A method and system for judging the degree of environmental reverberation based on speech signals
Through multi-dimensional acquisition and in-depth analysis of the energy attenuation and propagation characteristics of speech signals, the reverberation time is calculated, which solves the problems of low judgment accuracy and large results deviation caused by constant environmental data assumptions in traditional methods, and achieves a more accurate reverberation degree Judgment and excellent dereverberation effect.
Patent Information
- Application Number
- CN202410704261.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-31
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-05-31
AI Technical Summary
The traditional method of judging the degree of environmental reverberation based on speech signals assumes that the environmental data is constant, resulting in low judgment accuracy and large deviations in the result, which affects the dereverberation effect.
Using multi-dimensional acquisition module, reverb analysis module and evaluation and control module, by recording voice signals of different frequencies, collecting environmental constants and variable data, calculating waveform characteristics, spectrum characteristics and reverb time, deeply analyzing energy attenuation and propagation characteristics, and consolidating the To deal with environmental factors, determine the degree of reverberation.
It improves the accuracy of the judgment degree of reverberation, reduces the deviation of the judgment results, and improves the subsequent dereverberation effect.
Smart Images

Figure CN118609602B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of speech processing, and in particular to a method and system for determining the degree of environmental reverberation based on speech signals. Background Art
[0002] Environmental reverberation based on speech signals refers to the phenomenon that in a closed or semi-closed space, after the sound is emitted from the sound source, it is reflected multiple times on surfaces such as walls, ceilings and the ground, and finally reaches the receiver. The phenomenon that the voice signal is transmitted through multiple paths and causes the sound energy to be prolonged is called reverberation. Reverberation has a significant impact on the quality of voice communication, especially in the fields of teleconferencing, speech recognition, hearing aids and underwater acoustics. In extreme cases, environmental reverberation may cause distortion of the voice signal. The reduction in clarity increases the possibility of signal misunderstanding, making it difficult for the automatic speech recognition system to distinguish the voice signal, seriously affecting the recognition accuracy. The degree of reverberation is usually measured by the reverberation time (RT60), which represents the time required for the sound energy to decay by 60 decibels after a sound source stops making sound in the room. The longer the RT60 time value, the higher the reverberation degree, and the shorter the time value, the lighter the reverberation degree. The method of judging the degree of environmental reverberation based on speech signals involves multiple fields such as acoustics, signal processing and machine learning. Determining the degree of reverberation and solving the problem of speech signal distortion caused by reverberation requires the use of a combination of technologies and methods to ensure that voice communication and speech recognition systems can maintain good performance in various environments. As technology continues to develop, more efficient and intelligent solutions are expected to emerge in the future to further improve speech signal processing affected by reverberation.
[0003] At present, the traditional method of judging the degree of environmental reverberation based on speech signals assumes that the environmental data is constant. However, in the actual environment, the room impulse response may change due to factors such as human movement and the opening and closing of doors and windows, which will affect the accuracy of the judgment of the degree of environmental reverberation. The judgment result has a large deviation, which affects the subsequent speech dereverberation effect. Summary of the invention
[0004] 1. Technical issues to be solved
[0005] In view of the shortcomings of the prior art, the present invention provides a method and system for judging the degree of environmental reverberation based on speech signals, which have the advantages of more accurate judgment of the degree of reverberation due to normalized processing of environmental factors, small deviation in judgment results and good subsequent dereverberation effect. It solves the problem that the traditional method for judging the degree of environmental reverberation based on speech signals assumes that the environmental data is constant, the judgment accuracy is not high, the judgment result deviation is large and the subsequent dereverberation effect is poor.
