Fit and Comfort

Seen above is the MOONDROP COSMO placed on a mannequin head, to help show how these headphones would look on the human head. Note that the head is slightly under average sized, so account for the discrepancies accordingly. As with all headphones, getting a good fit and seal is crucial, so make sure to properly use the various pivot points, and the height adjustability of the headband and ear cups. This section is mostly a summary of the points expanded upon in more detail on the previous page, but I will point out that the COSMO is rated to weigh ~556 g, with my sample coming in at 545 g without the cable. This is an appreciable decrease from the VENUS, and we can see how that has come about thanks to the slimmer headband and suspension band, as well as the use of a plastic baffle. Still, I find the COSMO less comfortable owing to the more bottom-heavy design and the thin suspension band with minimal padding, and the slimmer ear pads don't help much either. Then there is the part where the headband sizing options feel more designed with larger heads in mind, and I was just glad to have obtained a good fit and seal around the ears with the lowest setting at least, but this feels like a misstep from MOONDROP on multiple fronts—especially for its new flagship headphones. Do keep in mind that the COSMO is an open-back set and thus will be best used in a quiet environment without others in the vicinity.

Audio Performance

Audio Hardware

MOONDROP is one of very, very few audio brands that makes in-house drivers, including for its headphones. These include dynamic drivers and planar magnetic drivers too, as well as the various electrostatic drivers that have been made in the pursuit of the yet-unreleased MoonZero. The planar magnetic drivers used in the COSMO are quite impressive from an engineering perspective, with the MOONDROP team arguably going further than anyone would realistically expect from a new-to-headphones brand, let alone for a product that is priced well under $1000. Take the diaphragm, for example, which is all of 500 nm (0.5 µm) thick. This ultrathin, light-weight, highly responsive diaphragm is on par with flagships such as the Audeze LCD-4 and HIFIMAN Susvara in terms of offering a snappy response to the same magnetic force as, say, thicker diaphragms. It can curtail the bass impact to an extent, but the 100 mm diameter diaphragm here should still allow for an impressive showing—provided it is tuned well. MOONDROP says it also developed a new tensioning method to further result in the diaphragm being uniform throughout its surface area, which in turn can reduce distortion, and the brand is using a thin film of pure silver as the conductive trace on the diaphragm. The magnet type/size/shape/placement has been optimized using computer simulations to result in a patented "Full Drive Technology" such that there is a larger fraction of the diaphragm now responding evenly to the magnetic force, and there are even 3D waveguides in the acoustic housing to reduce phase cancellation in the higher frequencies. The COSMO continues to use a whopping dual-sided set of 36 neodymium N52-class magnets (18 per side), and yet the diaphragm is clearly highly tensed and not as sensitive given it is rated for a low rated sensitivity of just 100 dB/Vrms paired with an equally low impedance of 15 Ω. Those numbers translate to 81.76 dB/mW which makes the MOONDROP COSMO ask for more current than even the notorious HIFIMAN Susvara/HE6se. In practice, it did feel slightly more demanding than the VENUS, but MOONDROP still appears to be on the conservative side here as the COSMO isn't exceptionally hard to drive. I'd still recommend going with a decently powerful solid state desktop amplifier given the low impedance here, although you could try out a tube pre-amp too. I mostly used the COSMO with the Aune S17 Pro amplifier fed off the Gustard A26 DAC for my listening sessions with the set.

Frequency Response Measurement and Listening

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


Our current headphones test setup uses a set of two custom in-ear microphones for the two channels. These microphones closely adhere to the IEC711 class, but have been tweaked to be more reliable in the >10 kHz frequency range, the precise issue with my previous setup, that is otherwise still very good and will continue to be used for IEMs and earphones. Two soft silicone pinnae are installed on the sides, separated by a distance matching my head, and multiple "height" adapters have been 3D-printed for further customization, based on fit, head size and shape. Each set of microphones has an XLR output I separately adapted to 3.5 mm. I used a transparent source—the FiiO K9 Pro ESS—for measurements after confirming it was not a bottleneck in any way, but also used a DAP, as seen above, to confirm it was no different for this purpose. This artificial head simulator feeds the microphone lines into a reference USB sound card, which in turn goes to a laptop that has ARTA and REW running. I begin with an impulse measurement to test for signal fidelity, FFT to test for headphone seal, 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, netting a good balance of detail and signal-to-noise ratio. The default tuning was used for testing, and no app or program-based EQ settings were chosen, unless specifically mentioned. Each sample of interest is measured at least thrice with separate mounts to account for any fit issues, and an average is taken of the individual measurements for statistical accuracy.


As before, you can find my headphone frequency response measurements on VSG.squig.link along with all the earphone measurements. Scroll to the bottom and choose different targets there, including two from Harman Kardon developed after years of R&D. The Harman 2018 over-ear target in particular, is a reference curve many headphone makers aim for now, but I find it too bass-boosted. As such, I am opting for the Harman 2018 curve with the bass target from the Harman 2013 curve, which is what is referred to as the "Harman Combined" target. Before we get talking about the sound signature of the MOONDROP COSMO, I want to briefly talk about the measurements themselves and the channel balance achieved here. I think a single look at the measurements above can confirm the excellent driver matching on my set, but I should mention that my review sample is one of the very first retail COSMO units ever made and well before MOONDROP started manufacturing them in high enough volumes to send out review samples galore—you will see soon why I ended up waiting to publish this review. So there is a possibility this particular unit may have hand-matched drivers, but MOONDROP does talk about more stringent QC tests and driver matching across its entire IEMs and headphones lineup lately, so take this for what you will.

