements, the user selects the mode in the parameter list in advance, adds it, sets the processing order of the audio stream, and saves it; Step 4: The calling VoIP phone starts the call. After the audio stream is processed by the effector chain based on the Mediastreamer2 framework, it is passed to module S for audio processing. The sampling rate is adapted to process the voice signal, and the dynamic control of the voice volume is accurately realized. The specific process is: The voice signal of the VOIP call is read through the virtual sound card. First, the voice signal is processed by the noise gate, and the RMS loudness value of each frame of input data is analyzed. The loudness value is compared with the set threshold. If it is lower than the set threshold, the audio data is cleared to make it silent. Then, the user switches the processing order and linking method of the compressor and expander by selecting the mode according to the actual usage scenario and usage requirements; The modes include single effector mode, dual effector series mode, dual effector parallel mode, and triple effector series-parallel mixed mode; Effects refers to compressors or expanders; Recalculate a new RMS loudness value using each frame of the input signal, and determine whether the value is higher than a threshold value. If yes, perform a compressor process to compress the loudness value that exceeds the threshold value; otherwise, perform an expander process to expand the loudness value that is lower than the threshold value. Then, the audio loudness is improved through the volume compensation module; Finally, add a limiter to recalculate the new RMS loudness value after gain, limit the maximum loudness value of each frame of data by the set maximum RMS loudness, and output the final signal.

2. A method for dynamically controlling the voice volume of a VOIP voice stream as claimed in claim 1, characterized in that: The controllable parameters of the compressor include: threshold, compression ratio, attack time and release time; the controllable parameters of the expander are: threshold and ratio; the context of the expander and the compressor depends on the mode selection; Both the limiter and the gate have only one parameter: the threshold; The setting of the gate threshold controls the threshold for small signal muting and is processed first in the workflow; the setting of the limiter threshold controls the maximum level of the output and is processed last.

3. A method for dynamically controlling the voice volume of a VOIP voice stream as claimed in claim 1, characterized in that: The dual-effector series mode is that two compressors or a compressor and an expander are processed in sequence, which has relatively stable audio dynamics; the three-effector series mode is that two compressors and an expander are processed in sequence, and then the output results are mixed by setting the ratio, which can have targeted control over excessive or insufficient signals and has a stronger dynamic control effect; the three-effector series-parallel mixed mode is that the dual effectors mix the output results by setting the ratio, and then connect an effector in series, which can stably control the audio dynamics.

4. A method for dynamically controlling the voice volume of a VOIP voice stream as claimed in claim 1, characterized in that: In the step 2, the parameter configuration includes: a compressor switch, an expander switch, a noise gate switch and a limiter switch.

5. The method for dynamically controlling the voice volume of a VOIP voice stream as claimed in claim 1, characterized in that: In the step 3, the specific modes include: Mode 1 dual effector series mode refers to the dual series structure of dual compressors or the dual series structure of compressor + expander. The dual compressor is usually a combination of a high-threshold and high-compression ratio compressor + a low-threshold and low-compression ratio compressor, or a normal combination of a compressor and an expander. Mode 2 three-effect series mode refers to the three-effect series structure of dual compressor + expander, usually a combination of high-threshold high-compression ratio compressor + low-threshold low-compression ratio compressor + expander; Mode 3 three-effector series-parallel mixed mode refers to the series + parallel structure of the effects, usually a combination of a parallel circuit group of effects a and b connected in series after effect c; Mode 4 three-effects parallel-series mixed mode refers to the parallel + series structure of the effects; usually it is a parallel circuit group of effects a and b, followed by a series effect c.

6. A method for dynamically controlling the voice volume of a VOIP voice stream as claimed in claim 1, characterized in that: In step 4, the compression amount depends on two parameters: the difference between the current frame loudness value and the threshold value and the compression ratio.

7. A method for dynamically controlling the voice volume of a VOIP voice stream as claimed in claim 1, characterized in that: In step 4, the expansion amount depends on two parameters: the current frame loudness value and the expansion ratio.

Images