[0006] (II) Technical solution
[0007] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a method for judging the degree of environmental reverberation based on a speech signal, wherein the method for judging the degree of environmental reverberation based on a speech signal is judged by a system for judging the degree of environmental reverberation based on a speech signal, wherein the system for judging the degree of environmental reverberation based on a speech signal comprises a multi-dimensional acquisition module, a reverberation analysis module and an evaluation and control module, and the method for judging the degree of environmental reverberation based on a speech signal comprises the following steps:
[0008] Step 1: Record multiple voice signals of different frequencies, and each voice signal has the same duration;
[0009] Step 2: In the test environment, the signal transmitting device is controlled to output a voice signal, and the signal receiving device is used to perform sampling, and the multi-dimensional acquisition module collects data sets according to different data sources;
[0010] Step 3: Comprehensively analyze the energy attenuation trend and propagation characteristics of the speech signal in the ambient space, and the reverberation analysis module calculates and generates waveform features Bxtz, spectrum features Putz and reverberation time Hxsj accordingly;
[0011] Step 4: The evaluation and control module determines the reverberation degree of the speech signal in the test environment according to the reverberation time Hxsj, and generates a corresponding control report.
[0012] Preferably, the types of recorded voice signals include low-frequency voice signals, medium-frequency voice signals and high-frequency voice signals, the frequency range of the low-frequency voice signals is set to 20Hz to 300Hz, the frequency range of the medium-frequency voice signals is set to 300Hz to 3400Hz, and the frequency range of the high-frequency voice signals is set to 3400Hz to 20000Hz.
[0013] Preferably, a system for determining the degree of environmental reverberation based on speech signals comprises a multi-dimensional acquisition module, a reverberation analysis module and an evaluation and control module;
[0014] The multi-dimensional acquisition module includes an environmental constant unit, an environmental variable unit, a constant signal unit and a variable signal unit. The environmental constant unit collects an environmental constant data set through a network connection to an environmental scanning device, and transmits it to the reverberation analysis module after numbering. The environmental variable unit collects an environmental variable data set through a network connection to an environmental monitoring device, and transmits it to the reverberation analysis module after numbering. The constant signal unit collects a constant signal data set through a network connection to a signal transmitting device, and transmits it to the reverberation analysis module after numbering. The variable signal unit collects a variable signal data set through a network connection to a signal receiving device, and transmits it to the reverberation analysis module after numbering. The multi-dimensional acquisition module is connected to the reverberation analysis module through a network.
[0015] The reverberation analysis module includes a time domain analysis unit, a spectrum analysis unit and a duration analysis unit. The reverberation analysis module collects a reference data set through a network connection to a database, and caches the data set to the reverberation analysis module after numbering. The time domain analysis unit calculates and generates a waveform feature Bxtz according to a constant signal data set, a variable signal data set and a reference data set. The spectrum analysis unit calculates and generates a spectrum feature Putz according to a constant signal data set, a variable signal data set and a reference data set. The duration analysis unit calculates and generates a reverberation time Hxsj according to an environmental constant data set and an environmental variable data set. The reverberation analysis module is connected to an evaluation and control module through a network connection.
[0016] The evaluation and control module determines the reverberation degree of the speech signal in the test environment according to the reverberation time Hxsj, and generates a control report accordingly.
[0017] Preferably, the environmental constant unit numbers the constant data of the total indoor volume, the total indoor surface area and the average sound absorption coefficient in the test environment space according to the characteristics of the environmental constant data set, and the environmental constant data set is numbered HJ tj , HJ mj and HJ xs The environmental variable unit numbers the thermodynamic temperature and pressure data in the test environment space according to the characteristics of the environmental variable data set, and the environmental variable data set is numbered HJ wd and HJ qy .
[0018] Preferably, the constant signal unit numbers the initial speech signal data output by the signal transmitting device according to the characteristics of the constant signal data set, and the constant signal data set is numbered as follows: a represents the frequency range data of the initial speech signal, and the variable signal unit numbers the final speech signal data received by the signal receiving device according to the characteristics of the variable signal data set, and the variable signal data set number is b represents the frequency range data of the final speech signal.