As for the sound itself, I am indifferent with the COSMO—at least as it comes out of the box. It sounds bass lite, but also relaxed in the upper mids and treble to where it almost feels like a studio monitoring set that you would add your own EQ filters on top, similar to the likes of the HEDDphone 2. The VENUS had a more typical neutral-bright planar magnetic signature, and the PARA went for more forward mids, making all three MOONDROP planar magnetic headphones sound distinct from each other. I can appreciate this to an extent, however the three compete against other sets on the market and are also priced quite differently too. The end result is the COSMO is arguably the least exciting sounding, and this is a shame because the drivers are capable of so much more. There's also the part where the COSMO has a noticeable bump in the 1-2 kHz range, making for some vocals to come off honky and nasally, which is compounded further by the treble being on the darker side overall. Bass extension is decent, and honestly you should absolutely EQ a bass shelf here because the drivers are fantastic. There's no distortion to hold you back even at extremely loud listening volumes, and you get so much detail too. The COSMO also has a really low resonant frequency, making it ideal for those wearing glasses. It can be quite punchy, even for sub-bass lovers, as long as you work on tuning it to your preference. The stock tonality unfortunately feels quite soft and lacking impact, mostly due to the subdued lower treble, and generally I felt this extremely open-back set was sounding closed-off as a result. I have no idea what MOONDROP was thinking, except that perhaps it wanted to appeal to those who find such sets to be overly bright and fatiguing. The end result is poor timbre and a generally narrow soundstage too, and I really can't recommend the COSMO to people who don't plan to EQ or at least try some pad rolling.


I was reminded of the HarmonicDyne G200 to a good degree here, with the COSMO also having excellent, highly competent drivers being held back by their tuning. I know a lot of people are still against EQ for various reasons—seriously, you are just losing free performance here—and I was in two minds on how to really judge the COSMO as it comes. In the time since I had received this sample, a few people online mentioned that the MOONDROP PARA ear pads made the COSMO sound much more lively. Unfortunately, MOONDROP does not sell these particular pads separately so I don't see a point in including them in a review, and I don't think you should be purchasing headphones just to use their pads on another product either. MOONDROP does sell an upgrade set of ear pads called the EP-100A, which are intended for the PARA but will fit the COSMO perfectly too with the same install mechanism—you will have to remove the plate off the stock pads and insert them on the back of these pads first though. You will note the EP-100A pads are also hybrid in design, thicker than the stock COSMO pads, and are perforated on the sides. Immediately, these aid in adding more upper mids and lower treble, which helps solve a lot of issues I had but not all. The bass feels arguably even lighter, but now we do have more meaty mids and even the mid-bass notes have some impact behind them as a result of the higher-order resonances being filled in. The sound also opens up in general, allowing room for vocals to not sound boxed in and instruments to have a more ethereal presentation. The 1-2 kHz honk is also reduced by comparison, although the COSMO leans bright and the upper treble presence is even more prominent. This can alienate a different customer base, although I think the best combination is the COSMO with these EP-100A pads and your own EQ profile. Out of curiosity, I also tried the ZMF Caldera stock hybrid pads, which were even thicker, but note that they are ever-so-slightly larger to where I had to stuff them into the ear cups. The COSMO now has slightly more lower treble energy, which also helps improve the listening experience, but not enough to warrant the $80 asking price for these pads in this use case.


The MOONDROP VENUS is the obvious comparison to make here, although it appears MOONDROP is no longer manufacturing the VENUS. I could type pages worth of my thoughts on the deliberately short life cycles of IEMs, and now headphones too, coming out of China but for now all I can say is that if you have the VENUS or find one being sold second hand, then the COSMO is not necessarily a direct upgrade out of the box. The COSMO weighs less but is likely going to be less comfortable regardless. It has significantly better detail retrieval and scales tremendously with EQ, yet sounds unexciting and potentially even wrong in some places with the stock tuning. The COSMO can be easily made to compete with the VENUS, as well as headphones several times more expensive, so it is worth considering if you go with this mindset. The HIFIMAN Arya Stealth is arguably the most popular planar magnetic set in this price range, and probably just outright too irrespective of driver type. It debuted at $1600 in 2021 and now sells for $650, offering HIFIMAN's best headband and a lighter chassis to make it easily the winner in comfort compared to the COSMO, and the build quality is decent too. Sound wise, the Arya Stealth will hit you hard in a lot of ways to where it's decidedly on the brighter side of things. Note the upper mids and lower treble being significantly more prominent here. I'd argue the COSMO + EP-100A pads are more balanced, but you will be paying nearly $300 more now compared to the Arya Stealth and I just can't see MOONDROP winning here. While there are many other open-back sets in this price, including those using dynamic drivers and e-stats, I'll give a shout-out to the iBasso SR3, which I feel is quite underrated even today. The SR3 sounds like an upgraded Focal Clear MG to me, and at a much lower cost too. It is quite dynamic, plenty resolving, very comfortable, and looks/feels premium. There is a noticeable peak at 6 kHz which can be a bit much for people, but otherwise this is a dynamic driver set that I would personally pick up over the COSMO too. If you are amenable to EQ however, then the COSMO can surpass all of these named sets.