Mundorf AMT21CM2.1-C
( Review )
Mundorf AMT21CM2.1-C
( Review )
The Mundorf AMT21CM2.1-C is an AMT-Type (Air Motion Transformer) tweeter manufactured by Mundorf, a German high-end audio manufacturer. It employs a unique accordion-shaped pleated diaphragm structure and is characterized by excellent transient performance, high resolution, and smooth, natural high-frequency reproduction.
It is widely recognized as one of the representative high-performance models among AMT-Type tweeters and has been adopted in many high-end loudspeaker systems.
Here, based on measurement results obtained using REW, the Impulse Response, ETC (Energy Time Curve), and STEP Response are analyzed, and a review is provided, including listening impressions, while comparing its characteristics with those of the VCD-type VCD-DT63.
For a detailed explanation of the characteristics of the AMT (Air Motion Transformer) principle itself, please refer to the AMT (Air Motion Transformer) Type section.
■Impulse Response
The Impulse Response describes how a loudspeaker responds over time to an instantaneous input signal, showing the time variation of amplitude (voltage).
The ideal response exhibits a sharp single peak, followed by a rapid decay of subsequent vibrations.
If significant subsequent oscillations remain, this indicates the presence of unwanted reflections or resonances within the system. These appear as components delayed relative to the direct sound (“delayed sound”), and as their amount increases, they can lead to blurred image outlines and reduced clarity of the soundstage.
【Measurement Conditions 】
●Acoustic Measurement Software: REW (Room EQ Wizard)
●Analysis Items: Impulse Response
●Measurement Distance: 10 cm
●Bandwidth: 3 kHz – 96 kHz (Butterworth HPF, 2nd order ×2; no LPF applied)
●Sampling Frequency: 192 kHz
●Normalization: Peak Normalization
The technical characteristics of the Mundorf AMT21CM2.1-C and the VCD-type VCD-DT63 are analyzed based on their Impulse Response waveforms, shown in the upper plot (overall view up to 1.00 ms) and the lower plot (expanded view up to 200 µs).
In addition, the relationship between these characteristics and the resulting sound quality is discussed in conjunction with listening impressions.
For reference, the manufacturer's published data for the AMT21CM2.1-C are also included below, enabling direct comparison with the VCD-type graphs under equivalent conditions.
■ Main Peak (Impulse 1st Peak)
(Energy Concentration and High-Frequency Extension )
The main peak is an important indicator of initial transient reproduction.
Although both drivers exhibit very steep rise characteristics, the energy of the AMT21CM2.1-C is distributed over a slightly wider time interval, resulting in a relatively lower degree of concentration in the main response. In contrast, the VCD-DT63 releases its energy within a shorter period of time, indicating a higher degree of response concentration in the time domain.
This difference is particularly evident in the expanded view shown in the lower graph (approximately 0–40 µs).
At the first positive peak (approximately 28 µs), the VCD-DT63 exhibits a significantly higher amplitude and a much steeper rise than the AMT21CM2.1-C. The subsequent transition to the first negative peak (approximately 52 µs) is also extremely rapid.
This demonstrates that the VCD diaphragm is exceptionally lightweight and that its effective high-frequency reproduction bandwidth (impulse-tracking capability) extends beyond that of the AMT21CM2.1-C. As a result, the driving force is transmitted directly to the diaphragm with minimal delay, enabling highly accurate tracking of the input signal.
[Impact on Listening Impressions]
For transient sounds such as the strike of a cymbal, the pluck of an acoustic guitar string, or the moment a violin bow contacts a string, the AMT21CM2.1-C delivers a smooth and easy-to-listen-to presentation. However, the edges of the sound are perceived as slightly softer.
In contrast, the VCD-DT63 exhibits excellent reproduction of the rise characteristics of low-level signals, producing exceptionally clear attacks with sharp leading edges and minimal smearing. As a result, the sense of speed and realism at the instant a sound is generated is perceived more vividly.