[0019] Preferably, the reference data set is composed of a reference time domain waveform algorithm model, a reference spectrum analysis algorithm model and a reference reverberation time standard data set, and the reverberation analysis module numbers the reference data set according to its characteristics, and the reference data set is numbered CK sy , CK pu and CK hx .
[0020] Preferably, the time domain analysis unit calculates and generates the waveform feature Bxtz according to the constant signal data set, the variable data set and the reference data set, and the calculation formula is as follows:
[0021]
[0022] In the formula, Bxtz represents the waveform characteristics, It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the constant signal data set are substituted into the reference time domain waveform algorithm model to calculate the amplitude intensity feature data. It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the variable signal data set are substituted into the reference time domain waveform algorithm model to calculate the amplitude intensity feature data.
[0023] Preferably, the spectrum analysis unit calculates and generates a spectrum feature Putz according to the constant signal data set, the variable data set and the reference data set, and the calculation formula thereof is as follows:
[0024]
[0025] In the formula, Putz represents the spectrum characteristics, It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the constant signal data set are substituted into the reference spectrum analysis algorithm model to calculate the frequency distribution characteristic data. It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the variable signal data set are substituted into the reference spectrum analysis algorithm model to calculate the frequency distribution characteristic data.
[0026] Preferably, the duration analysis unit calculates and generates the reverberation time Hxsj according to the environmental constant data set and the environmental variable data set, and the calculation formula is as follows:
[0027]
[0028] In the formula, Hxsj represents the reverberation time, r represents the ideal gas constant in the air density calculation formula, It represents the air density of the current test environment obtained by substituting the thermodynamic temperature and pressure values in the environmental variable data set into the air density calculation formula. k represents the adiabatic index of air, and R represents the ideal gas constant in the sound speed calculation formula. It represents the sound speed of the current test environment obtained by substituting the thermodynamic temperature value in the environmental variable data set into the sound speed calculation formula. μ represents the conversion coefficient for air density and sound speed in the reverberation time calculation formula. It represents the ideal reverberation time calculated based on the total indoor volume, total indoor surface area and average sound absorption coefficient in the environmental constant data set under ideal temperature and sound velocity. It means that the precise reverberation time is calculated by converting the temperature and sound speed variables in the actual detection environment and combining them with the reverberation time under ideal conditions.
[0029] Preferably, the evaluation and control module compares the reverberation time Hxsj with the reference reverberation time standard data set CK pu , judge the reverberation degree of the speech signal in the test environment, the reverberation time Hxsj is lower than the reference reverberation time standard data set CK pu When the reverberation degree of the test environment is judged to be slight, no control report is generated, and the reverberation time Hxsj is greater than or equal to the reference reverberation time standard data set CK pu When the reverberation degree of the test environment is judged to be serious, a control report is generated, which includes an environmental constant data set, an environmental variable data set, a constant signal data set, a variable signal data set, a reference data set, a waveform feature Bxtz, a spectrum feature Putz and a reverberation time Hxsj.
[0030] Compared with the prior art, the present invention provides a method and system for determining the degree of environmental reverberation based on speech signals, which has the following beneficial effects:
[0031] 1. The present invention records multiple segments of voice signals with the same duration and different frequencies, controls a signal transmitting device to output voice signals of different frequency stages in a test environment, and performs sampling through a signal receiving device. A multidimensional acquisition module sets an environmental constant unit, an environmental variable unit, a constant signal unit, and a variable signal unit, classifies and collects environmental constant data sets, environmental variable data sets, constant signal data sets, and variable signal data sets according to different data sources. A reverberation analysis module sets a time domain analysis unit, a spectrum analysis unit, and a duration analysis unit, comprehensively analyzes the energy attenuation trend and propagation characteristics of the voice signal in the environmental space, the time domain analysis unit calculates and generates a waveform feature Bxtz, compares and analyzes the amplitude intensity features of the initial voice signal and the final voice signal, the spectrum analysis unit calculates and generates a spectrum feature Putz, compares and analyzes the frequency distribution features of the initial voice signal and the final voice signal, the duration analysis unit calculates and generates a reverberation time Hxsj, deeply analyzes the correlation between the constant data and the variable data, and the normalized processing of the environmental factors makes the reverberation degree judgment more accurate.
