CN118018940A - A loudness perception model and a round window excitation loudness calculation method based on its application - Google Patents

Abstract

The invention discloses a loudness perception model and an applied round window excitation loudness calculation method, wherein an auditory peripheral model is established, and the auditory peripheral model comprises an outer ear filter model, a middle ear dynamics model and a cochlear dynamics model, and can calculate the basal membrane speed under acoustic excitation and round window excitation; the data processing backend stage converts the base film speed to loudness. Based on the loudness perception model, the round window excitation loudness calculation method provided by the invention comprises the following steps: 1) The exciting force of the actuator is input to a round window membrane of the middle ear dynamics model to obtain cochlear fluid acceleration; 2) Substituting the cochlear fluid acceleration into a cochlear dynamics model to obtain the basal membrane speed under round window excitation; 3) And substituting the substrate film speed into the data processing rear end to obtain the excitation loudness of the round window. The invention can provide a loudness prediction value for the development of a round window excitation type artificial middle ear hearing test and match algorithm.

Description

A loudness perception model and a round window excitation loudness calculation method based on its application

Technical Field

The present invention relates to the field of hearing aid technology, and in particular to a loudness perception model and a round window excitation loudness calculation method using the model.

Background technique

The actuator in the traditional artificial middle ear implant acts on the ossicles, which requires the patient's ossicular chain to be intact, as shown in Figure 3. However, many patients also have lesions such as ossicular chain deformity and ossicular chain corrosion, which cannot provide intact ossicles, making it impossible to implant the traditional artificial middle ear. In response to this problem, Colletti et al. avoided the damaged ossicular chain during clinical implantation in 2006 and directly implanted the actuator of the VIBRANT SOUNDBRIDGE brand artificial middle ear produced by MED-EL, Austria, at the other entrance of the patient's cochlea, the round window, and compensated for hearing by mechanically exciting the round window membrane 7 by the actuator, and achieved good clinical results. Its principle is shown in Figure 4. This round window excitation hearing loss compensation mode expands the treatment field of the traditional artificial middle ear, enabling it to treat mixed deafness accompanied by abnormalities of the tympanic cavity or ossicular chain (such as congenital external middle ear malformation, ossicular corrosion caused by otitis media, etc.). Due to the above advantages, the round window excitation artificial middle ear has been widely used in clinical practice, but there are problems such as large initial fitting errors and difficulty in later fine-tuning during its fitting process. This is mainly because there is no hearing fitting algorithm for this excitation mode so far. In clinical practice, fitting is completed by borrowing hearing aid fitting algorithms (such as NAL proposed by the National Acoustic Laboratory of Australia, DSL proposed by the National Audiology Center of Canada, etc.). These classic hearing aid fitting algorithms are specially developed for the sound transmission characteristics of hearing aids. However, for round window excitation hearing compensation, the input energy enters the cochlea through the round window of the cochlea; for hearing aid hearing compensation, the input energy enters the cochlea through the eardrum, ossicular chain, and oval window of the cochlea. These two types of energy are transmitted into the cochlea in different ways and have different loudness perception mechanisms, so borrowing hearing aid fitting algorithms for round window excitation is not effective. Therefore, it is necessary to develop a hearing fitting algorithm specifically for round window excitation.

Loudness is an important parameter of sound quality, which reflects the subjective perception of the human auditory center to the strength of sound. In order to accurately calculate the sound perception effect of the human ear and design electroacoustic equipment such as hearing aids, researchers at home and abroad have established a variety of loudness perception calculation methods. The classic one is the Moore-Glasberg method (ISO 532-2:2017) listed as a standard by the International Organization for Standardization. The typical hearing aid fitting algorithm is based on the loudness perception calculation method. For example, the NAL algorithm proposed by the Australian National Acoustic Laboratory is designed based on the Moore-Glasberg loudness perception calculation method. As shown in Figure 5, in the Moore-Glasberg loudness perception calculation method, after the sound signal passes through the outer ear and middle ear filters, a filter group is used to decompose the sound signal in frequency, and then the excitation on each equivalent rectangular bandwidth is calculated, and finally the loudness under sound excitation is calculated.

However, this loudness perception calculation method cannot be used for round window excitation because it has the following shortcomings:

Insufficient 1: This loudness perception calculation method uses filters to directly simulate the sound transmission characteristics of the outer ear and middle ear, without considering the physiological structures in the middle ear such as the eardrum, hammer, incus, stapes, tendons, etc. It is similar to a black box model and cannot simulate conductive hearing loss such as ossicular chain damage, and mixed hearing loss accompanied by conductive hearing loss. These two types of hearing loss are the main targets of round window excitation, which makes the traditional loudness calculation model unable to simulate the main hearing loss status of round window excitation patients.

