Metamaterial‐Augmented Head‐Mounted Audio Module

Recent developments in metamaterial research have shown promising progress in overcoming the fundamental physical limitations of acoustics. While the majority of this research is curiosity‐driven and focused on fundamental physics, there has been a scarcity of acoustic metamaterials research that can have a direct and immediate impact on real‐world applications. In this study, an acoustic metamaterial inside a head‐mounted audio module is incorporate to achieve a 3–7 dB broadband sound pressure level (SPL) improvement in the bass range (50–700 Hz), which is the most challenging frequency range for radiation enhancement. The adoption of the acoustic metamaterial not only enhances voltage sensitivity near the intrinsic metamaterial resonance frequency, but also extends the broadband enhancement to a lower transducer resonance frequency by carefully engineering the metamaterial‐transducer resonant coupling. The improvement of the audio module is demonstrated through fully coupled electrical‐mechanical‐acoustical numerical simulations using finite element analysis, which are validated against comprehensive measurements including electrical impedance analysis, speaker diaphragm displacement analysis, voltage sensitivity and power sensitivity analysis, and spectrogram. This study not only offers a promising path to improve the audio module quality without increasing the size, but also represents an important milestone toward using acoustic metamaterial research to solve audio industry challenges.


Introduction
Advancements in audio modules have led to the development of successful consumer electronics products such as loudspeakers, cellphones, headphones, and smart speakers.More recently, as a crucial component of head-mounted devices (HMDs), audio DOI: 10.1002/admt.202300834module quality plays a significant role in augmented reality (AR) or virtual reality (VR) experiences.These products revolutionize how humans make immersive interaction with surroundings [1] in several domains such as entertainment, health care, [2] and education. [3]Audio radiation, which is commonly used in conjunction with visual displays and other sensors, is one of the major interfaces that humans use to communicate with HMDs.However, the radiation performance of these devices has reached a plateau in recent years.Improving loudness or power efficiency without increasing the device's size remains to be extremely challenging.This is due to the fact that the power of a monopole source radiated to the far field is proportional to (D/) 2 (D is the diameter and  is the wavelength).With the trend of miniaturization for HMDs, the power budget is very limited (a small battery), and the audio radiation module is confined to a small space where low-frequency radiation is even more challenging.There is an increasing need for a compact audio module that can produce a high sound pressure level while consuming low power, particularly at low frequencies ranging from 20 Hz to 1 kHz.It has been shown [4] that the bass frequencies (below 100 Hz) contributed to the largest increase of loudness in popular music recordings over time.Additionally, the mid band (octave band centered ≈500 Hz) is a major part of the long-term average spectrum of speech, [5] which pertains to speech intelligibility and various audiological applications.We therefore chose the bass and mid band frequency as the main target frequency band for the audio enhancement.
Typical audio module designs are electrical-mechanicalacoustical coupled systems consisting of transducers, front and back acoustical cavities, and surrounding environments.Examples of classical loudspeaker designs include closed boxes, [6] vented boxes, [7] horns, [8] and transmission line. [9,10]Recently, some folded transmission line designs [11,12] have been used for improving the radiation performance of small audio devices.These concepts are interesting, and the acoustic domain and transducer domain has been connected using equivalent circuit method.However, in previous studies, the acoustic structural resonance and transducer resonance were nearly independent, leading to an intrinsically weak coupling between the two resonances and thus suboptimal designs especially at low frequencies.
Acoustic metamaterials [13] are artificial structures that are capable of controlling the propagation of sound in unconventional ways, and have led to interesting phenomena such as acoustic cloaking, [14][15][16] acoustic subwavelength imaging, [17][18][19] acoustic rainbow trapping, [20,21] and acoustic unidirectional transparency. [22,23]However, most acoustic metamaterials are designed in a way that makes it challenging for them to have a direct and immediate impact on real-world applications pertaining to the audio industry, with one possible exception being sound absorption. [24]On the one hand, despite the reduction of audio distortion [25] realized by the mitigation of high-frequency acoustical cavity-resonances through the addition of acoustic metamaterial-based absorbers in loudspeaker enclosures, the attempt to incorporate passive acoustic metamaterials [26][27][28][29][30][31][32][33] into the loudspeaker designs has seen very limited radiation enhancement at low frequencies due to the bulky size and the lack of consideration on the metamaterial-transducer coupling.On the other hand, though active acoustic metamaterial may offer a solution to a broadband radiation enhancement, [34] they often involve designs that are complicated to avoid instability and also require more power consumption.
This study proposes a metamaterial-transducer resonant coupling, by carefully designing coiled acoustic metamaterials (CAMM) inside a head-mounted acoustic module, to enhance the sound radiation performance (both loudness and power efficiency) without changing the transducer or the overall exterior enclosure dimensions, which is generally a challenging requirement for product development in the audio industry.In contrast to previous studies, the resonant coupling effect between the transducer and metamaterial resonances is utilized as a critical tool to achieve the improvement of voltage sensitivity (dB SPL re 0.1 Vrms input excitation).We numerically show that, though the metamaterial resonance frequency is ≈ 600 Hz, a properly tuned metamaterial-transducer resonant coupling can lead to sound enhancement at a much lower frequency (≈100 Hz) close to the transducer resonance, which is the key to achieving broadband sound enhancement.We prototyped the design using 3D printing and validated the simulation results using free field measurements, where good agreements between the measured and simulated voltage sensitivity, excursion (the displacement of the diaphragm), electrical impedance, and power sensitivity (SPL re 0.1 W) are achieved.Finally, we recorded the audio output by mounting the audio module with CAMM on a head and torse simulator (HATS).The results showed 3 dB improvements at 100 Hz and 2.5 dB RMS improvements when using a male speech as input, and 2.6 dB improvements at 117 Hz and 0.8 dB RMS improvements when playing a music.It is important to note that although these improvements may seem moderate, they are actually significant by industrial standards.These results collectively demonstrate the substantial improvement of the sound radiation from the transducer to acoustic domain enabled by acoustic metamaterials.