[0032] 2. The present invention compares the reverberation time Hxsj with the reference reverberation time standard data set CK through the evaluation control module pu , judge the reverberation degree of the speech signal in the test environment, and the reverberation time Hxsj is lower than the reference reverberation time standard data set CK pu When the reverberation degree of the test environment is judged to be slight, no control report is generated, and the reverberation time Hxsj is greater than or equal to the reference reverberation time standard data set CK puWhen the reverberation degree of the test environment is judged to be serious, a control report is generated. The control report includes the environmental constant data set, environmental variable data set, constant signal data set, variable signal data set, reference data set, waveform feature Bxtz, spectrum feature Putz and reverberation time Hxsj. The smaller the deviation of the judgment result, the better the subsequent dereverberation effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 It is a step diagram of the method of the present invention;
[0034] Figure 2 Schematic diagram of the system of the present invention. DETAILED DESCRIPTION
[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0036] See also Figure 1-2 A method for judging the degree of environmental reverberation based on a speech signal is provided. The method for judging the degree of environmental reverberation based on a speech signal is judged by a system for judging the degree of environmental reverberation based on a speech signal. The system for judging the degree of environmental reverberation based on a speech signal comprises a multi-dimensional acquisition module, a reverberation analysis module and an evaluation and control module. The method for judging the degree of environmental reverberation based on a speech signal comprises the following steps:
[0037] Step 1, recording multiple voice signals of different frequencies, and each voice signal has the same duration. The recorded voice signals include low-frequency voice signals, medium-frequency voice signals and high-frequency voice signals. The voice frequency range that can be heard by human ears is approximately between 20Hz and 20000Hz. The frequency range of low-frequency voice signals is set to 20Hz to 300Hz, with a lower pitch. The frequency range of medium-frequency voice signals is set to 300Hz to 3400Hz. The sound within this range is the main energy concentration part of human speech. The frequency range of high-frequency voice signals is set to 3400Hz to 20000Hz, with a higher pitch.
[0038] Step 2: In the test environment, the signal transmitting device is controlled to output speech signals at different frequency stages, and the signal receiving device is used to sample the speech signals. The multi-dimensional acquisition module collects data sets according to different data sources, and the subsequent comprehensive evaluation of the degree and characteristics of reverberation is more accurate.
[0039] Step 3: Comprehensively analyze the energy attenuation trend and propagation characteristics of the speech signal in the environmental space. The reverberation analysis module calculates and generates waveform features Bxtz, spectrum features Putz and reverberation time Hxsj accordingly. The reverberation degree of the normalized environmental factors is more accurately judged.
[0040] Step 4: The evaluation and control module determines the reverberation degree of the speech signal in the test environment according to the reverberation time Hxsj, and generates a corresponding control report. The smaller the deviation of the judgment result, the better the subsequent dereverberation effect.