Disadvantage 2: This loudness perception calculation method is a fit to the results of normal human ear audiometry experiments under acoustic excitation. The round window excitation artificial middle ear is an implantable hearing aid device and cannot be used for normal human ear audiometry experiments. Therefore, the method of using filters to fit the loudness perception of the human ear under round window excitation cannot be used to establish a loudness model.

Insufficient 3: This loudness perception calculation method fits the loudness perception of the human ear under free-field acoustic excitation. The sound is transmitted from the ear canal to the middle ear, and then from the stapes of the middle ear to the cochlea through the oval window of the cochlea. It can be applied to hearing aids and other hearing aids that have the same sound transmission path as the normal human ear. However, the input energy of the round window-excited artificial middle ear crosses the middle ear and is directly transmitted to the cochlea through another window (round window) of the cochlea. The sound transmission path of the round window excitation is different from the transmission path of normal sound perception. Therefore, this loudness model cannot be applied to the round window-excited artificial middle ear.

Summary of the invention

In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a loudness perception model and a round window excitation loudness calculation method using the model. The model can simulate and analyze the loudness perception characteristics of patients with conductive hearing loss and mixed hearing loss by changing the material properties of the tissue in the middle ear of the model, and can calculate the basilar membrane velocity under acoustic excitation and round window excitation, and convert the basilar membrane velocity into loudness. The algorithm can solve the current problem of being unable to accurately calculate the loudness of round window excitation.

In order to solve the above technical problems, the present invention adopts the following technical solutions:

The present invention provides a loudness perception model and a circular window excitation loudness calculation method using the model, including a model processing stage and a model processing backend stage. The model processing stage includes the following steps:

Step 1: Establish a loudness perception model, including the outer ear filter model, the middle ear dynamics model, and the cochlear dynamics model;

Step 2: input the excitation force of the round window excitation artificial middle ear actuator into the round window membrane in the middle ear dynamics model in the constructed loudness perception model, and calculate the acceleration of the cochlear fluid under the round window excitation;

Step 3: Substitute the cochlear fluid acceleration into the cochlear dynamics model in the constructed loudness perception model to calculate the basilar membrane velocity on each cochlear segment under round window excitation.

The model processing backend stage includes the following steps:

Step 4: Calculate the absolute value of the basilar membrane velocity on each cochlear segment; integrate it over time, and then obtain the initial excitation on each cochlear segment;

Step 5: Divide the cochlea into multiple excited segments, each of which contains multiple cochlear segments, average the initial excitations of the cochlear segments in each excited segment, and then obtain the excitation of each excited segment;

Step 6: convert the excitement of each excitement segment into the characteristic loudness of each excitement segment; sum the characteristic loudness of all excitement segments to obtain the initial loudness;

Step 7: According to the function S, the initial loudness _LI is converted into the loudness level _LL ; the function S is _LL = 52.45tanh(1.427( _LI -0.68)) + 37.2;

Step 8: According to the function P, convert the loudness level L _L into the loudness L; the function P is L = 8 × 10 ^-4 (0.1L _L +1.2) ^4.4 .

Preferably, in step 1, the method for establishing the middle ear dynamics model is: simplifying the malleus, incus, stapes and cochlear fluid into mass, simplifying the vestibular aqueduct and the cochlear aqueduct into damping, and simplifying the tympanic membrane, the anterior ligament of the malleus, the posterior ligament of the incus, the annular ligament of the stapes base, the anvil-hammer joint, the anvil-stapedial joint and the round window membrane into stiffness and damping; the model can simulate the sound transmission characteristics of the human ear under acoustic excitation and round window excitation while having fast calculation.

Preferably, in step 1, the method for establishing the cochlear dynamics model is: simplifying the cochlea into a one-dimensional fluid-coupled conical cochlea with a round window and an oval window; the model divides the cochlea into 100 segments, each segment consisting of basilar membrane mass, reticular plate mass, tectorial membrane mass, basilar membrane longitudinal coupling stiffness and damping, basilar membrane bending stiffness and damping, outer hair cell stiffness and damping, ciliary bundle stiffness and damping, and tectorial membrane stiffness and damping; the model can accurately calculate the basilar membrane velocity on each cochlear segment under acoustic excitation and round window excitation.

Preferably, in step 1, a digital filter is used to establish an outer ear filter model, specifically: Where _PE is the sound pressure at the eardrum, _PFree is the sound pressure at the free-field sound source, n is the nth sampling point of the time signal, k is the order of the filter, and a(k) is the gain parameter of the kth order filter.