Schematic of a Dipole Audio Module
A dipole source is typically employed to maximize audio output at the user's ears located at the near field and reduce the far field audio signal leakage (for privacy concerns) towards people around the user of HMD. [35]Figure 1a shows that a representative dipole audio module is located at the 'arm' of a commercial HMD (i.e., a VR device Quest 2). Figure 1b depicts a 3D schematic of the audio module, comprising a flat waveguide and a cylindrical shell where a customized and well-characterized moving-coil transducer (See Experimental Section for details) is mounted.The transducer is located further away from the ear  due to industrial design constraint, and the flat waveguide is used to direct sound from the transducer to a sound port (which is defined as the front port) as close as possible to the ear location.The waveguide is normally empty and thus provides a natural harbor for the installation of acoustic metamaterials.The sound port, which is situated at the left-hand side of the waveguide as shown in Figure 1b, emits the sound that originates from the bottom of the transducer and is referred to as the rear port.While one end of the shell is connected to the waveguide, the other end is connected to the main body (e.g., the optical display) of the HMD, as shown in Figure 1a.The cylindrical shell has a thickness of T cyl = 11.8 mm, an inner radius of R in = 10.43 mmn and an outer radius of R out = 13 mm.The thickness of the waveguide is T arm = T + 2t, where T is the inner thickness and t is the wall thickness 0.8 mm.The cross-sectional view of the waveguide is presented in Figure 1c, where the mechanical outline is a rectangle (L arm × W arm = L arm × (W + 2t) = 72 mm × (24.6 mm + 2×0.8 mm)) with two semi-circles (R arm = R + t = 12.3 mm + 0.8 mm).The front sound port (L p × W p = 9.5 mm ×3.5 mm) is covered by a thin layer of acoustic mesh with an impedance of 32 Ralys.A particular point of interest throughout this study is located at a distance of 5 cm and an angle of 45 degrees from the front sound port, on the same plane (z = 0), where a microphone is placed.This location is defined as the listening position (LP) where the radiation performance is targeted to be enhanced.The radiation from the front and rear sound ports behaves approximately as a dipole, as evidenced by the wave field illustrated in Figure 1d.