[0041] A system for judging the degree of environmental reverberation based on speech signals, comprising a multi-dimensional acquisition module, a reverberation analysis module and an evaluation and control module;
[0042] The multi-dimensional acquisition module includes an environmental constant unit, an environmental variable unit, a constant signal unit and a variable signal unit. The environmental constant unit collects environmental constant data sets through a network connection to an environmental scanning device. The constant data of the total indoor volume, the total indoor surface area and the average sound absorption coefficient in the test environment space are numbered and transmitted to the reverberation analysis module. The environmental constant data set is numbered as HJ tj , HJ mj and HJ xs The environmental variable unit collects environmental variable data sets through a network connection to the environmental monitoring device, numbers the thermodynamic temperature and air pressure data in the test environment space, and transmits them to the reverberation analysis module. The environmental variable data set number is HJ wd and HJ qy The constant signal unit collects the constant signal data set through the network connection signal transmitter, and transmits the initial speech signal data output by the signal transmitter to the reverberation analysis module after numbering. The constant signal data set is numbered as a represents the frequency range data of the initial speech signal. The variable signal unit collects the variable signal data set through the network connection signal receiving device, and transmits the final speech signal data received by the signal receiving device to the reverberation analysis module after numbering. The variable signal data set number is b represents the frequency range data of the final speech signal, and the multi-dimensional acquisition module is connected to the reverberation analysis module through the network;
[0043] The reverberation analysis module includes a time domain analysis unit, a spectrum analysis unit and a duration analysis unit. The reverberation analysis module collects reference data sets through a network connection database, and caches them to the reverberation analysis module after numbering. The reference data set consists of a reference time domain waveform algorithm model, a reference spectrum analysis algorithm model and a reference reverberation duration standard data set. The reference data set number is CK sy , CK pu and CK hxThe time domain analysis unit calculates and generates the waveform feature Bxtz based on the constant signal data set, the variable signal data set and the reference data set. The calculation formula is as follows:
[0044]
[0045] In the formula, Bxtz represents the waveform characteristics, It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the constant signal data set are substituted into the reference time domain waveform algorithm model to calculate the amplitude intensity feature data. It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the variable signal data set are substituted into the reference time-domain waveform algorithm model to calculate the amplitude intensity feature data, and the amplitude intensity features of the initial speech signal and the final speech signal are compared and analyzed according to the waveform feature Bxtz;
[0046] The spectrum analysis unit calculates and generates the spectrum feature Putz based on the constant signal data set, the variable signal data set and the reference data set. The calculation formula is as follows:
[0047]
[0048] In the formula, Putz represents the spectrum characteristics, It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the constant signal data set are substituted into the reference spectrum analysis algorithm model to calculate the frequency distribution characteristic data. It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the variable signal data set are substituted into the reference spectrum analysis algorithm model to calculate the frequency distribution feature data, and the frequency distribution features of the initial speech signal and the final speech signal are compared and analyzed according to the spectrum feature Putz;
[0049] The duration analysis unit calculates and generates the reverberation time Hxsj based on the environmental constant data set and the environmental variable data set. The reverberation analysis module is connected to the evaluation control module through the network. The calculation formula is as follows:
[0050]
[0051] In the formula, Hxsj represents the reverberation time, r represents the ideal gas constant in the air density calculation formula, It represents the air density of the current test environment obtained by substituting the thermodynamic temperature and pressure values in the environmental variable data set into the air density calculation formula. k represents the adiabatic index of air, and R represents the ideal gas constant in the sound speed calculation formula. It represents the sound speed of the current test environment obtained by substituting the thermodynamic temperature value in the environmental variable data set into the sound speed calculation formula. μ represents the conversion coefficient for air density and sound speed in the reverberation time calculation formula. It represents the ideal reverberation time calculated based on the total indoor volume, total indoor surface area and average sound absorption coefficient in the environmental constant data set under ideal temperature and sound velocity. It means converting the temperature and sound velocity variables in the actual detection environment, combining them with the reverberation time under ideal conditions to calculate the precise reverberation time. According to the reverberation time Hxsj, the correlation between the constant data and the variable data is deeply analyzed, and the reverberation degree of the environmental factors is judged more accurately after normalization.
[0052] The evaluation and control module compares the reverberation time Hxsj with the reference reverberation time standard dataset CK pu , judge the reverberation degree of the speech signal in the test environment, and the reverberation time Hxsj is lower than the reference reverberation time standard data set CK pu When the reverberation degree of the test environment is judged to be slight, no control report is generated, and the reverberation time Hxsj is greater than or equal to the reference reverberation time standard data set CK pu When the reverberation degree of the test environment is judged to be serious, a control report is generated. The control report includes the environmental constant data set, environmental variable data set, constant signal data set, variable signal data set, reference data set, waveform feature Bxtz, spectrum feature Putz and reverberation time Hxsj. The smaller the deviation of the judgment result, the better the subsequent dereverberation effect.
[0053] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for determining the degree of environmental reverberation based on a speech signal, characterized in that: The method for judging the degree of environmental reverberation based on speech signals is judged by a system for judging the degree of environmental reverberation based on speech signals. The system for judging the degree of environmental reverberation based on speech signals includes a multi-dimensional acquisition module, a reverberation analysis module, and an evaluation and control module. The method for judging the degree of environmental reverberation based on speech signals includes the following steps: Step 1: Record multiple voice signals of different frequencies, and each voice signal has the same duration; Step 2: In the test environment, the signal transmitting device is controlled to output a voice signal, and the signal receiving device is used to perform sampling, and the multi-dimensional acquisition module collects data sets according to different data sources; Step 3: Comprehensively analyze the energy attenuation trend and propagation characteristics of the speech signal in the ambient space, and the reverberation analysis module calculates and generates waveform features Bxtz, spectrum features Putz and reverberation time Hxsj accordingly; Step 4: The evaluation and control module determines the reverberation degree of the speech signal in the test environment according to the reverberation time Hxsj, and generates a corresponding control report.
2. The method for determining the degree of environmental reverberation based on a speech signal according to claim 1, characterized in that: The types of recorded voice signals include low-frequency voice signals, medium-frequency voice signals and high-frequency voice signals. The frequency range of the low-frequency voice signal is set to 20Hz to 300Hz, the frequency range of the medium-frequency voice signal is set to 300Hz to 3400Hz, and the frequency range of the high-frequency voice signal is set to 3400Hz to 20000Hz.
3. A system for determining the degree of environmental reverberation based on speech signals, characterized in that: Includes multi-dimensional acquisition module, reverberation analysis module and evaluation and control module; The multi-dimensional acquisition module includes an environmental constant unit, an environmental variable unit, a constant signal unit and a variable signal unit. The environmental constant unit collects an environmental constant data set through a network connection to an environmental scanning device, and transmits it to the reverberation analysis module after numbering. The environmental variable unit collects an environmental variable data set through a network connection to an environmental monitoring device, and transmits it to the reverberation analysis module after numbering. The constant signal unit collects a constant signal data set through a network connection to a signal transmitting device, and transmits it to the reverberation analysis module after numbering. The variable signal unit collects a variable signal data set through a network connection to a signal receiving device, and transmits it to the reverberation analysis module after numbering. The multi-dimensional acquisition module is connected to the reverberation analysis module through a network. The reverberation analysis module includes a time domain analysis unit, a spectrum analysis unit and a duration analysis unit. The reverberation analysis module collects a reference data set through a network connection to a database, and caches the data set to the reverberation analysis module after numbering. The time domain analysis unit calculates and generates a waveform feature Bxtz according to a constant signal data set, a variable signal data set and a reference data set. The spectrum analysis unit calculates and generates a spectrum feature Putz according to a constant signal data set, a variable signal data set and a reference data set. The duration analysis unit calculates and generates a reverberation time Hxsj according to an environmental constant data set and an environmental variable data set. The reverberation analysis module is connected to an evaluation and control module through a network connection. The evaluation and control module determines the reverberation degree of the speech signal in the test environment according to the reverberation time Hxsj, and generates a control report accordingly.
4. The system for determining the environmental reverberation degree based on speech signals according to claim 3, characterized in that: The environmental constant unit numbers the constant data of the total indoor volume, the total indoor surface area and the average sound absorption coefficient in the test environment space according to the characteristics of the environmental constant data set. The environmental constant data set is numbered as HJ tj , HJ mj and HJ xs The environmental variable unit numbers the thermodynamic temperature and pressure data in the test environment space according to the characteristics of the environmental variable data set, and the environmental variable data set is numbered HJ wd and HJ qy .
