The impulse response waveforms of three tweeters—Mundorf AMT21CM2.1-C, Dayton Audio AMT2-4, and the VCD-type VCD-DT63—are overlaid and displayed, and their respective characteristics are described below.
For reference, the impulse response waveform attached to the Mundorf AMT21CM2.1-C product documentation is shown in Fig. 1. For comparison, the characteristics of the VCD-DT63 are shown in Fig. 2 using the same scale as Fig. 1.

■Main Peak (Impulse 1st Peak)  
The sharpness of the initial rise is an important indicator that directly reflects the reproduction of initial transients.

The peak amplitude (energy concentration) is highest for the VCD-DT63, indicating that the energy is released in a more concentrated manner within a shorter time.

The sharpness of the rise is evaluated by the temporal slope up to the peak (dA/dt), with the VCD-DT63 showing the steepest rise. In comparison, the AMT21CM2.1-C exhibits a slightly broader temporal spread, and the AMT2-4 shows an even greater spread.

Characteristic of the AMT design: due to multiple vibration modes originating from the pleated diaphragm structure and non-uniform in-plane velocity distribution, the energy tends to be distributed over time, resulting in a broader peak.

■Impulse 1st Valley (First Negative Peak)  
　Rise Immediately After the 1st Valley(approx. 80–150 µs)
AMT21CM2.1-C shows 2–3 cycles of residual oscillation after the valley.

AMT2-4 exhibits larger residual oscillation with a longer period.

VCD-DT63 shows very rapid convergence after the valley, indicating the strongest damping.

Characteristic of the AMT type: In both AMT21CM2.1-C and AMT2-4, a relatively large positive peak appears again after the valley. This is considered to result from the folded diaphragm structure of the AMT driver, which operates by “pushing air out → returning.” Due to the combined effects of air loading and the elasticity of the diaphragm, the energy is not dissipated in a single event but is partially re-radiated.

■100–300 µs Region (Residual Oscillation)  
This is the region where performance differences between tweeters become most apparent; in high-end tweeters this region is significantly lower.

AMT21CM2.1-C shows clearly larger residual energy than VCD-DT63 in the 100–300 µs region, and additional peaks remain around 500–600 µs. In the “initial region that is particularly important in auditory perception” (100–300 µs), AMT21CM2.1-C performs better than the AMT2-4.

AMT2-4 shows the largest initial peaks and valleys, and the ripple in the 150–300 µs region is also the largest; however, the longer tail in the later region appears slightly shorter than that of the AMT21CM2.1-C.

VCD-DT63 shows that the major components have largely converged by approximately 230 µs.

Characteristic of the AMT type: Multiple periodic peaks and valleys continue to appear. This is considered to result from the folded structure of the AMT diaphragm, where several folds form the vibrating element. Each fold does not stop simultaneously, and local vibration modes are released with slight time dispersion. As a result, the impulse energy is not concentrated at a single point but appears distributed over several cycles.

In addition, the small peaks and valleys around 200–300 µs are thought to be partly caused by sound reflected within the back chamber, which acts again as pressure on the diaphragm and passes through the diaphragm to the front side.

■After 500 µs  
AMT21CM2.1-C shows a distinct secondary rise around 500–600 µs, indicating that residual energy continues into the later part of the response.

AMT2-4 exhibits the largest initial residual oscillation, but the behavior of the long tail after 500 µs differs from that of the AMT21CM2.1-C.

VCD-DT63 shows very rapid convergence after the valley, indicating the strongest damping.

Characteristic of the AMT type (delayed components): In AMT21CM2.1-C, a clear peak appears around 500–600 µs, while in AMT2-4 residual energy remains around 400–500 µs. This phenomenon is considered to result from delayed release of energy associated with in-plane vibration of the AMT diaphragm and the compression–expansion of the air cavity, and it is a feature relatively often observed in AMT designs. In piston-type tweeters, the energy in this time region is usually much smaller.

In addition, the peak observed in AMT21CM2.1-C around 500–600 µs is particularly distinct and difficult to explain solely by primary vibration. It is thought to include components in which sound reflected multiple times within the back chamber subsequently passes through the diaphragm and is radiated to the front. 

■Impulse Convergence Speed (Visual Estimate)  
AMT21CM2.1-C shows residual energy beyond 600 µs.    
AMT2-4 shows residual energy beyond 500 µs.  
VCD-DT63 shows convergence of the major components at approximately 230 µs.  

■Acoustic Significance  
These differences mainly affect the sharpness of image outlines, the transparency of the soundstage, and the clarity of cymbals and consonant sounds.

AMT21CM2.1-C is high quality but slightly soft in character.

AMT2-4 tends to sound somewhat smeared.

VCD-DT63 produces sharper image outlines and clearer spatial information.

■Overall Evaluation (Impulse Response)
Compared with the Mundorf AMT21CM2.1-C (a high-end AMT), the VCD-DT63 shows clearly faster impulse convergence, suggesting the possibility that it exceeds the limitations of conventional tweeters.

Characteristic of the AMT type: In both AMT21CM2.1-C and AMT2-4, the response does not drop sharply to zero after the first peak but instead decays gradually through several peaks. This is considered to result from the fact that the AMT diaphragm is not a single piston but is composed of multiple folds, causing the vibrational energy to be released over time rather than in a single instant. 

Similarly, the ETC (Energy Time Curve) responses of three tweeters—Mundorf AMT21CM2.1-C, Dayton Audio AMT2-4, and the VCD-type VCD-DT63—are overlaid, and the characteristics of each are described below.

■Initial Decay (0–200 µs)
In terms of coherence, the AMT21CM2.1-C outperforms the AMT2-4, while the AMT2-4 tends to exhibit longer persistence of initial energy.
VCD-DT63 decays steeply immediately after the peak and has already dropped to nearly −30 dB at around 200 µs. 

■Periodic Peaks and Valleys Appearing Around 100–400 µs

Characteristic of the AMT type: multiple peaks and valleys appear repeatedly. This is considered to occur because the AMT diaphragm is not a single piston but is composed of multiple folds, causing the vibration modes of each fold to be radiated with temporal dispersion. As a result, the energy does not concentrate at a single point in time but appears distributed across multiple time components. 

■Mid-Term Decay (200–500 µs)

This is one of the most important regions among the Impulse, ETC, and STEP responses. The decay characteristic in this region appears as the ETC −40 dB arrival time and has a significant influence on spatial transparency, openness, and the clarity of the soundstage.

Characteristic of the AMT type: in this region, the residual energy of the AMT type becomes clearly visible. AMT21CM2.1-C and AMT2-4 show multiple repeating peaks and valleys in this region, indicating that energy is being intermittently re-radiated. This is considered to result from the multiple vibration modes of the AMT folded diaphragm and internal reflections.
VCD-DT63 decays rapidly in this region, reaching −40 dB at approximately 420 µs and dropping to nearly −50 dB at around 500 µs.

 
■Intermittent Energy Re-Rise Around 300–800 µs

Characteristic of the AMT type: multiple re-rise peaks appear during the decay process. This is considered to result from in-plane vibrations of the folded diaphragm and sound-pressure fluctuations inside the back chamber acting again on the diaphragm, causing energy to be intermittently radiated. In piston-type diaphragms, this region typically exhibits a more monotonic decay.

■Late Region (0.7–2.2 ms)

Characteristic of the AMT type: in the AMT21CM2.1-C, relatively large peaks continue up to around 1.7 ms, while in the AMT2-4, periodic residual components persist up to approximately 2.2 ms. This is considered to result from internal vibrations of the diaphragm and reflections within the back chamber being radiated to the front with temporal delay, causing a longer energy decay. As a result, the energy appears distributed over multiple releases rather than decaying in a single event.

■Overall Evaluation

The direct-sound peak is comparable among the three, but the subsequent energy convergence speed differs clearly.

AMT21CM2.1-C shows relatively good initial decay, but residual peaks continue up to around 1.7 ms.

AMT2-4 exhibits the slowest initial decay, with periodic residual components continuing to approximately 2.2 ms.

VCD-DT63 shows the strongest temporal concentration of energy and the fastest overall decay.

Characteristic of the AMT type: in the AMT type, periodic peaks and valleys appear continuously in the decay curve. This is considered to result from the multiple vibration modes of the folded diaphragm structure, which distribute the energy over time. 

Furthermore, the STEP response waveforms of three tweeters—Mundorf AMT21CM2.1-C, Dayton Audio AMT2-4, and the VCD-type VCD-DT63—are overlaid and displayed, and their respective characteristics are described below.
For reference, the STEP response waveform attached to the Mundorf AMT21CM2.1-C product documentation is shown in Fig. 3. For comparison, the characteristics of VCD-DT63 are shown in Fig. 4, with the horizontal axis using the same scale as Fig. 3 and the vertical axis normalized to the peak value. 

■Rise (0–50 µs)
The peak amplitude is higher for VCD-DT63 and AMT21CM2.1-C, indicating that the energy is concentrated within a short time. AMT2-4 shows a lower peak.

■Depth of the 1st Valley (First Negative Peak)

AMT2-4 drops to nearly −100%, and AMT21CM2.1-C also exhibits a deep valley. This indicates that the transient response of the vibration system is somewhat oscillatory. In contrast, VCD shows a shallower valley, with transient oscillation well suppressed.
Characteristic of the AMT type: both AMT21CM2.1-C and AMT2-4 exhibit very deep valleys, reaching approximately −80% to −100%. This is considered to result from the folded diaphragm structure, where after pushing air outward, the air load and diaphragm elasticity produce a strong reverse velocity component, leading to the oscillatory transient behavior characteristic of AMT. 

■Initial Residual Oscillation (100–300 µs)

This region in the STEP response is particularly important among all responses, similar to the ETC −40 dB arrival time. The speed of recovery after overshoot has a significant influence on soundstage clarity and the sense of silence.

AMT2-4 exhibits large, sustained periodic oscillations, and AMT21CM2.1-C also shows relatively large residual vibrations. In contrast, VCD-DT63 has smaller amplitudes and shorter oscillation periods.
Characteristic of the AMT type: multiple peaks and valleys appear in succession. This is considered to result from the folded diaphragm not stopping as a single piston, but rather each fold ceasing vibration at slightly different times due to different vibration modes. As a result, periodic oscillations appear in the STEP response.

■Convergence Process (300–800 µs)

AMT21CM2.1-C and AMT2-4 still exhibit multiple oscillation cycles in this region. In particular, AMT21CM2.1-C shows a relatively large peak around 600 µs, indicating temporal re-release of energy. In contrast, VCD-DT63 shows significantly smaller amplitudes in this region and faster convergence.

Characteristic of the AMT type: relatively large peaks are observed around 600 µs in AMT21CM2.1-C and around 400–700 µs in AMT2-4. This is considered to result from in-plane vibrations of the diaphragm and sound-pressure fluctuations in the back chamber acting again on the diaphragm, causing delayed radiation of energy.  

■Residuals Beyond 800 µs

VCD-DT63 has nearly converged to zero in this region.

Characteristic of the AMT type: in both AMT21CM2.1-C and AMT2-4, periodic oscillations remain beyond approximately 1 ms. This is considered to result from the superposition of multiple vibration modes of the folded diaphragm structure and reflections within the internal cavity, causing the energy to be distributed over time.  

■Overall Evaluation

From the perspective of STEP response, VCD-DT63 shows the highest concentration of rise energy, with the smallest subsequent oscillations and the fastest convergence. AMT21CM2.1-C, as a high-performance AMT, exhibits relatively well-controlled behavior in the initial region, but a re-rise component remains around 600 µs. AMT2-4 shows the largest initial oscillations and tends to have the slowest convergence.

Characteristic of the AMT type: in the AMT type, vibrational energy does not dissipate at once but tends to be distributed across multiple time components and persist over a longer duration, which is also reflected in the STEP response.