Simulations
We begin with an audio module with an empty waveguide at a thickness of T = 5.0 mm and conduct numerical simulations utilizing COMSOL Multiphysics 6.0, a finite element analysis (FEA) software.The simulation utilizes an equivalent circuit method (ECM) [36] to simulate the electrical-mechanical domain of the electromagnetic transducer and calculate the speaker diaphragm displacement as an input to the thermoacoustic domain of the front and back cavities as well as the surrounding air in the far field, in order to compute the radiated sound pressure.Figure 2a illustrates the ECM model for the audio module.The electrical domain is modeled using an inductor L e and resistor R e , while the mechanical domain is represented by the mechanical resistance R m , moving mass M m , and mechanical capacitor C m .A gyrator with a force factor of BL couples the electrical and mechanical domains.Notably, R m , C m , and BL, which are typically nonlinear, [36] are considered constants (linear approximation) for this study as shown in Table 1 (See Numerical and Experimental Methods for details), as a small voltage input signal 0.1 Vrms is utilized to ensure linearity.The coupling effect of the The waveguide resonance is f w = 900 Hz, which is a quarter wavelength resonance approximated by f w = c 0 /(4×L eff ), [12] where L eff is the effective acoustic propagation length from the diaphragm to the front sound port.Below the transducer resonance, voltage sensitivity at LP rolls off 12 dB per octave because the radiated pressure is proportional to the square of frequency (see Section S1, Supporting Information for details).Between the transducer resonance and the waveguide resonance, voltage sensitivity is normally independent of frequency for an empty waveguide (see Section S1, Supporting Information for details).
With the introduction of the CAMM inside the flat waveguide, the resonant frequency of the waveguide gets closer to that of the transducer and the resonant coupling between them must be taken into consideration.Under this scenario, a delicate balance is necessary to resolve the competition between needing a lower waveguide resonance frequency and a smaller added mass on the transducer resonance, both of which can lead to improved sound radiation at lower frequencies.The CAMM is illustrated in Figure 1b, which introduces a coiled channel in the waveguide characterized by a partition distance d, to effectively increase the physical length of wave propagation path without changing the waveguide volume.A parametric study of different partition distances d = 4, 8, and 12 mm is utilized to demonstrate the coupling effect, with SPL shown in Figure 2b and the differences between CAMM designs and the benchmark shown in Figure 2c.A lower waveguide resonance frequency can be obtained as the partition distance d becomes smaller, since L eff increases due to the smaller partition distance under a fixed waveguide space.However, voltage sensitivity around the transducer resonance may not benefit from CAMM with a smaller d.For a small d of 4 mm, Figure 2b indicates that the sensitivity drops above the transducer resonance, even though the sensitivity is improved below the transducer resonance.Treated as a single Degree-of-Freedom forced vibration problem, the frequency region right above the transducer resonance is a mass-controlled region (see Section S1, Supporting Information for details).In this region, as the added mass from CAMM (approximated by M a =  0 l eff dT A dT ) increases because of a decrease in d, the excursion and SPL frequency response decrease.It was found that an optimal partition distance of 8 mm achieves a continuous and broadband (20 Hz -700 Hz) enhancement of the voltage sensitivity, with peaks of 2.6 dB at 100 Hz and 7.1 dB at 600 Hz.It is important to note that unlike the conventional vented box design [7] which normally introduces a resonance lower than the transducer resonance to extend the bass response, resulting in a bulky size, our CAMM design introduces a resonance above the transducer resonance.