5. The system for determining the environmental reverberation degree based on speech signals according to claim 4, characterized in that: The constant signal unit numbers the initial speech signal data output by the signal transmitting device according to the characteristics of the constant signal data set, and the constant signal data set is numbered as follows: a represents the frequency range data of the initial speech signal, and the variable signal unit numbers the final speech signal data received by the signal receiving device according to the characteristics of the variable signal data set, and the variable signal data set number is b represents the frequency range data of the final speech signal.
6. The system for determining the environmental reverberation degree based on speech signals according to claim 5, characterized in that: The reference data set is composed of a reference time domain waveform algorithm model, a reference spectrum analysis algorithm model and a reference reverberation duration standard data set. The reverberation analysis module numbers the reference data set according to its characteristics. The reference data set is numbered CK sy , CK pu and CK hx .
7. The system for determining the environmental reverberation degree based on speech signals according to claim 6, characterized in that: The time domain analysis unit calculates and generates the waveform feature Bxtz according to the constant signal data set, the variable data set and the reference data set. The calculation formula is as follows: In the formula, Bxtz represents the waveform characteristics, It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the constant signal data set are substituted into the reference time domain waveform algorithm model to calculate the amplitude intensity feature data. It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the variable signal data set are substituted into the reference time domain waveform algorithm model to calculate the amplitude intensity feature data.
8. The system for determining the environmental reverberation degree based on speech signals according to claim 6, characterized in that: The spectrum analysis unit calculates and generates the spectrum feature Putz according to the constant signal data set, the variable data set and the reference data set. The calculation formula is as follows: In the formula, Putz represents the spectrum characteristics, It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the constant signal data set are substituted into the reference spectrum analysis algorithm model to calculate the frequency distribution characteristic data. It means that the low-frequency speech signal, the medium-frequency speech signal and the high-frequency speech signal in the variable signal data set are substituted into the reference spectrum analysis algorithm model to calculate the frequency distribution characteristic data.
9. The system for determining the environmental reverberation degree based on speech signals according to claim 6, characterized in that: The duration analysis unit calculates and generates the reverberation time Hxsj according to the environmental constant data set and the environmental variable data set, and the calculation formula is as follows: In the formula, Hxsj represents the reverberation time, r represents the ideal gas constant in the air density calculation formula, It represents the air density of the current test environment obtained by substituting the thermodynamic temperature and pressure values in the environmental variable data set into the air density calculation formula. k represents the adiabatic index of air, and R represents the ideal gas constant in the sound speed calculation formula. It represents the sound speed of the current test environment obtained by substituting the thermodynamic temperature value in the environmental variable data set into the sound speed calculation formula. μ represents the conversion coefficient for air density and sound speed in the reverberation time calculation formula. It represents the ideal reverberation time calculated based on the total indoor volume, total indoor surface area and average sound absorption coefficient in the environmental constant data set under ideal temperature and sound velocity. It means that the precise reverberation time is calculated by converting the temperature and sound speed variables in the actual detection environment and combining them with the reverberation time under ideal conditions.
10. The system for determining the environmental reverberation degree based on speech signals according to claim 9, characterized in that: The evaluation and control module compares the reverberation time Hxsj with the reference reverberation time standard data set CK. pu , judge the reverberation degree of the speech signal in the test environment, the reverberation time Hxsj is lower than the reference reverberation time standard data set CL pu When the reverberation degree of the test environment is judged to be slight, no control report is generated, and the reverberation time Hxsj is greater than or equal to the reference reverberation time standard data set CK pu When the reverberation degree of the test environment is judged to be serious, a control report is generated, which includes an environmental constant data set, an environmental variable data set, a constant signal data set, a variable signal data set, a reference data set, a waveform feature Bxtz, a spectrum feature Putz and a reverberation time Hxsj.