In addition, the contours of vocal consonants and instrumental attack components are reproduced with greater clarity, and the separation between individual sounds is perceived to be excellent.
■ 1st Valley(First Negative Peak)
Both drivers reach the first negative peak at nearly the same time, showing no significant difference in timing. Furthermore, both responses approach nearly −100%, demonstrating extremely fast transient characteristics and a high level of fundamental vibrational-system performance.
■ Second Peak (Approximately 70–100 µs)
The second peak is the region where the difference between the two drivers becomes most pronounced. It is significantly larger in the AMT21CM2.1-C, while it is noticeably smaller in the VCD-DT63.
This region represents the re-concentration of energy following the main response.
In the AMT21CM2.1-C, it indicates that a portion of the energy not released during the main peak reappears at a later time. In AMT drivers, the energy in this region tends to be larger due to multiple vibration modes associated with the folded diaphragm structure, as well as non-uniform in-plane velocity distribution.
In contrast, the VCD-DT63 exhibits a much smaller re-concentration of energy, indicating that the energy is released more completely and within a shorter time interval.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, the sound image tends to be slightly broader, and the edges of sounds are perceived as somewhat softer.
In contrast, the VCD-DT63 presents a more sharply focused sound image, with improved separation and definition between individual sounds.
■ 100–300 µs Region (Most Important)
This is the time region that has the strongest correlation with perceived sound quality.
In the AMT21CM2.1-C, even after the initial large waveform (0–100 µs) has decayed, small oscillatory residual vibrations continue to appear at approximately 100 µs, 200 µs, 300 µs, and beyond. In contrast, the VCD-DT63 exhibits much lower amplitude in this region and decays more rapidly, with the difference becoming increasingly apparent beyond approximately 200 µs.
This indicates that, in the AMT21CM2.1-C, vibrational components persist over multiple cycles, causing energy to be distributed over time. In contrast, the energy in the VCD-DT63 decays rapidly, demonstrating a much higher degree of energy concentration in the time domain.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, a somewhat richer sense of resonance and a smoother texture could be perceived, giving the impression that sounds were spread slightly over time. As a result, there were occasions when a slight sense of haze and additional sonic artifacts, not present in the original signal, seemed to be superimposed on the sound.
In contrast, the VCD-DT63 delivered a highly transparent soundstage, a remarkably quiet background, and exceptionally precise image localization. Because sounds decay naturally without leaving unnecessary residual energy, the sense of silence between notes is perceived as exceptionally deep.
In addition, reverberation and echoes contained in live or hall recordings could be heard fading smoothly into very low levels without being masked by unwanted resonances from the loudspeaker itself. As a result, the reproduced soundstage exhibits a high degree of resolution and transparency, allowing the listener to perceive not only spatial depth but also the subtle ambience and air of the recording venue.
■ 300–600 µs Region
While the AMT21CM2.1-C exhibits a noticeable re-emergence of energy around 400–550 µs, the VCD-DT63 has already reached a near-settled state by this point. The difference in this region is substantial and represents one of the most important factors in the overall evaluation of the Impulse Response.
The delayed energy components observed in the AMT21CM2.1-C are believed to result from the combined effects of multiple vibration modes associated with the folded diaphragm structure, in-plane vibrations, and reflections within the back chamber.
In contrast, the VCD design does not rely on high stiffness to forcibly suppress breakup modes. Instead, it employs a structure that inherently inhibits the propagation of unwanted vibrations, thereby reducing the temporal dispersion of vibrational energy itself. As a result, excess energy decays rapidly, producing exceptionally fast settling behavior.
The measurements further indicate that this excess energy is dissipated within an extremely short period of time, resulting in settling characteristics that may be considered among the highest levels currently achievable.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, the sound was perceived as having a rich sense of resonance and a longer decay, characteristics that sometimes manifested as the sparkling high-frequency quality and distinctive tonal character often associated with AMT drivers.