A dipole model is used to further understand how metamaterial-transducer resonant coupling can boost radiation in a continuous and broadband manner, especially at a lower frequency (the transducer resonance) by employing a structure with a higher resonance frequency (CAMM resonance).Figure 3a-c  Figure 3d shows the pressure phase difference between the front port and the rear port (referred to as phase difference throughout the paper for simplification).It can be observed that SPL at LP is highly correlated (similar shape for the SPL curves) to the front port since the LP is closer to the front port than the rear port.It can also be observed that the phase difference turns from 180 degrees to 0 degree when the working frequency goes from a frequency below the transducer resonance to the waveguide resonance, which induces a smaller cancellation between two sound ports and consequently a larger SPL at LP. Based on these two observations, we will show that SPL at the front port and phase difference between the two ports are two major contributing factors to the voltage sensitivity enhancement at LP near both the transducer resonance and waveguide resonance.Let us first focus on the voltage sensitivity near the transducer resonance.On the one hand, the properly designed CAMM has a 1.4 dB improvement on the front port SPL compared with the benchmark (from 81.8 to 83.2 dB) at 100 Hz, which is one factor contributing to the overall enhancement of 2.6 dB (from 52.1 to 54.7 dB) at the LP at 100 Hz.On the other hand, the properly designed CAMM introduces a slightly different phase difference compared with the benchmark (from 176˚to 175.5˚) at 100 Hz, resulting in a smaller dipole cancellation, which is the other factor contributing to the enhanced SPL at the LP.Note that instead of directly employing a structure resonating at 100 Hz, such a radiation enhancement of 2.6 dB at 100 Hz, which is due to the metamaterial-transducer resonant coupling from a 600 Hz metamaterial resonator, has never been reported and can prove valuable for the audio industry.Next, let us focus on the voltage sensitivity near the waveguide resonance.For the properly designed CAMM, as the magnitude of the front port response reaches a peak and the phase difference becomes 0 degree, the audio module is transformed from a quasi-dipole to an efficient monopole radiator, as confirmed by the peak of SPL at LP.If the metamaterial-transducer resonant coupling is improperly designed such as CAMM50_4mm, there is a significant drop of SPL at the LP between the transducer resonant frequency and metamaterial resonant frequency (100 -600 Hz), which is primarily caused by the drop of SPL from the front port.Based on the analysis of a single Degree-of-Freedom forced vibration (see Section S1, Supporting Information for details), acoustic pressure radiated from the transducer right above the transducer resonant frequency is inversely proportional to the added mass from CAMM (approximated by M a =  0 l eff dT A dT ).Thus, the excessive added mass from CAMM50_4mm on the transducer resonance is responsible for the drop of SPL at the LP for the improperly coupled resonance design.

Measured Radiation Enhancement induced by CAMM in Free Field
In this section, we use comprehensive analyses and experiments to demonstrate the coupled resonances that complement the acoustic domain analysis presented in the previous sections.Power sensitivity is another important metric for audio modules, particularly for power-constrained devices.Figure 4e,f demonstrate the real part of electrical impedance and power sensitivity in both simulations and experiments, respectively.Power sensitivity can be derived as SPL re 0.1W = SPL re 0.1V + 10log 10 (Re), where Re is the real part of the electrical impedance.By adding the CAMM into the waveguide, two positive effects are imposed on the real part of the impedance as shown in Figure 4e: the overall impedance profile shifts to a lower frequency, and a larger real part of impedance is induced at the waveguide resonance.Because of the redshift of the real part of impedance and the enhancement of voltage sensitivity ≈100 Hz shown in Figure 4c, power sensitivity of CAMM50 also achieves a broadband (50 -800 Hz) enhancement (except a few dB loss near 200 Hz) compared to WG50, with peaks of 4.2 dB (4.2 dB) at 100 Hz and 10.1 dB (9.4 dB) at 600 Hz (560 Hz) in simulations (experiments).

Measured Radiation Enhancement induced by CAMM on HATS
In addition to measuring audio output in free field, we have also conducted experiments in which we mounted the WG50 and CAMM50 modules in a pair of glasses to simulate an HMD, as shown in Figure 5b.Then we mount the glasses on a HATS and measure audio output from the eardrum reference point (DRP) using in-ear microphones, as shown in Figure 5a.We play both male voice signals and music signals to the WG50 and CAMM50 prototypes separately using 0.1 Vrms flat excitation without equalization (EQ).For the first example, the standard recording of Harvard Sentences according to the IEEE standard [37,38] from a male speech is played, which represents discrete and low-frequency sound.As shown in Figure 5c, the measured digital signals (with 0 dB full scale (dBFS) representing the maximum value of 1) show that our CAMM50 achieves a continuous and broadband (80-700 Hz) enhancement of voltage (except <1 dB loss ≈200 Hz) compared to WG50, with improvement peaks of 3 dB at 100 Hz and 6.9 dB at 570 Hz, and 2.5 dB RMS improvement in the time domain.As shown in Figure 5d,e, the spectrograms of WG50 and CAMM50 from 50 to 2000 Hz demonstrate the radiation enhancement as the signal varies in time.More signals marked in deep red indicate an enhancement for the frequency bands of 400 to 700 Hz and 1100 to 1400 Hz.A close-up view of the lower frequency range (50 -200 Hz) is plotted for WG50 and CAMM50, with more signals marked in green and yellow indicating the enhancement.For the second example, a music is played and Figure 5f shows that our CAMM50 again achieves a continuous and broadband (50 -700 Hz) enhancement of the voltage sensitivity (except ≈1 dB loss ≈200 Hz) compared with WG50, with improvement peaks of 2.6 dB at 117 Hz and 6.7 dB at 586 Hz, and 0.8 dB RMS improvement in the time domain.Similarly, the spectrograms of WG50 and CAMM50 in Figure 5g,h suggest a radiation enhancement in the frequency bands of 60 -200 Hz, 300 -700 Hz, and 1200 -1800 Hz, with more signal points marked in red and yellow indicating enhancement.The measurement of total harmonics distortion shows that CAMM does not introduce a strong distortion to the signal (see Section S2, Supporting Information for details).The slightly lower improvement in the music sound compared to the male voice is due to the concentration of higher frequencies in the music, which are less enhanced by the current audio module design.The enhanced radiation measured on HATS confirms that the CAMM offers a promising solution for HMD improvement and could potentially have an immediate impact on the audio industry.

Conclusion
In conclusion, we have proposed a coiled acoustic metamaterial to enhance broadband voltage sensitivity and power sensitivity in a head mounted audio module.The metamaterial design is easily manufacturable and more importantly, it doesn't consume additional space.Through a careful tuning of the coupling between the transducer resonance and metamaterial resonance, CAMM introduces a proper added mass on the front port and a smaller phase difference between the front port and the rear port.A combination of these factors leads to a 3-7 dB radiation enhancement in a broad frequency range from 50 to 700 Hz.Our findings constitute a major step toward using acoustic metamaterial research to solve practical audio industry challenges, which will benefit consumer electronics and AR/VR technologies.

