Audio IEMs Reviews Universal-Fit

Gojira! Or, how the DUNU DN-2000J changed my mind about hybrid IEMs

Electroacoustic Measurements & Technical Analysis

Editor’s Note: Once again, we’re indebted to speakerphone’s @farbic amazing work. We spent a good number of hours discussing and troubleshooting measurement artifacts, in addition to discussing technical details. This effort, though, would not have been possible had it not been for DUNU‘s incredible support, providing us with two different copies of the DN-2000J to perform analyses on, as well as affording us absolute full reign over our test parameters. They want to gain as much constructive criticism from us as they can, whether it be with respect to sound signature tuning or quality control. This type of forthcoming attitude is to be applauded and set as an example to all manufacturers in the audio industry.

Frequency Response

Once again, all measurements follow an acquisition protocol as close to IEC 60268-7 as possible. While insert earphones are normally to be used at the reference plane of ear simulator couplers, the DN-2000J was placed at a “realistic” distance away from the reference plane — about 5.8 mm. This ensures a practical response of ear canal resonances for most ears, given the larger diameter housings of the DN-2000J. Please bear in mind that, as with all ear simulator measurements, these results are only a representation of what can be currently modeled with current methods.

Left and right channels are well-matched, almost always within ±0.5 dB of one another, and mostly inside of ±0.2 dB. Our first test unit exhibited a ~4 dB bass response mismatch at 20 Hz in the right channel, accompanied by increased overall harmonic distortion (mostly due to increased 4th order harmonic distortion), suggestive of a partial blockage of the pressure excursion vent, as channel matching above 200 Hz was near perfect (within ±0.2 dB). Through lengthy discussions, DUNU decided to send us a second test unit, which tested without any mismatch issues.

Raw, uncompensated frequency response of the DN-2000J, as measured by an IEC 603018-4 (formerly IEC 60711) compliant ear simulator.
Raw, uncompensated frequency response of the DN-2000J, as measured by an IEC 603018-4 (formerly IEC 60711) compliant ear simulator. Red trace is the right channel, blue is the left channel. Channel matching is found to be quite good in this unit (the second tested).

The first resonant peak has a locus of approximately 3.2 kHz. Traditionally, earphones have had a large, broad peak centered between 2.5-2.7 kHz, in order to compensate for ear canal resonance, but it is also well-known that the loci of peak canal resonance differ from person to person, deviating between 2.5-3.5 kHz. Thus, it should still be acceptable for most people. Depending on the style, depth of wear, and choice of ear tips, this first resonant peak may also shift slightly.

The DN-2000J takes on a distinctively U-shaped sound signature response — moderately boosted lows, centered between 80-125 Hz, and boosted highs. This is a common acoustic treatment for earphones that attempt to adjust for Fletcher-Munson volume thresholds, making the DN-2000J especially good for low volume listening. As noted in our review, the DN-2000J does sound especially remarkable at very low volumes.

This type of U-shaped tuning is especially apparent when target equalization is applied, either for Diffuse Field or the Harman International target by Drs. Sean Olive and Todd Welti. When the bass extending silicone ring is applied, the bass of the DN-2000J is nearly on par with the Olive-Welti target curve for flat equalization below 200 Hz.

This slideshow requires JavaScript.

The treble response, however, is contrary to perfectly linear, and may pose problems for users that prefer smooth highs. The main treble resonances are deemed to be too strong when normalized to a equalization target such as the Olive-Welti target curve.

However, bandwidth on the DN-2000J is quite good, reaching most targets end to end.

Effects of Differing Ear Tips on Perceived Response

Being the first DUNU product to actively advertise a strategic partnership with Comply Foam, the DN-2000J seems to have been designed for these tips, taking advantage of foam tips’ tendency to smooth out peaky treble. With Ts-500 foam tips, the DN-2000J is smoother across the board, de-wrinkling most issues people might have with the capricious nature of the TWFK tweeter response. It does sound a little more prosaic as a consequence, but for long-term listening, I wouldn’t hesitate to recommend a pack of Comply tips to any individuals wishing to use the DN-2000J for longer listening sessions.