Preferably, in step 1, the motion differential equation of the middle ear dynamics model is:

where m _M , m _I , m _S and m _CF are the mass of the malleus, the mass of the incus, the mass of the stapes and the mass of the cochlear fluid, respectively; k _AML , k _PIL , k _AL , k _E , k _IMJ , k _ISJ and k _RW are the stiffness of the anterior malleus ligament, the stiffness of the posterior incus ligament, the stiffness of the annular ligament at the base of the stapes, the stiffness of the tympanic membrane, the stiffness of the anvil-hammer joint, the stiffness of the incus-stapedial joint and the stiffness of the round window membrane, respectively; c _AML , c _PIL , c _AL , c _E , c _IMJ , c _ISJ , c _VA , c _CA and c _RW are the damping of the anterior malleus ligament, the damping of the posterior incus ligament, the damping of the annular ligament at the base of the stapes, the damping of the tympanic membrane, the damping of the anvil-hammer joint, the damping of the incus-stapedial joint, the damping of the vestibular aqueduct, the damping of the cochlear aqueduct and the damping of the round window membrane, respectively; _AE is the area of the tympanic membrane; _RW is the force exerted by the round window excitation artificial middle ear on the round window membrane.

Preferably, in step 1, the cochlear dynamics model simplifies the cochlea into a one-dimensional fluid-coupled cone cochlea, and the transmission of traveling waves in the cochlear dynamics model is described as:

Where p is the pressure difference across the cochlear diaphragm, is the radial acceleration in the cochlear fluid, ρ is the density of the cochlear fluid, H is the effective height of the scala vestibuli and scala tympani, and H = π ² _AC /(8W _B ); _AC is the cross-sectional area of the cochlea, W _B is the width of the basilar membrane;

The cochlear dynamics model is driven by the pressure difference between the oval window membrane and the round window membrane;

The cochlear dynamics model divides the cochlea into 100 segments, and the force analysis of each segment is very similar. The motion differential equation of the i-th segment of the cochlear dynamics model is:

Among them, m _Bi , m _Ri and m _Ti are the mass of the basilar membrane, the mass of the reticular plate and the mass of the tectorial membrane on the i-th segment, respectively; k _Li , k _Bi , k _OHCi , k _HBi and k _Ti are the longitudinal coupling stiffness of the basilar membrane, the bending stiffness of the basilar membrane, the stiffness of the outer hair cells, the stiffness of the ciliary bundle and the stiffness of the tectorial membrane, respectively; c _Li , c _Bi , c _OHCi , c _HBi and c _Ti are the longitudinal coupling damping of the basilar membrane, the bending damping of the basilar membrane, the damping of the outer hair cells, the damping of the ciliary bundle and the damping of the tectorial membrane, respectively.

Preferably, in step 4, the absolute value and time of the basilar membrane dormitories on each segment of the cochlea are calculated.

E _i,0 ＝0

Integrate to find the initial excitement:

Wherein, E _i,t is the initial excitation of the ith basilar membrane segment at time t, vi _,t is the velocity of the ith basilar membrane segment at time t, the sampling frequency f _s = 200 kHz, and the time constant τ = 15 ms.

Preferably, in step 5, according to the 1/3 octave, the cochlear segments with characteristic frequencies of 88.4-11313.7 Hz are divided into 21 excited segments, and the cochlear segments with characteristic frequencies less than 88.4 Hz and greater than 11313.7 Hz are respectively regarded as an excited segment, and the cochlea is divided into 23 excited segments in total; the excitation _Ex on the x-th excited segment is calculated as:

Where M is the number of cochlear segments contained in the xth excited segment.

Preferably, in step 6, the excitement on the x-th excitement segment is converted into the characteristic loudness on the x-th excitement segment by the following formula:

Among them, _Cx , _αx and _Ax are constants for calculating characteristic loudness on the xth exciting segment, and they are different in each exciting segment; the initial loudness L _I is obtained by summing the characteristic loudness on all exciting segments:

The beneficial effects of the present invention are:

1. The loudness perception model of the present invention is based on the physiological structure of the human ear. It can simulate and analyze the loudness perception characteristics of patients with conductive hearing loss and mixed hearing loss by changing the tissue material properties of the middle ear and cochlea in the model.

2. The auditory peripheral model parameters in the loudness perception model of the present invention are based on existing experimental reports, and the optimization of the parameters in the model processing backend stage is based on the loudness perception data under acoustic stimulation. Therefore, the establishment of the loudness perception model of the present invention does not require a round window excitation human ear audiometry experiment.