In contrast, the VCD-DT63 exhibited much faster energy settling, resulting in a cleaner and more organized presentation of musical information, with a more tightly focused sound image. In addition, because unwanted residual energy was minimized, coloration was perceived to be lower across the entire frequency range, giving the impression that the sound contained in the source material was reproduced more faithfully.
Furthermore, the high frequencies reproduced even the finest details with excellent clarity while remaining free from excessive sharpness or listening fatigue. As a result, the sound was perceived as smooth and natural, remaining comfortable to listen to even at high playback levels.
■ Overall Evaluation (Impulse Response)
The difference in design philosophy in the time domain is clearly reflected in the waveforms of the two drivers.
The Mundorf AMT21CM2.1-C exhibits excellent transient performance and high resolution while maintaining a smooth and refined character. At the same time, energy dispersion over multiple cycles and relatively large re-radiation components following the second peak can be observed. These characteristics are believed to contribute to its rich sense of air, natural resonance, and overall listening comfort.
In contrast, the VCD-DT63 demonstrates a high degree of energy concentration in the main peak, a significantly smaller second peak, rapid decay in the 100–300 µs region, and extremely low residual energy beyond 300 µs. As a result, it delivers exceptional transparency, precise image localization, excellent separation, and remarkably accurate transient reproduction.
If the Mundorf AMT21CM2.1-C can be regarded as one of the representative time-domain responses of today's high-performance tweeters, the VCD-DT63 may be viewed as representing a new generation of response characteristics, further pursuing energy concentration and rapid settling in the time domain.
Particularly beyond 100 µs, a substantial difference in settling behavior becomes evident, with the VCD-DT63 exhibiting an exceptionally short settling time. This difference is not only apparent in the measurements but was also clearly perceived during listening tests, manifesting itself as greater image clarity, enhanced spatial transparency, and superior reproduction of fine musical details.
■ETC(Energy Time Curve)
ETC is a metric derived from the Impulse Response that shows how the output sound energy is distributed and decays over time.
While the Impulse Response shows the time variation of amplitude (voltage), the ETC visualizes the decay behavior of the remaining acoustic energy in dB on a logarithmic scale. This makes differences in unwanted resonances and energy decay rates much more apparent.
These differences in ETC characteristics are perceived during listening as differences in the clarity of sound images and the transparency of the soundstage.
Differences in sound quality and soundstage reproduction are determined by the amount of components that arrive later than the direct sound (“delayed sound”). In the ETC, the energy components that appear after the direct sound are observed as sounds arriving with delay (“delayed sound”).
【Measurement Conditions 】
●Acoustic Measurement Software: REW (Room EQ Wizard)
●Analysis Items: Impulse Response / ETC (Energy Time Curve)
●Measurement Distance: 10 cm
●Bandwidth: 3 kHz – 96 kHz (Butterworth HPF, 2nd order ×2; no LPF applied)
●Sampling Frequency: 192 kHz
●Normalization: Peak Normalization
The technical characteristics of the ETC responses of the Mundorf AMT21CM2.1-C and the VCD-type VCD-DT63 are analyzed based on the waveforms shown in the upper graph (overall view up to 2.0 ms) and the lower graph (expanded view of the initial reflections and decay up to 250 µs).
In addition, the relationship between these characteristics and the resulting sound quality is examined in conjunction with listening test results.
In the ETC plots, the differences already visible in the Impulse Response become even more apparent as differences in time-domain energy behavior.
■ Initial Decay (0–200 µs)
This region is an important time interval that reflects the energy-settling characteristics immediately following the direct sound.
Although both drivers exhibit nearly identical direct-sound peaks (0 dB), the VCD-DT63 (red) remains at a consistently lower level throughout the 50–200 µs range.
In contrast, the Mundorf AMT21CM2.1-C (blue) retains energy in the −20 to −30 dB region for a relatively longer period, indicating that a portion of the main energy is distributed over time rather than being released instantaneously. As a result, its decay tendency is somewhat more gradual than that of the VCD-DT63.
■ 100–400 µs Region (Most Important)
This is the region where the difference between the two drivers becomes most apparent.
In the AMT21CM2.1-C, energy components of approximately −30 dB are still present around 300 µs, and a series of periodic peaks and valleys can be observed. Multiple energy re-emergence components are also evident, indicating that the energy is being released as multiple time-domain components distributed over time.
In contrast, the VCD-DT63 has already decayed to nearly −40 dB within the same time interval, demonstrating much faster energy settling. As a result, delayed energy components are significantly reduced, indicating a high degree of energy concentration in the time domain.
[Impact on Listening Impressions]
The decay characteristics in this region have a significant influence on soundstage transparency and the ability to perceive spatial details.
With the AMT21CM2.1-C, a rich sense of air and natural resonance can be perceived, giving the impression of a somewhat broader soundstage.
In contrast, the VCD-DT63 exhibits fewer unwanted delayed components, resulting in a clearer view into the recorded space, a quieter background, and a soundstage reproduction characterized by a higher degree of transparency.
■ ■ Intermittent Energy Components (Approximately 300–800 µs)
In the AMT21CM2.1-C, intermittent re-emergence peaks continue to appear even after the main energy release, and a relatively large peak can still be observed around 500 µs. This behavior may result from the combined effects of multiple vibration modes associated with the AMT folded-diaphragm structure, non-uniform in-plane velocity distribution, and reflections within the back chamber.
In contrast, these re-emergence components are substantially reduced in the VCD-DT63, indicating that the energy settles within a much shorter period of time.
[Impact on Listening Impressions]
As a result of these characteristics, the AMT21CM2.1-C conveys a rich sense of resonance, a longer decay, and a greater sense of body in the sound.
In contrast, the VCD-DT63 produces a more tightly focused sound image, with clearer source localization and a higher degree of separation between individual sounds.
■ Late Residual Energy (After 800 µs)
Late residual energy does not represent the direct sound itself, but rather energy components that appear with a time delay following the main response.
In the AMT21CM2.1-C, energy can be seen to remain distributed and released over an extended period of time. Periodic peaks and valleys continue to be observed in the 1.0–1.6 ms region, and measurable energy components remain present up to approximately 1.7 ms, indicating a prolonged decay process.
In contrast, the VCD-DT63 exhibits extremely low levels of late residual energy, with the response settling within a much shorter period of time. The energy approaches the noise floor at approximately 700–900 µs, and delayed components are substantially suppressed.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, the soundstage is perceived as spacious, with a rich sense of decay and reverberation, resulting in an overall smooth and natural presentation.
In contrast, the VCD-DT63 is characterized by a remarkably quiet background, highly precise image localization, and clearly defined sound images.
■ Overall Evaluation (ETC)
The Mundorf AMT21CM2.1-C exhibits excellent performance as a high-end AMT driver. However, its energy is released as multiple time-domain components, resulting in intermittent re-radiation throughout the decay process. In addition, noticeable late residual energy remains beyond 1 ms.
In contrast, the VCD-DT63 is characterized by a much faster initial decay, a significantly shorter ETC −40 dB decay time, and remarkably low residual energy beyond 300 µs. This behavior reflects the design philosophy of the VCD type, which does not rely on high stiffness to forcibly suppress vibrations. Instead, it minimizes the propagation of unwanted vibrations themselves, resulting in a structure that inherently reduces the reflection and accumulation of energy within the driver.
Human hearing perceives spatial width and depth by detecting very low-level reverberation and ambient information immediately following louder sounds. Therefore, when energy in the −30 to −45 dB range persists for an extended period, as observed in the AMT21CM2.1-C between approximately 0.5 and 1.5 ms, subtle reverberation cues and delicate low-level details contained in the original recording may be partially masked by the loudspeaker's own residual energy.
By contrast, the VCD-DT63 exhibits rapid energy decay beyond approximately 0.5 ms, preserving the sense of silence after sounds have ceased. As a result, fine reverberation cues, the quiet ambience of the recording venue, and subtle musical nuances can be reproduced without being obscured. This contributes to a soundstage presentation characterized by exceptional transparency, high resolution, and a clear sense of spatial depth.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, a rich sense of air, smoothness, and natural resonance could be perceived. At the same time, the numerous fine peaks and valleys visible in the ETC indicate that vibrational energy is being released as multiple time-domain components. These characteristics are believed to reflect the inherent sonic character of the driver, arising from its materials and structural design. Such behavior is often perceived positively as the delicate and sparkling sound commonly associated with AMT drivers. However, from the standpoint of fidelity to the original signal (Hi-Fi reproduction), it may also be regarded as a form of coloration.
In contrast, the VCD-DT63 releases its energy within a shorter period of time, while residual energy components are suppressed to extremely low levels. Furthermore, because there is very little evidence of intermittent energy storage or re-radiation, the driver exhibits minimal sonic character of its own, resulting in a highly neutral presentation. During listening tests, a high degree of transparency, excellent image localization, clearly defined sound images, and outstanding reproduction of low-level details were perceived. These impressions are also strongly supported by the ETC characteristics shown here.
■STEP Response
STEP Response represents how the output changes over time when an input signal rises instantaneously and is then maintained at a constant level.
In an ideal response, the waveform rises rapidly and then settles into a stable state without subsequent oscillation.
However, if unwanted reflections or delays exist within the system, oscillations and fluctuations appear in the output. These are observed as the effects of components that occur later than the direct sound (“delayed sound”).
【Measurement Conditions 】
●Acoustic Measurement Software: REW (Room EQ Wizard)
●Analysis Items: Impulse Response / Step Response
●Measurement Distance: 10 cm
●Bandwidth: 3 kHz – 96 kHz (Butterworth HPF, 2nd order ×2; no LPF applied)
●Sampling Frequency: 192 kHz
●Normalization: Peak Normalization
The technical characteristics of the STEP Responses of the Mundorf AMT21CM2.1-C and the VCD-type VCD-DT63 are analyzed based on the upper graph (overall view up to 2.0 ms) and the lower graph (expanded view of the initial rise up to 220 µs). Their relationship to perceived sound quality is then discussed in conjunction with listening test results.
For reference, the manufacturer's published data for the AMT21CM2.1-C are also included below, enabling direct comparison with the VCD-type graphs under equivalent conditions.
In the attached STEP Response plots, the differences observed in the Impulse Response and ETC appear as time-integrated behavior, clearly revealing the differences in the energy-settling characteristics of the two designs.
■ Initial Rise (0–50 µs)
Both drivers exhibit extremely steep rise characteristics, with no significant differences in the main peak amplitude or the peak arrival time.
In this region, both drivers demonstrate exceptionally good initial response performance. However, clear differences begin to emerge in their behavior immediately after the main peak.
[Impact on Listening Impressions]
Both drivers exhibit a very high level of high-frequency transient response and attack reproduction capability.
■1st Valley(First Negative Peak)(50~100µs)
The AMT21CM2.1-C drops to nearly −100%, forming a very deep valley. This deep valley in the STEP Response indicates a large reversal component following the main response, showing that the energy does not settle all at once and instead exhibits an oscillatory transient response.
In contrast, although the VCD-DT63 drops to approximately −50%, the reversal component is significantly suppressed, resulting in a more stable settling behavior.
When the waveform swings close to −100% immediately after the initial decay, energy cancellation can occur immediately after the attack. Perceptually, this may be heard as a thinner sound body or as a sound that is sharp in the high frequencies but somewhat lacking in substance.
On the other hand, when the reverse amplitude is suppressed to less than half of this level, the sharpness of the attack can be maintained while preserving the physical presence of instruments and vocals (their energy density and core body). As a result, the reproduced sound tends to be perceived as more natural and substantial.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, the edges of sounds tend to be slightly softer, and sound images may appear somewhat broader.
In contrast, the VCD-DT63 provides more clearly defined sound-image contours, better coherence after the attack, and a more tightly focused presentation.
■ Initial Residual Vibration (100–300 µs) (Particularly Important)
This is the region where the differences between the two drivers become most apparent, and it is also a particularly important time interval because of its strong correlation with perceived sound quality.
In the AMT21CM2.1-C, a large positive peak appears, and relatively large vibration components of approximately 20–25% continue to be observed even in the 250–350 µs range. This indicates a re-concentration of energy, and the re-radiation components previously observed in the Impulse Response and ETC are also clearly visible in this STEP Response.
In contrast, the VCD-DT63 exhibits significantly smaller amplitudes and less energy re-concentration, indicating that the response settles much more rapidly.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, the soundstage tends to be perceived as slightly wider, with richer resonance and a longer sense of decay, although sound images may appear somewhat broader.
In contrast, the VCD-DT63 provides clear image localization, a high degree of separation between individual sounds, and a more clearly defined sound body.
■ Settling Process (300–800 µs)
In the AMT21CM2.1-C, large oscillation cycles continue to be observed, and re-emergence components are still present around 700–800 µs. This indicates that energy is being intermittently re-radiated and distributed as multiple time-domain components. This behavior is consistent with the effects previously identified in the Impulse Response and ETC, including multiple vibration modes, in-plane vibrations, and reflections within the back chamber.
In contrast, the VCD-DT63 has already settled to nearly zero beyond 400 µs, confirming that delayed components are extremely well suppressed.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, rich resonance and a strong sense of air can be perceived, giving the impression of a somewhat broader soundstage.
In contrast, the VCD-DT63 provides a clearer view into the recorded space, a remarkably quiet background, and more precise source localization.
■ Late Residual Energy (After 800 µs)
This region does not represent the direct sound itself, but rather late residual energy components that appear with a time delay following the main response.
In the AMT21CM2.1-C, periodic oscillations continue up to approximately 1.6 ms, with residual vibration components of about ±5% still remaining. This indicates that energy is being distributed and released over an extended period of time along the time axis.
In contrast, the VCD-DT63 has already reached a nearly settled state by this point, and late residual energy components are suppressed to extremely low levels.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, a rich sense of decay, greater resonance, and a softer texture can be perceived.
In contrast, the VCD-DT63 provides a strong sense of silence, stable image localization, and clear spatial reproduction.
■ Overall Evaluation (STEP Response)
Although the Mundorf AMT21CM2.1-C exhibits exceptionally good initial response performance, it also shows a deep 1st Valley, large re-emergence components, long-lasting residual vibrations, and energy distributed over multiple oscillation cycles.
In contrast, the VCD-type VCD-DT63 is characterized by a high degree of energy concentration in the main peak, a shallower 1st Valley, smaller re-emergence components, rapid settling behavior, and extremely low levels of late residual energy.
It should be noted that when large oscillations are present in the 200 µs to 1.5 ms region, the unwanted decay components of the tweeter can overlap with those of the woofer or midrange when crossed over, potentially causing a loss of clarity around the crossover frequency region.
Furthermore, if these oscillations persist beyond 1.5 ms, the residual components of the previous sound can overlap with the onset of the next sound, slightly degrading the sense of tempo and rhythmic precision, and making the overall musical presentation sound somewhat heavier.
The most significant difference observed in this STEP Response is that, while the AMT21CM2.1-C releases its energy as multiple time-domain components distributed over time, the VCD-DT63 concentrates its energy into a much shorter time interval and then settles rapidly. This clearly reflects the different design philosophies of the two approaches in the time domain.
[Impact on Listening Impressions]
With the AMT21CM2.1-C, a rich sense of air, smoothness, and a spacious soundstage could be perceived.
In contrast, the VCD-DT63 exhibited a high degree of transparency, excellent image localization, clearly defined sound images, and outstanding reproduction of micro-transients.