Experimental Section
Numerical Simulations: Three Dimensional numerical simulations were carried out by the commercial finite element software COMSOL Multiphysics v6.0.The electrical and mechanical domains were modeled in Electrical Circuit module using TS parameters measured by Klippel R&D as shown in Table 1.The acoustic domain was modeled in Pressure Acoustics and Thermoviscous Acoustics to accurately account for the thermoviscous boundary loss.The coupling between the electrical domain and the mechanical domain was realized by 'Current-Controlled Voltage Source' feature under Electrical Circuit, while the coupling between the mechanical domain and the acoustic domain was realized by 'Interior Lumped Speaker Boundary' feature under Pressure Acoustics.In the lowfrequency regime of interest, the movement of the transducer diaphragm was approximated as a rigid piston.The frameworks of the acoustic module were modeled as acoustically rigid.The background medium was air with mass density  0 = 1.21 kg m −3 and sound velocity c 0 = 343.2m s −1 .Perfectly matched layers were imposed on the outer boundaries to eliminate sound reflections.The thin layer of acoustic mesh 32 Ralys was modeled using 'Interior Normal Impedance' feature under Thermoviscous Acoustics module, with the normal impedance set as 32 Pa s m, slip tangential velocity, and isothermal thermal condition.To ensure the accuracy of simulation, a mesh independent test was carried out.Numerical solutions converge, when the largest element size was set as 1/6 of the smallest working wavelength, the number of boundary layer was set as 6, and the thickness of first layer was set as 1/5 of the viscous boundary layer thickness.
Fabrication and Assembly of the Audio Module: The CAMM sample was fabricated using Clear Resin through Form 3+ printer, which utilizes Low Force Stereolithography (LFS) and provides a nominal precision of 25 μm.Loctite vinyl plastic and fabric adhesive was used to assemble the transducer to the 3D printed sample.To ensure a full and even seal, the adhesive was applied carefully using a syringe and needle under a microscope.
Measurement of SPL in Free Field: The speaker assembly was placed at the center of the room.A B&K 4190 microphone was placed 5 cm and 45 degrees away from the assembly, per the setup diagram.Based on Listen's standard loudspeaker measurement, a signal of stepped-sine sweep was generated, with the exciting voltage set as 0.1 Vrms.The room used for testing complies with ETSI standard TS 103 224 V1.5.1.For reverberation, the ETSI calls for a Clarity 80 average to be larger than 20dB.And C80 of the room was measured as 23.5 dB.The low reverberation level and the short measurement distance of 5 cm were deemed acceptable for this measurement.A time-gated measurement would have been employed if the measurement distance were much greater, such as 50 cm.
Measurement of Thiele-Small (TS) Parameters, Excursion, and Electrical Impedance: The TS parameters (or the small-signal parameters) of the transducer were measured using the Linear Parameter Measurement (LPM) module of the KLIPPEL Analyzer System by a small excitation signal of 0.1 Vrms.The excursion was measured using Klippel DA-2 analyzer and LK-H052 laser sensor.The speaker was contained in a transparent CAMM sample, which was held in place with the Klippel Microspeaker Clamping Jig.The laser was pointed through the CAMM sample and fixed on the center of the transducer diaphragm.The impedance was measured using the current sense and voltage monitoring capabilities of the DA-2 analyzer.
Measurement of SPL on HATS: The pair of glasses was mounted on a 5128C Head and Torse Simulator (HATS), and the audio module prototypes were positioned 5 cm and 45 degrees away from the ears per the diagram in Figure 5b.The room used for testing complies with ETSI standard TS 103 224 V1.5.1.Signals were measured at a 48kHz sample rate and 24-bit fixed bit depth.A 20 dB gain was added through 1704 2-channel CCLD amplifier for both male voice and music signals.A 30 dB gain was added to music recordings only to make it hearable.The measured digital signals, using Reaper DAW software, were plotted in Figure 5c,f with a size of 4096 and a Hann window.To convert the wav file sample values to pressure, a 95.2 dBSPL (1.151 Pa RMS/1.628Pa) 1kHz reference signal was generated by the sound calibrator type 4231 with DI-0658 adaptor to create a calibrated wave file recording (with 20dB mic amp gain and no other gain added to the recordings).For the reference signal, the average peak sample value was 1333.602, from which the conversion factor can be calculated as 1333.602/1.628= 0.00122 Pa/wav sample unit.The music recording has an additional 30 dB gain, which turns the conversion factor to 3.860e-5 Pa/wav sample unit.After converting the digital signal to pressure, power spectrum density was made in Figure 5d,e,g,h with a size of 2ˆ14, 50% overlap and a Hann window.

Figure 1 .
Figure 1.a) A dipole audio module on a HMD (a VR device called Quest 2).b) 3D, c) 2D schematics of the dipole audio module with CAMM, d) The normalized pressure distribution of the dipole configuration outside the audio module on the plane z = 0.

Figure 2 .
Figure 2. a) Block diagram of the audio module, only the front port with CAMM is included.b) Numerical voltage sensitivity of the audio module, when coupled resonances, which are induced by CAMM at different partition distances d, are introduced into an empty waveguide with a thickness of T = 5.0 mm.c) The improvement of voltage sensitivity compared with the empty waveguide at T = 5.0 mm.
show the near field frequency response in the acoustic domain at different locations (at a distance of 0.5 cm from the front port, at a distance of 0.5 cm from the rear port and at LP) for three different designs.These include the benchmark design (T = 5.0 mm without CAMM, denoted as WG50), improperly coupled resonances (T = 5.0 mm and d = 4 mm, denoted as CAMM50_4mm), and properly coupled resonances (T = 5.0 mm and d = 8 mm, denoted as CAMM50).

Figure 3 .
Figure 3. Numerical results of the dipole response for WG50 (black), CAMM50_4mm (red), and CAMM50 (green).The voltage sensitivity at three locations: a) at a distance of 0.5 cm from the front port, b) at a distance of 0.5 cm from the rear port and, c) LP. d) shows the pressure phase difference between the front port and rear port.

Figure 4 .
Figure 4. Photograph of measuring a) voltage sensitivity and b) excursion and electrical impedance of the audio module in free field under 0.1 Vrms.Measurements (dash lines) and simulations (solid lines) of c) voltage sensitivity at LP, d) excursion at the center of the speaker diaphragm, e) the real part of the electrical impedance, and f) power sensitivity per 0.1 W at LP for WG50 (black) and CAMM50 (green).

Figure
Figure 4a,b depict the experimental setups for measuring voltage sensitivity and excursion in free field (See Numerical and Experimental Methods for details).Figure 4c,d illustrate the comparisons between simulations and measurements of voltage sensitivity and excursion, respectively.The measurement is in good Figure 4a,b depict the experimental setups for measuring voltage sensitivity and excursion in free field (See Numerical and Experimental Methods for details).Figure 4c,d illustrate the comparisons between simulations and measurements of voltage sensitivity and excursion, respectively.The measurement is in good agreement with the simulation.In Figure 4c, CAMM50 achieves a broadband (50 -700 Hz) enhancement of output SPL with constant voltage excitation compared to WG50, with peaks of 2.6 dB (3.1 dB) at 100 Hz (95 Hz) and 7.1 dB (7.1 dB) at 600 Hz (500 Hz) in simulations (experiments).The full spectrum

Figure 5 .
Figure 5. a) Schematic of measuring the voltage sensitivity on HATS when b) WG50/CAMM50 is mounted on the left/right ear under 0.1 Vrms excitation voltage.Measured digital signals for the voltage sensitivity using in-ear microphone when c) male voice signals and f) a music signal are played by WG50 (black) and CAMM50 (green).Improvement of the voltage sensitivity compared with WG50 for the male voice signals and the music signal are plotted in the figures right below the voltage sensitivity.For the male voice signals, the spectrogram of WG50 and CAMM50 are shown in d) and e) from 50 to 2000 Hz using the same colormap (−140 to −97 dB).A close-up view with another colormap range (−130 to −107 dB) is plotted for WG50 and CAMM50 right below to show the enhancement from 50 to 200 Hz.For the music signal, the spectrogram of WG50 and CAMM50 are shown in g) and h) from 50 to 2000 Hz using the same colormap (−170 to −125 dB).A close-up view with another colormap range (−150 to −125 dB) is plotted for WG50 and CAMM50 right below to show the enhancement from 50 to 200 Hz.

Table 1 .
TS parameters of the moving-coil transducer from Klippel LPM measurements.
e /p ref ), p ref = 2 × 10 −5 Pa) at LP is marked in black in Figure 2b as a benchmark.SPL is substantially affected by two resonances at low frequency range of interest, i.e., transducer resonance and waveguide resonance.The transducer resonance is f s = 195.7 Hz, which can be calculated by (2 × sqrt (C m