The plethora of ear tips included yield slightly different measured responses in our measurement tests. While the silicone ear tips mostly yielded identical results (more later), the different Comply foams yielded some surprising results. Surprisingly, the T-500 tips managed to attenuate treble even more than the Tx-500. The circular Ts-500 tips attenuate treble the least. All the Comply tips seem to smooth out the response of the treble, lessening resonance peaks.

Zoomed view of the effect of various Comply foams on frequency response. Blue trace denotes Ts-500 response, red is Tx-500, and orange is T-500.
Zoomed view of the effect of various Comply foams on frequency response. Blue trace denotes Ts-500 response, red is Tx-500, and orange is T-500.

The final resonant peak is attenuated a lot more greatly than is the resonant peak at 6.5 kHz, correlating with the subjective experience that the Comply foam tips control harsh treble and sibilance. The resonant peak at 6.5 kHz also shifts downward in frequency, suggestive of an improvement in sibilance control while maintaining clarity.

The measurements also show that absolute bandwidth decreases with Comply foams, but the subjective experience a posteriori doesn’t corroborate these findings. The different silicone tips, made largely of the same material, hold the frequency response similarly.

However, there are signs that the tips have differing resonant frequencies, as shown by the bass artifacts that show up. Although the effect was not measured, I found that wide bore, bi-flange silicone tips also tend to help facilitate both insertion depth and resonant peak smoothness.

Effects of the Silicone Ring on Bass Response & Spacer Rings on Treble Response

The silicone bass rings included with the packet are intended to raise the bass extension of the DN-2000J, at the expense of driver flex (caused by the inability for the earphone to equalize pressure between the sealed interface of the ear canal and the rear chamber of the earphone).

While the rings clearly show improved bass response below 250 Hz, increasing approximately +4 dB at 100 Hz, +6 dB at 50 Hz, and +8 dB at 20 Hz, the subjective experience suggests a more sluggish sounding bass response compared to the response sans occluded vents.

Spacer rings are a carryover from both the DN-1000 and DN-2000 — originally intended as novel means of solving fit issues some people had with wider housings, they also modify perceived treble response on the principle that residual ear canal volume affects resonant quarter-wave frequencies in the treble region.

As measured, with the spacer rings lower the final resonant peak indicative of (approximate) treble extension by a minor amount, but seems to affect other frequencies in relatively minor ways. Perhaps the most pronounced effect would be that the overhang of the ear tips is increasingly greater, but ear tips’ effect on frequency response is not convincingly measurable given our current testing methodology.

A combined diagram showing the effect of the silicone bass extension ring (magenta trace), and the tip spacer rings. The orange trace shows the effect of the thinnest, silver-colored ring. The green trace shows the effect of the medium thickness, blue-colored ring. Lastly, the royal blue trace shows the effect of the thickest, red-colored spacer ring.
A combined diagram showing the effect of the silicone bass extension ring (magenta trace), and the tip spacer rings. The orange trace shows the effect of the thinnest, silver-colored ring. The green trace shows the effect of the medium thickness, blue-colored ring. Lastly, the royal blue trace shows the effect of the thickest, red-colored spacer ring.
Measurement Artifacts with Differing Ear Tips

When measured with translucent, white, frosted silicone tips (red trace), there is a distinct bass artifact centered at around 150 Hz, while the foam-less grey silicone tips (blue trace) exhibit a similar one closer to 200 Hz. The foam-filled (for isolation) grey silicone tips (green trace) do not exhibit any artifacting.

Silicone tips exhibit notch artifacts centered between 150-200 Hz.
Silicone tips exhibit notch artifacts centered between 150-200 Hz.

The figure below zooms in on the region in question:

The notch artifact is non-existent with both the foam-filled grey silicone tips and with Comply foam tips.
The notch artifact is non-existent with both the foam-filled grey silicone tips (green) and with Comply foam tips.

We have a possible explanation for these artifacts: sound waves are longitudinal waves, and are therefore subject to interaction with compliant materials such as silicone ear tips. The pulsatile response of sound waves will “bounce back” and create a distinctive pattern when earphones are measured inside an IEC711 coupler (and perhaps others).

The best analogy for this phenomenon would be in the human circulatory system. The aorta, the largest artery of the human body, is composed largely of elastic fibers that serve to provide the circulatory system a “reservoir” of sorts to buffer the flow of blood across all tissues of the body — a phenomenon known as the Windkessel Effect. Reflected blood flow will impress on the aortic semilunar valve and create a “dicrotic notch” very similar in pattern to what is shown here.

Thus, it’s reasonable to assume that the artifacts seen in our measurements are manifestations of destructive/constructive bounce-back due to varying elastic compliance of tips of different materials and shapes. It’s furthermore reasonable to surmise that silicone tips of different varieties (which a multitude of companies sell aftermarket) will affect users’ sonic experiences considerably.

We recommend any and all users to continually search for the best pair of tips for their personal ear size/shape, but also for tips of varying materials, as they may very well impact sound quality, both negatively and positively. At the same time, ear tips size and wear depth factors to be a much more significant variable. As we tested larger-sized ear tips that forced an increasingly shallow position in our ear simulator with respect to the reference plane, we noticed the expected downward shifts in quarter-wave resonances for treble frequencies.

The measured response of differing ear tip sizes and their effect on the measured distance from the reference plane. The black trace is the original response, measured at 5.8 mm, while the orange trace is at 6.5 mm, and lastly, the green trace is at 7.4 mm away (approximately the shallowest insertion possible, about the same as the entrance of the external acoustic meatus).
The measured response of differing ear tip sizes and their effect on the measured distance from the reference plane. The black trace is the original response, measured at 5.8 mm, while the orange trace is at 6.5 mm, and lastly, the green trace is at 7.4 mm away (approximately the shallowest insertion possible, about the same as the entrance of the external acoustic meatus).
Electrical Impedance & Source Matching

This slideshow requires JavaScript.

The DN-2000J is probably one of the lowest impedance earphones on the market currently in existence. Measuring at just over 5.5 Ω at the low end of the audio spectrum, it rises to just over 6 Ω at 1 kHz and to only about 11 Ω even at 20 kHz.

With such low load impedance, the DN-2000J will be ironically a dramatically challenging load for most headphone amplification sources, despite its high sensitivity. While it would take very little voltage/power to get the DN-2000J up to listenable volumes, expect the earphone to require great transient control, thus demanding a source that delivers adequate current swing.

We also suggest an amplifier of very low output impedance for the finest damping control over the bass driver, as the frequency response will drift by about +1.5-3 dB in the upper midrange and treble frequencies when 5-15 Ω of extra serial resistance is added to the output amplifier.

Luckily, the change won’t alter too much from there; even with an extra 400 Ω of serial resistance, the DN-2000J deviates by only a maximum of an extra 2-3 dB — no small amount, but nothing serious to write home about.

The effect of resistance added to the output amplifier at 5, 15, 30, and 400 Ω, simulating the effect of source output impedance on frequency response.
The effect of resistance added to the output amplifier at 5 (green), 15 (orange), 30 (teal), and 400 (pink) Ω, simulating the effect of source output impedance on frequency response.

Distortion

Perhaps because of the immensely low impedance (probably a little too low) voice coils of the dynamic and balanced armature drivers, the THD response of the DN-2000J isn’t as clean as we had hoped, but is still pretty good in the grand scheme of things, and is by far the least distorting hybrid design we’ve seen aside from the AKG K3003.

We’d be more concerned about THD variance from unit to unit, as the two we received and tested exhibited fairly different results, though both were never in the realm of danger for audible distortion, even under loud conditions. The subjective experience is certainly not hurt by these readings, as the DN-2000J is one of the most cleanly resolving universal-fit IEMs we’ve ever had the pleasure of experiencing.

This slideshow requires JavaScript.

 Cumulative Spectral Decay (CSD)

From the CSD plot, we see that the treble contains typical ringing of the TWFK driver. We’re not thrilled with it, but it’s expected from anything sporting that driver family.

The more important area to notice is the bass — nearly all of the bass decay is neatly confined to frequencies below 250 Hz, with the DN-2000J‘s gentle mid-bass hump decaying away fairly quickly and giving way to sub-bass decay. These results hint at what we’re hearing from the truly excellent bass characteristics of this earphone.

Cumulative Spectral Decay plot of the right channel response; the left channel exhibits no significant differences.
Cumulative Spectral Decay plot of the right channel response; the left channel exhibits no significant differences.

Concluding Thoughts

Like the 911 Turbo meeting Godzilla on the proving grounds of the Nürburgring, the AKG K3003 has met its East Asian counterpart in the DUNU DN-2000J. At an MSRP of $349 USD, the DN-2000J has made true flagship-level hybrid sound available to the public at way under $500.

At the time of its release, the $1299 AKG K3003 was widely considered “overpriced”, but even its harshest critics (like me) pegged it at least at a value of $799. That the DN-2000J comes very, very close to the performance of the K3003 in the midrange and treble, while noticeably surpassing it in the bass — for less than half the price, no less — is nothing short of astounding. DUNU should be very proud of what they’ve accomplished; their ability to create such advancement in performance at the $350 price point is nothing short of astounding. If you’re a “clarity head” or anyone who craves immense levels of resolution, the DN-2000J is basically a no-brainer purchase.

PROS
  • Flagship-level bass and overall sound quality for $349
  • Retains resolution and clarity even at very low volumes
  • Improved fit and well-rounded accessory kit
CONS
  • Brighter, analytical, resolution-driven sound signature won’t please everyone
  • Some may perceive treble to be harsh
  • Extremely low impedance, will be very picky with sources for hiss/hum/rub/THD/IMD

Special thanks goes to the incredible support of Mr. Hsiao and Mr. Wu of the DUNU team in their New Taipei City office; without their many hours of assistance and accommodation, we would not have been able to test the DN-2000J so early and so extensively, both subjectively and objectively.

Furthermore, we thank DUNU for their extremely forthcoming attitude and willingness to troubleshoot technical snafus, regardless of how small they were. Their investigative, assiduous spirit will only enable them to raise the bar on their future products ever higher.

For more information about the DUNU DN-2000J, please visit:
http://www.dunu-topsound.com/DN-2000J.html

About Mr. T

Mr. T is an in-ear fanatic by day, and writes SOAP notes by night. He pities the fool who actually has the patience to read through his stuff. (Full Author Bio)

17 comments

    1. Many people really like the A83, but my opinion (Mr. T) is less optimistic; while the A83 is warmer and gentler than the 2000J, it has a large suckout in the upper midrange that robs it of the detail it should have in the human hearing system’s most sensitive region. The 2000J is almost assuredly more detailed than the A83, but it does come at the price of brightness.

      Like

  1. One of the most complete review thanks 🙂

    I really love the Dunu DN-1000, it’s detailed, bright and V shapped,nice treble extension, lot of bass and rumble nicely but mid are recessed after listen my GR07 classic, I just want more mid and less emphasis on mid bass region with same soundstage or even better than DN-1000, did you consider the DN-2000J like an upgrade base on my taste ?

    Liked by 1 person

    1. The DN-2000J definitely seems to possess the improvements you’re after. It has a good balance between the mid- and sub-bass — the entire bass region is extremely high quality, on par with $1K+ flagships. There should also be more mid presence as well, though it is not “mid-forward”.

      Like

  2. This is the best review I have ever read in my life. And I have read ALOT of reviews on Headfonia, Headfonics, Headfi, Headphonelist etc.

    Very enjoyable, technical, informing, great vocabulary, charismatic and never a dull moment.

    Please just carry on doing what you’re doing!

    Thanks.

    Liked by 1 person

  3. Hi great review! Very detailed and descriptive. I’m strongly considering purchasing these IEMs but the main concern I have is its low impedance. Most consumer portable amps and dac-amps that I see are 16ohm and higher. I personally own an iPhone 6+ which I’ve read has an output impedance of around 3.5 so it wouldn’t be ideal for these IEMs but perhaps ok.

    My question is whether you could recommend a few portable dac-amps or amps to pair with this IEM to get the best sound. Also, given recent improvements to premium cell phone hardware, does a cheaper dac really serve a purpose in improving audio quality or clean amplification over the analog headphone out of iPhone 6+?

    Thanks!!

    Like

    1. It’s difficult to recommend any one device, and don’t worry too much about the impedance rating — that stuff is all very approximate and really doesn’t apply to an ultra-low impedance device like the DN-2000J; the major reason to upgrade over something like an iPhone 6 Plus would be to improve the residual noise performance, as well as crosstalk. The new iPhones are not as clean with residual noise as they were in the past, so you’ll get occasional crackling and some hiss-related issues when you insert your earphones or switch apps — issues that aren’t major, but can become annoying from time to time.

      The general guideline for purchasing a DAC/amp or standalone amplifier that would work well with the DN-2000J would be to go for something that uses a good current buffer so that a current-hungry earphone like the 2000J would be able to perform as well as possible. So, devices with opamps like the LME49600 are *probably* good choices. But that’s not the sole consideration — the amp must also, as mentioned before deliver that current quietly. Residual noise (these ratings are rarely ever published, but should show up in units on the order of microvolts) should be at a minimum, and DC offset (another rarely published spec) should be at a minimum as well. The power rating (measured in mW, or Voltage at full-scale as an approximation) is really not important at all, though in general, the more powerful the amp, the *less* suitable it would be for the 2000J because a powerful amp is less concerned with clean noise floor, than it is about driving intensely difficult to drive full-size headphones, such as planar magnetic ones.

      It’s not to say that you should be following these guidelines absolutely, however. Not all theory works perfectly in the real world, so it’s best to try out multiple pairings yourself, and if the pairing with the 6+ is adequate for you, then don’t worry about it. We’ve heard good things about pairings with devices like the FiiO X5-II and Cowon P1, however, and have had personal success even with a higher-powered, low-cost device like the FiiO E11K (now A3). The Oppo HA-2 has also been a decent pairing.

      Like

  4. Very nice review. Recently i just bought dunu titan 1, for the price the performance is really impress me, and its fit with my type of music wich is rock and heavy metal.

    I am planning to buy dunu dn2000j so my question is whether its fit with my sound preference? (i like the sound of grado).

    Thanks…

    Like

  5. I use the 2000j’s with my Sony ZX2, and they are fantastic, the j’s are a bit harder to drive than my custom JH16’S, but the ZX2 takes care of that, for the price the 2000j’s can’t be beat.

    Like

  6. Hey! After saw your mensurements and compare between the innerfidelity I’ve checked the D2K has low distortion below <0.1 and 2000j's above 1.0%, this distortion can be because the low impedance? Or on papel does it not matter much?

    Like

    1. InnerFidelity’s measurements are a bit iffy for the DN-2000 —- if you look carefully, it has lower measured distortion at 100 dB, compared to 90 dB! Paradoxical, isn’t it?

      Second, the measurements for the DN-2000J’s distortion at 100 dB are nearly perfectly smooth between 20-16k, which is odd, because these measurements are THD+N, rather than THD, and is almost surely done via discrete Fourier transform. We should expect sawtoothed patterns like the rest of the measurements. These findings suggest that there’s something going on with Tyll’s analyzer, so I’m not sure if these distortion measurements are fully reliable.

      In general, don’t take absolute stock in measurements. They can reveal a lot about a product, but they’re not the final word, either. And because most of us do these measurements on our own time for hobbyist pursuits and not for peer-reviewed scientific publication, we don’t triple- quadruple-check the integrity of our results.

      Like

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: