Fit and Comfort

Seen above is the right channel of the XENNS Mangird Top placed into an artificial ear mold with the included size M black bore silicone ear tip installed. I have average-sized ears, and the ear mold above represents my own experiences well enough as a proxy. Size M silicone tips are my go-to for testing, since foam tips are not included by some brands. The Mangird Top is larger-than-average in physical size owing to the high driver count inside, and it's bulbous to an extent that does help keep the inside more manageable. The ergonomic shape helps fill out the concha but it will likely stick out of the ears for you as it did for me—those with smaller ears might struggle with comfort. On the other hand, the manageable nozzles fit well in my ear canals and this helped provide a secure and isolating fit to where I had no problem keeping and using these IEMs for hours on end. The shells weigh ~6.5 g each to make for a low density and minimize any potential of being physically fatiguing. The pre-formed ear hooks combined with the relatively supple cable also makes for an additional point of contact and support for the IEMs, although once again you can see where having flush-mount connectors or even an angled housing would have been better. The one thing to note here is the pronounced ridge on the inside of the shells that may or may not work for you and can even result in a weak/uncomfortable fit depending on the shape of your concha and not just the size. I will also mention this shell design coupled with the single vent at the top makes for driver flex for some folks—especially with the dynamic driver—but I did not experience it with the Mangird Top.

Audio Performance

Audio Hardware

The Mangird Top employs a hybrid driver configuration in that there are two different driver types here for a total of nine drivers per side! This makes for one of the more intricate hybrid driver designs I've tested to date and goes back to explaining the relatively larger nature of these IEM shells. XENNS uses a single 10 mm dynamic driver using a beryllium-coated diaphragm for the lower frequencies, with the coating aiming to make for a stiffer and more responsive driver for good bass reproduction. Two Sonion 2600 series balanced armature drivers handle the mids and the transition to the higher frequencies where two E50D dual balanced armature drivers take over for the treble. There is also a Knowles 32406 dual balanced armature driver for ultra-high frequency, effectively acting as a super tweeter. The drivers are part of a 4-way crossover system and XENNS also publishes a factory frequency response curve albeit without clarifying the test system and test conditions to where it's not the most useful. Overall I'd describe the XENNS Mangird Top as being fairly average among IEMs when it comes to being easy to drive with its rated impedance of 25 Ω and sensitivity of 107 dB (presumably this is dB/mW @1 kHz)—it can be easily driven off a basic dongle, let alone necessitating a higher end portable DAC/amp such as the Questyle M15 which does make for a great sounding, highly portable combination to show simply getting loud isn't always enough!

Frequency Measurement and Listening

I will mention that I have a general preference for a warm-neutral signature with a slightly elevated bass, smooth treble range, detailed mids, and good tonal separation. I also generally prefer instrumental music over vocals, with favored genres including jazz and classical music.


Our reproducible testing methodology begins with a calibrated IEC711 audio coupler/artificial ear that IEMs can feed into enough for decent isolation. The audio coupler feeds into a USB sound card, which in turn goes to a laptop that has ARTA and REW running and the earphones connected to the laptop through a capable and transparent DAC/amp—I used the iKKO Heimdallr ITB03 here. I begin with an impulse measurement to test for signal fidelity, calibrate the sound card and channel output, account for floor noise, and finally test the frequency response of each channel separately. Octave smoothing is at the 1/12th setting, which nets a good balance of detail and noise not being identified as useful data. Also, the default tuning was used for testing, and no app-based settings were chosen unless specifically mentioned. Each sample of interest is tested thrice with separate mounts to account for any fit issues, and an average is taken of the three individual measurements for statistical accuracy. For IEMs, I am also using the appropriate ear mold fitted to the audio coupler for a separate test to compare how the IEMs fare when installed in a pinna geometry instead of just the audio coupler. The raw data is then exported from REW and plotted in OriginPro for easier comparison.


The IEC711 is such that you can't really compare these results with most other test setups, especially those using a head and torso simulator (HATS). The raw dB numbers are also quite contingent on the set volume, gain levels, and sensitivity of the system. What is more useful information is how the left and right channels work across the rated frequency response in the XENNS Mangird Top. The left channel was separately tested from the right one, and colored differently for contrast. I did my best to ensure an identical fit for both inside the IEC711 orifice, so note how the two channels are very similar across the entire useful 20 Hz to 20 kHz range! The measured discrepancy at ~6-8 kHz wasn't one I noticed much admittedly, so I am leaning towards calling them measurement artifacts borne out of aiming to match the coupler resonance to 8 kHz, or trying to anyway! Given the price range the Mangird Top operates in, I expect to see driver matching with decent quality control and, at least when it comes to this review unit which happens to be a randomly chosen retail unit, things are satisfactory enough. Measurements taken after 25 hours of testing, which included these playing a mix of various songs as well as white or pink noise and sine sweeps, showed no difference. There was no perceived burn-in effect thus, and none was measurable, either. The response with the anthropomorphic pinna in place matched the ideal scenario in the coupler quite well too and this is an indicator of how good the seal was when installed in the artificial ear.


Here is the average frequency response for both channels of the XENNS Mangird Top plotted against my personal target taken from VSG.squig.link, which also gives you an idea of my personal preferences to better correlate any possible biases. The tuning of a set of headphones or earphones does not have to match my target as long as it is tuned with some direction, makes sense, and is executed well. After all, no one set will appeal to everyone, and having different options is what makes this hobby so interesting and hard to quantify. Indeed, given the nature of this tonality, I decided to also show how the Mangird Top looks compared to the Harman in-ear 2019v2 target which it matches better against. This gives you a good idea of what to expect already but things can be deceiving.

XENNS does seem to like its bass and the Mangird Top provides quite a lot of it. How you perceive it will depend on your preferences and the fit achieved, and I know people who absolutely love it while others think its just not good. There appears to be a bass tuck to an extent, which some would argue is more the ear gain starting from ~200 Hz as most HRTF studies recommend. This, combined with the clearly prominent sub- and mid-bass presence, can be too much volume without a lot of detail to back it up. The Mangird Top has more bass for my preference too, although I do find it quite dynamic and resolving. Electronic music in particular is a delicious feast here whereas those who prefer both rock or metal music with bass guitars and growly deep male vocals may want to look elsewhere.

Indeed, one of the issues of this tonality is the mids come off quite recessed to where male vocals aren't the best represented here in general. It also can affect brass instrument reproduction and imaging but I'd still say the Mangird Top has great instrument separation and above par imaging. The overall detail retrieval is also plenty fine in my books and the same can be said about the soundstage that feels just about right—think a jazz club with music surrounding you in a more intimate manner but still having the room to appreciate it. Some female vocals came off brighter than others, however—Whitney Houston and Dua Lipa, for example—and Kpop music can be on the shrill side too. I do think the Harman target is also on the shoutier side so in that regard the Mangird Top comes off more forgiving. The upper mids leading to the lower treble is one of the better executed performances here which makes for piano keys to be rendered well and the overall timbre coming off only slightly metallic but close enough to make drums and string instruments feel at home. The set also has a pretty smooth treble response all things considered—it makes for a set that isn't fatiguing and quite pleasant for long listening sessions with genres such as acoustic jazz. That said, the Mangird Top does not have the most technical performance here compared to sets using ESTs and planar tweeters perhaps, given some fundamental tones don't shine in a sea of mixed harmonics—this is especially noticeable with cymbal clashes, harps, as well as certain piano keys. Given it is far more amenable to treble-sensitive folk too, I'd say overall the treble response is a win.

Comparisons

I have been fortunate enough to have spent plenty of time with all of XENNS's works since the rebranding and I will give the team at XENNS props for not doing the same tuning style each time. Note how they go from U-shaped with the oldest Mangird Tea all the way to Harman-like with the Mangird Top covered here. The XENNS UP is the most expensive here being a tribrid set with EST drivers and yet I would claim the Mangird Top to be the best from XENNS to date! It may not be to my exact preferences given the nature of the bass response compared to the mids, but does what it set to do better than the others to where, if you wanted to try out this style of tuning, then you can't do much wrong going with the Mangird Top. I'd also say this is one of very few sets where I would not do much EQ either since it's already set up to do a certain task well.


The Harman-like tonality begs the obvious comparison to the MOONDROP Variations—another tribrid set—which I do not have any experience with. Instead, I can talk about three other sets competing near the same price point that you may want to also consider if you are more open to tuning styles. The LETSHUOER EJ07M comes off more V-shaped relative to the Mangird Top, and ends up with more sub-bass emphasis too. It also comes in CIEM form if you wish to go that route. The standard version employs a metal shell and is easier to fit in the ears to where I'd give the comfort win to the EJ07M. I do think the Mangird Top is overall a better set though, especially given that smooth treble and bass response that is a better take on the Harman shelf for those looking at a more fun, bassy sound. Then there's the ThieAudio Oracle MKII that tilts things towards a brighter signature with a more tame bass shelf that I handle, some of the better vocals out of any IEMs I've heard to date, also more comfortable to wear, and then undoes all the good things by going with an unfortunately shouty, even fatiguing and slightly sibilant upper mids and treble response that is also hard to solve by EQ. If only the Oracle MKII and Mangird Top could combine together would make for an unstoppable force for the money! DUNU then comes in the picture with another hybrid set—the VULKAN—that is another take on this slightly colored tonality but is far more balanced than the rest. This is a tougher competition given also the better build, albeit the Mangird Top is more aesthetically pleasing, but once again the higher frequencies come to XENNS' rescue to where you now need to decide what is more important. There are aspects the VULKAN handles better than the Mangird Top and it does cost less too.