3. The loudness model based on the physiological structure of the human ear of the present invention takes the round window membrane into consideration, and can calculate the velocity of the basilar membrane under the round window excitation. The model processing back-end stage can calculate the loudness based on the basilar membrane velocity. Therefore, the loudness perception model of the present invention can be used to calculate the loudness of the round window excitation, which makes up for the deficiency that the existing loudness model can only be used to analyze the loudness perception characteristics of the normal sound transmission path.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

Figure 1 is a simplified schematic diagram of a human ear;

Figure 2 is a schematic diagram of a cross section of the cochlea

Fig. 3 is a schematic diagram of a conventional artificial middle ear;

FIG4 is a schematic diagram of a round window excitation artificial middle ear;

FIG5 is a calculation flow chart of the Moore-Glasberg method in the international standard ISO 532-2:2017;

FIG6 is a flow chart of a loudness perception model and a round window excitation loudness calculation method using the model provided by an example of the present invention;

FIG7 is a schematic diagram of a function S provided by an example of the present invention;

FIG8 is a schematic diagram of a function P provided by an example of the present invention;

FIG9 is a comparison diagram of the outer ear transfer function provided by an example of the present invention and the standard ANSIS3.4;

FIG10 is a schematic diagram of a middle ear dynamics model provided by an example of the present invention;

FIG11 is a comparison diagram of the middle ear transfer function under acoustic excitation provided by an example of the present invention and the experimental data of Nakajima et al.;

12 is a comparison diagram of stapes velocity under 94 dB sound pressure level excitation provided by an example of the present invention and experimental data statistically obtained by Koch et al.;

FIG13 is a comparison diagram of the round window excitation transfer function provided by an example of the present invention and the experimental data of Nakajima et al.;

FIG14 is a simplified schematic diagram of a one-dimensional fluid-coupled conical cochlea;

FIG15 is a schematic diagram of the i-th segment of the cochlear dynamics model provided by an example of the present invention;

FIG16 is a comparison diagram of the frequency-selective characteristics of the basement membrane provided by the example of the present invention and the experimental data of Kringlebotn et al.;

FIG17 is a comparison diagram of the basilar membrane frequency response characteristics provided by the example of the present invention and the experimental data of Gundersen et al. and Stenfelt et al.;

FIG18 is a schematic diagram of an outer hair cell circuit model provided by an example of the present invention;

FIG19 is a comparison diagram of the basilar membrane displacement of the active model and the passive model provided by the example of the present invention and the experimental data of the living cochlea and the dead cochlea by Johnstone et al.;

FIG20 is a comparison diagram of equal loudness curves of the loudness perception model of the present invention and the international standard ISO 226;

FIG21 is a comparison diagram of the loudness of broadband noise under acoustic excitation of the loudness perception model of the present invention and experimental data of Zwicker et al.;

FIG22 is a comparison chart of the loudness of pure tones under frequency domain masking and Zwicker experimental data.

Detailed ways

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.

Referring to Figures 1 to 4, the human ear is mainly composed of three parts: the outer ear, the middle ear, and the inner ear. In normal human ears, the outer auricle of the outer ear first collects the sound into the external auditory canal 1, causing the eardrum 2 to vibrate; then it drives the hammer 3 of the middle ear to move, and transmits it to the incus 4 and the stapes 5; the stapes transmits the vibration energy through the cochlear oval window membrane 30 through its bottom plate to the cochlea of the inner ear; the cochlea senses the input vibration energy through the fluid-solid coupling effect of the lymph 6 and the basilar membrane 13 inside it, and the active amplification function of the outer hair cells 15 (actively amplifying the tiny vibrations induced by the basilar membrane), so that the inner hair cells in the cochlea sense the input vibration energy, generate nerve pulses and transmit them to the auditory nerve, so that people can hear the sound. Sensorineural hearing loss is mainly caused by the damage of the outer hair cells 15, which cannot amplify the input weak vibration signals, so that the patient cannot hear the external low-intensity sound. The hearing aid device enables the patient to hear by amplifying the weak sound signal before it is input into the cochlea, thereby compensating for the patient's hearing loss.

The artificial middle ear is mainly composed of a microphone 20, a signal processing unit 21, an actuator 22 implanted in the body, and a power supply 23. Among them, the actuator 22 is usually coupled to the ossicles of the middle ear, such as the incus body, the long process of the incus, the stapes, etc. Its working process is as follows: the microphone 20 first collects the sound, converts it into an electrical signal and transmits it to the signal processing unit 21; the signal processing unit 21 performs corresponding amplification and other processing on the signal according to the patient's hearing loss, and then outputs the signal to the actuator 22; under the action of the driving electrical signal, the actuator 22 performs mechanical movement to drive the ear tissue it acts on. Finally, the vibration energy is input into the cochlea 19 in the inner ear through the oval window membrane 30 of the cochlea to achieve the purpose of hearing compensation. Compared with traditional hearing aids that compensate for hearing loss through sound excitation, the artificial middle ear, which uses mechanical excitation, has the advantages of not blocking the ear canal, no sound feedback, high speech clarity and strong high-frequency gain, which makes up for the shortcomings of traditional hearing aids.

The present embodiment provides a method for calculating the loudness of acoustic excitation based on a human ear loudness perception model, including a model processing stage and a model processing backend stage. The model processing stage includes: establishing a loudness perception model, the method of which is shown in the following step 1, including an outer ear filter model, a middle ear dynamics model and a cochlear dynamics model; under acoustic excitation, the sound pressure at a free-field sound source is converted into the sound pressure at the eardrum after passing through the outer ear filter model; the sound pressure at the eardrum is input into the eardrum of the middle ear dynamics model to obtain the cochlear fluid acceleration; the cochlear fluid acceleration is used as the drive of the cochlear dynamics model to calculate the basilar membrane velocity on each cochlear segment.

The model processing backend stage is the same as the model processing backend stage method in the following loudness perception model and the round window excitation loudness calculation method used.

Referring to FIG6 , this embodiment provides a loudness perception model and a round window excitation loudness calculation method for the application. The loudness model of this embodiment can simulate sound excitation in order to verify the model using sound excitation loudness data. The main purpose of the study is to calculate the round window excitation loudness. The method includes a model processing stage and a model processing backend stage. The model processing stage includes the following steps:

Step 1: Establish a loudness perception model (auditory peripheral model), including the outer ear filter model, the middle ear dynamics model, and the cochlear dynamics model;

Step 2: input the excitation force of the round window excitation artificial middle ear actuator into the round window membrane in the middle ear dynamics model in the constructed loudness perception model, and calculate the acceleration of the cochlear fluid under the round window excitation;

Step 3: Substitute the acceleration of the cochlear fluid into the cochlear dynamics model in the constructed loudness perception model to calculate the velocity of the basilar membrane on each cochlear segment under round window excitation;

Step 4: Substitute the basilar membrane velocity of each cochlear segment into the data processing backend of the constructed loudness perception model to calculate the round window excitation loudness.

The outer ear filter model is:

Where, _PE is the sound pressure at the eardrum, _PFree is the sound pressure at the free-field sound source, n is the nth sampling point of the time signal, k is the order of the filter, and a(k) is the gain parameter of the kth order filter.

The outer ear transfer function of the outer ear filter model is compared with the standard ANSIS3.4 as shown in Figure 9. The outer ear transfer function of the present invention is very consistent with the standard ANSIS3.4. The results show that the outer ear filter model of the present invention can accurately simulate the collection and resonance functions of the outer ear during the transmission of sound pressure from the free field to the eardrum.

The schematic diagram of the middle ear dynamics model is shown in FIG10 , and its specific motion differential equation is:

where m _M , m _I , m _S and m _CF are the mass of the malleus, the mass of the incus, the mass of the stapes and the mass of the cochlear fluid, respectively; k _AML , k _PIL , k _AL , k _E , k _IMJ , k _ISJ and k _RW are the stiffness of the anterior malleus ligament, the stiffness of the posterior incus ligament, the stiffness of the annular ligament at the base of the stapes, the stiffness of the tympanic membrane, the stiffness of the anvil-hammer joint, the stiffness of the incus-stapedial joint and the stiffness of the round window membrane, respectively; c _AML , c _PIL , c _AL , c _E , c _IMJ , c _ISJ , c _VA , c _CA and c _RW are the damping of the anterior malleus ligament, the damping of the posterior incus ligament, the damping of the annular ligament at the base of the stapes, the damping of the tympanic membrane, the damping of the anvil-hammer joint, the damping of the incus-stapedial joint, the damping of the vestibular aqueduct, the damping of the cochlear aqueduct and the damping of the round window membrane, respectively; _AE is the area of the tympanic membrane; _RW is the force exerted by the round window excitation artificial middle ear on the round window membrane.

The middle ear transfer function and stapes velocity of the middle ear dynamics model under acoustic excitation are compared with the experimental data of Nakajima et al. and the experimental data of Koch et al., as shown in Figures 11 and 12. The middle ear transfer function of the middle ear dynamics model under acoustic excitation is very consistent with the experimental data of Nakajima et al., and the stapes velocity under 94dB sound pressure level excitation is also within the experimental data range of Koch et al. The results show that the middle ear dynamics model of the present invention can accurately simulate the sound transmission characteristics of the middle ear under acoustic excitation.

The round window excitation transfer function of the middle ear dynamics model is compared with the experimental data of Nakajima et al., as shown in Figure 13. The round window excitation transfer function of the middle ear dynamics model is consistent with the experimental data of Nakajima et al. The results show that the middle ear dynamics model of the present invention can simulate the sound transmission characteristics of the human ear under round window excitation.

The cochlear dynamics model simplifies the cochlea into a one-dimensional fluid-coupled cone cochlea, as shown in Figure 14. The transmission of traveling waves in the cochlear dynamics model is described as:

Where p is the pressure difference across the cochlear diaphragm, is the radial acceleration in the cochlear fluid, ρ is the density of the cochlear fluid, H is the effective height of the scala vestibuli and scala tympani, and H = π ² _AC /(8W _B ). _AC is the cross-sectional area of the cochlea, and W _B is the width of the basilar membrane.

The cochlear dynamics model is driven by the pressure difference between the oval window membrane and the round window membrane.

The cochlear dynamics model divides the cochlea into 100 segments, and the force analysis of each segment is very similar. The schematic diagram of the i-th segment of the cochlear dynamics model is shown in Figure 15. The motion differential equation of the i-th segment of the cochlear dynamics model is:

Among them, m _Bi , m _Ri and m _Ti are the mass of the basilar membrane, the mass of the reticular plate and the mass of the tectorial membrane on the i-th segment, respectively; k _Li , k _Bi , k _OHCi , k _HBi and k _Ti are the longitudinal coupling stiffness of the basilar membrane, the bending stiffness of the basilar membrane, the stiffness of the outer hair cells, the stiffness of the ciliary bundle and the stiffness of the tectorial membrane, respectively; c _Li , c _Bi , c _OHCi , c _HBi and c _Ti are the longitudinal coupling damping of the basilar membrane, the bending damping of the basilar membrane, the damping of the outer hair cells, the damping of the ciliary bundle and the damping of the tectorial membrane, respectively.

Furthermore, the comparison of the frequency selection characteristics and frequency response characteristics of the basilar membrane in the cochlear dynamics model with the experimental data is shown in Figures 16 and 17, respectively. The frequency selection characteristic curve of the basilar membrane in the cochlear dynamics model is very consistent with the experimental data of Kringlebotn et al., and the frequency response characteristics of the basilar membrane at 12 mm from the stapes are relatively consistent with the experimental data of Gundersen et al. and Stenfelt et al. The results show that the cochlear dynamics model of the present invention can simulate the sound transmission characteristics of the cochlea more accurately.

Furthermore, the cochlear dynamics model includes an outer hair cell circuit model (as shown in FIG18 ) to simulate the active amplification function of the cochlea caused by the electrokinetic movement of the outer hair cells. The transmission of traveling waves in the cochlea causes the displacement of the basilar membrane, the reticular plate, and the tectorial membrane, which further causes the deflection of the ciliary bundle and the contraction of the outer hair cells; the deflection of the ciliary bundle and the contraction of the outer hair cells respectively cause the generation of transduction current and piezoelectric current, which in turn causes the electrokinetic movement of the outer hair cells; the electrokinetic movement of the outer hair cells generates an active excitation force, which further amplifies the vibration of the basilar membrane.

Furthermore, the cochlear model including the outer hair cell circuit model is generally referred to as an active cochlear model, and the cochlear model without considering the active amplification of the outer hair cells is generally referred to as a passive cochlear model. As shown in FIG19 , the basilar membrane displacement of the active model and the passive model of the present invention is compared with the experimental data of the living cochlea and the dead cochlea of Johnstone et al. The basilar membrane displacement of the active model and the passive model of the present invention is very consistent with the experimental data of Johnstone et al. The results show that the outer hair cell circuit model of the present invention can accurately simulate the active amplification of the outer hair cells, and the cochlear dynamics model can accurately simulate the movement of the basilar membrane under different excitation amplitudes.

Furthermore, the transduction current and piezoelectric current in the outer hair cell circuit model are respectively:

where V _EP and _VOHC are the resting potentials of the cochlear duct and outer hair cells, respectively, f is the characteristic frequency of the cochlear segment, ε is the piezoelectric coupling coefficient, _G1 is the conductivity of the ciliary bundle, and θ is the inclination angle of the reticular plate.

Furthermore, in the outer hair cell circuit model, the outer hair cell inner potential ( ) and extracellular potential (/> ) are:

Where R _ML and R _TL are the ground resistances of the cochlear duct and scala tympani, respectively; _Ra and _Rb are the resistances of the top and base of the outer hair cells, respectively; and _Ca and _Cb are the capacitances of the top and base of the outer hair cells, respectively.

Furthermore, the active excitation force generated by the outer hair cell circuit model is

where _NOHC is the number of outer hair cells per segment.

The model processing back-end stage includes the following steps: the absolute value and time integration of the basilar membrane dormitories on each segment of the cochlea are calculated to obtain the initial excitation:

E _i,0 ＝0

Wherein, E _i,t is the initial excitation of the ith basilar membrane segment at time t, vi _,t is the velocity of the ith basilar membrane segment at time t, the sampling frequency f _s = 200 kHz, and the time constant τ = 15 ms.

The data processing backend stage includes the following steps: According to the 1/3 octave, the cochlear segments with characteristic frequencies of 88.4-11313.7Hz are divided into 21 excited segments, and the cochlear segments with characteristic frequencies less than 88.4Hz and greater than 11313.7Hz are each regarded as an excited segment, and the cochlea is divided into 23 excited segments in total. The excitation ( _Ex ) on the xth excited segment can be calculated as:

Where M is the number of cochlear segments contained in the xth excited segment.

Furthermore, the excitement on the x-th excitement segment is converted into the characteristic loudness on the x-th excitement segment by the following formula:

Wherein, _Cx , _αx and _Ax are all constants for calculating characteristic loudness on the xth excitement segment, and are different in each excitement segment. Preferably, the values of these parameters are calculated using an adaptive particle swarm optimization algorithm, and the optimization target is the equal loudness curve in the international standard ISO 226. The comparison of the equal loudness curve of the loudness perception model of the present invention and the international standard ISO 226 is shown in FIG20.

Furthermore, the characteristic loudness of all excited segments is summed to obtain the initial loudness (L _I ):

See Figures 7 and 8; the data processing backend stage includes the following steps: according to the function S, the initial loudness _LI is converted into the loudness level _LL ; the loudness level is defined as: the loudness level of the sound is the sound pressure level of 1000Hz pure tone excitation under the condition of equal loudness. According to the above definition of loudness level, the initial loudness ( _LI ) of the loudness model of the present invention can be converted into the loudness level ( _LL ): the function S is

L _L =52.45tanh(1.427(L _I -0.68))+37.2;

According to the relationship between loudness and loudness level described in the American standard ANSI S3.4, the loudness level L _L is converted to loudness L according to the function P; the function P is:

L＝8×10 ^-4 (0.1L _L +1.2) ^4.4 .

The loudness perception model of the present invention is compared with the experimental data for the loudness of bandwidth noise under acoustic excitation and the loudness of pure tone under frequency domain masking, as shown in Figures 21 and 22. The loudness of the bandwidth noise under acoustic excitation of the loudness perception model of the present invention is consistent with the experimental data of Zwicker and others, and the loudness of pure tone under frequency domain masking is very consistent with the experimental data of Zwicker. The results show that the loudness perception model of the present invention can more accurately calculate the loudness of bandwidth noise and the loudness of pure tone under frequency domain masking.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims (9)

1. A loudness perception model and a round window excitation loudness calculation method using the model, wherein the loudness perception model includes an outer ear filter model, a middle ear dynamics model and a cochlear dynamics model; the model processing stage and the model processing backend stage are included, wherein the model processing stage includes the following steps: Step 1: Establish the outer ear filter model, the middle ear dynamics model and the cochlear dynamics model; Step 2: input the excitation force of the round window excitation artificial middle ear actuator into the round window membrane in the middle ear dynamics model in the constructed loudness perception model, and calculate the acceleration of the cochlear fluid under the round window excitation; Step 3: Substitute the cochlear fluid acceleration into the cochlear dynamics model in the constructed loudness perception model to calculate the basilar membrane velocity on each cochlear segment under round window excitation. The model processing backend stage includes the following steps: Step 4: Calculate the absolute value of the basilar membrane velocity on each cochlear segment; integrate it over time, and then obtain the initial excitation on each cochlear segment; Step 5: Divide the cochlea into multiple excited segments, each of which contains multiple cochlear segments, average the initial excitations of the cochlear segments in each excited segment, and then obtain the excitation of each excited segment; Step 6: convert the excitement of each excitement segment into the characteristic loudness of each excitement segment; sum the characteristic loudness of all excitement segments to obtain the initial loudness; Step 7: According to the function S, the initial loudness _LI is converted into the loudness level _LL ; the function S is _LL = 52.45tanh(1.427( _LI -0.68)) + 37.2; Step 8: According to the function P, convert the loudness level L _L into the loudness L; the function P is L = 8 × 10 ^-4 (0.1L _L +1.2) ^4.4 . 2. The method of claim 1, wherein in step 1, the method for establishing the middle ear dynamics model is as follows: the malleus, incus, stapes and cochlear fluid are simplified to mass, the vestibular aqueduct and the cochlear aqueduct are simplified to damping, and the tympanic membrane, the anterior ligament of the malleus, the posterior ligament of the incus, the annular ligament of the stapes base, the anvil-hammer joint, the anvil-stapedial joint and the round window membrane are simplified to stiffness and damping; the model can simulate the sound transmission characteristics of the human ear under acoustic excitation and round window excitation while having fast calculation. 3. The method as claimed in claim 2 is characterized in that in step 1, the method for establishing the cochlear dynamics model is: simplifying the cochlea into a one-dimensional fluid-coupled cone cochlea with a round window and an oval window; the model divides the cochlea into 100 segments, each segment is composed of basilar membrane mass, reticular plate mass, tectorial membrane mass, basilar membrane longitudinal coupling stiffness and damping, basilar membrane bending stiffness and damping, outer hair cell stiffness and damping, ciliary bundle stiffness and damping, and tectorial membrane stiffness and damping; the model can accurately calculate the basilar membrane velocity on each cochlear segment under acoustic excitation and round window excitation. 4. The method according to claim 3, characterized in that, in step 1, a digital filter is used to establish an outer ear filter model, specifically: Where _PE is the sound pressure at the eardrum, _PFree is the sound pressure at the free-field sound source, n is the nth sampling point of the time signal, k is the order of the filter, and a(k) is the gain parameter of the kth order filter. 5. The method according to claim 4, characterized in that in step 1, the motion differential equation of the middle ear dynamics model is: where m _M , m _I , m _S and m _CF are the mass of the malleus, the mass of the incus, the mass of the stapes and the mass of the cochlear fluid, respectively; k _AML , k _PIL , k _AL , k _E , k _IMJ , k _ISJ and k _RW are the stiffness of the anterior malleus ligament, the stiffness of the posterior incus ligament, the stiffness of the annular ligament at the base of the stapes, the stiffness of the tympanic membrane, the stiffness of the anvil-hammer joint, the stiffness of the incus-stapedial joint and the stiffness of the round window membrane, respectively; c _AML , c _PIL , c _AL , c _E , c _IMJ , c _ISJ , c _VA , c _CA and c _RW are the damping of the anterior malleus ligament, the damping of the posterior incus ligament, the damping of the annular ligament at the base of the stapes, the damping of the tympanic membrane, the damping of the anvil-hammer joint, the damping of the incus-stapedial joint, the damping of the vestibular aqueduct, the damping of the cochlear aqueduct and the damping of the round window membrane, respectively; _AE is the area of the tympanic membrane; _RW is the force exerted by the round window excitation artificial middle ear on the round window membrane. 6. The method according to claim 5, characterized in that in step 1, the cochlear dynamics model simplifies the cochlea into a one-dimensional fluid-coupled cone cochlea, and the transmission of traveling waves in the cochlear dynamics model is described as: Where p is the pressure difference across the cochlear diaphragm, is the radial acceleration in the cochlear fluid, ρ is the density of the cochlear fluid, H is the effective height of the scala vestibuli and scala tympani, and H = π ² _AC /(8W _B ); _AC is the cross-sectional area of the cochlea, W _B is the width of the basilar membrane; The cochlear dynamics model is driven by the pressure difference between the oval window membrane and the round window membrane; The cochlear dynamics model divides the cochlea into 100 segments, and the force analysis of each segment is very similar. The motion differential equation of the i-th segment of the cochlear dynamics model is: Among them, m _Bi , m _Ri and m _Ti are the mass of the basilar membrane, the mass of the reticular plate and the mass of the tectorial membrane on the i-th segment, respectively; k _Li , k _Bi , k _OHCi , k _HBi and k _Ti are the longitudinal coupling stiffness of the basilar membrane, the bending stiffness of the basilar membrane, the stiffness of the outer hair cells, the stiffness of the ciliary bundle and the stiffness of the tectorial membrane, respectively; c _Li , c _Bi , c _OHCi , c _HBi and c _Ti are the longitudinal coupling damping of the basilar membrane, the bending damping of the basilar membrane, the damping of the outer hair cells, the damping of the ciliary bundle and the damping of the tectorial membrane, respectively. 7. The method according to claim 6, characterized in that in step 4, the absolute value and time integration of the basilar membrane dormitories on each segment of the cochlea are calculated to obtain the initial excitation: E _i,0 = 0 Wherein, E _i,t is the initial excitation of the ith basilar membrane segment at time t, vi _,t is the velocity of the ith basilar membrane segment at time t, the sampling frequency f _s = 200 kHz, and the time constant τ = 15 ms. 8. The method of claim 7, characterized in that in step 5, according to the 1/3 octave, the cochlear segment with a characteristic frequency of 88.4-11313.7 Hz is divided into 21 excited segments, and the cochlear segments with characteristic frequencies less than 88.4 Hz and greater than 11313.7 Hz are respectively regarded as an excited segment, and the cochlea is divided into 23 excited segments in total; the excitement _Ex on the x-th excited segment is calculated as: Where M is the number of cochlear segments contained in the xth excited segment. 9. The method according to claim 8, characterized in that, in step 6, the excitement on the x-th excitement segment is converted into the characteristic loudness on the x-th excitement segment by the following formula: Among them, _Cx , _αx and _Ax are constants for calculating characteristic loudness on the xth exciting segment, and they are different in each exciting segment; the initial loudness L _I is obtained by summing the characteristic loudness on all exciting segments: