Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 March 2023

Reviewed on: SoundStage! Solo, March 2023

I measured the Sennheiser Momentum 4 headphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 QC audio analyzer. A Reiyin WT-HD06 Bluetooth transmitter was used to send signals from the Clio 12 QC to the headphones. A Samsung Galaxy S10 smartphone served as a source for certain measurements. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

This chart shows the Momentum 4 headphones’ frequency response in Bluetooth mode with noise canceling on full. There’s nothing really out of the ordinary here. The left- and right-channel measurements don’t match all that well; this could be the result of the gating needed to overcome Bluetooth’s latency, but I tried many times to get them to match, and couldn’t.

This chart (done using the Clio’s FFT function with white noise, so it looks somewhat different) shows the difference in response with the noise canceling on and off, Transparency mode, and with a wired connection with power off. Admirably, the noise canceling has no effect on the frequency response. However, Sennheiser doesn’t seem to have put any work into acoustical tuning, because the wired response with power off (which can’t exploit the internal digital signal processing) is a mess—adequate for plugging into an airplane seat and watching old Seinfeld episodes, but not in any case where you care about what the material sounds like.

This chart shows the Momentum 4 headphones’ response compared with a few competitors, all in Bluetooth mode with noise canceling on—except for the AKG K371s, which I’m using as a Harman curve proxy. The Momentum 4s seem a little light in the midrange, but otherwise largely in the ballpark with other good noise-canceling headphones.

The Momentum 4 headphones’ right-channel spectral-decay plot (measured with the wired connection) has a bit of resonance in the bass, but it’s well-damped and basically gone in a few milliseconds.

Here’s the THD vs. frequency, measured using the wired connection at 90dBA and 100dBA (both levels set with pink noise). The distortion gets pretty high in the bass, but only at extremely loud levels, and bass distortion is much less audible because the distortion harmonics are well below the human ear’s range of greatest sensitivity.

In this chart, the external noise level is 85dB SPL (the red trace), and numbers below that indicate the degree of attenuation of outside sounds. The lower the lines, the better the isolation. I threw in the Momentum 4s’ Transparency mode, which, strangely, doesn’t seem to let in much sound above 1kHz. The Momentum 4s’ noise canceling isn’t the best, but it’s pretty close to the best.

Note that the transmitter showed it was in aptX Low Latency mode. The headphones are equipped with aptX and aptX Adaptive.

The impedance magnitude, measured in wired mode with power off, measures about 75 to 80 ohms up to 1kHz, and then falls to a minimum of about 52 ohms, with a corresponding electrical phase shift.

Sensitivity with the wired connection, power off, averaged between 300Hz and 3kHz, with a 1mW signal calculated for the rated 80 ohms impedance, is 105.6dB, plenty high enough to get loud levels from any source.

Bottom line: The Momentum 4 headphones’ measured performance looks good, with a pretty safe and sane frequency response and very good noise canceling. The only sore spot is the wackadoodle frequency response in wired/power-off mode.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 20 January 2023

Reviewed on: SoundStage! Solo, January 2023

I measured the Focal Utopia 2022 headphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and a Clio 10 FW audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. The headphones were amplified using a Musical Fidelity V-CAN. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

The above chart shows the Utopia 2022s’ frequency response. This is fairly ordinary for audiophile headphones, but for two anomalies. First, that unusual little peak at about 1.4kHz, (also seen in the original Utopias), which will likely add a slight emphasis in the upper midrange. Second, the low energy above 8kHz, which isn’t surprising after hearing how the Utopia 2022s seem to damp extreme high-frequency sounds.

This chart shows how the Utopia 2022s’ tonal balance changes when they’re used with a high-impedance (75 ohms) source, such as a cheap laptop or some cheap professional headphone amps, or some tube amps. There’s a substantial bass boost with the high-impedance source—about +3.5dB, centered at about 65Hz—so these may get bassier when used with an amp that has a tube output stage.

This chart shows the Utopia 2022s’ right-channel response compared with the original Utopias and two other well-regarded high-end models: the Meze Elites (used with the supplied Alcantara earpads) and the Sennheiser HD 800 S headphones. You can see how the Meze and Sennheiser models have more energy in the top octave-and-a-half.

Just as with the original Utopias, the Utopia 2022s’ spectral-decay plot looks like the plots I see with planar-magnetic headphones, with a lot of very high-Q resonances between about 1 and 5kHz. That’s weird because planar drivers are flat, but the Utopia 2022s’ drivers are M-shaped. I associate these resonances with a nice sense of space, rather than an audible coloration, but to my knowledge, there’s no science on that—although a famous audio scientist was the one who gave me the idea.

There’s a little bit of distortion in the deep bass with the Utopia 2022s at very high levels, but you’ll never hear it because the distortion harmonics are still well below the most sensitive range of human hearing.

In this chart, the external noise level is 85dB SPL, and numbers below that indicate the degree of attenuation of outside sounds. The lower the lines, the better the isolation. The isolation of the Utopia 2022s, like that of most open-back headphones, is negligible. I added the original Utopias so you could see the effects of the changes in the outer grille (it does look like the new one is somewhat more transparent at high frequencies), and the Focal Stellias so you can see how a closed-back model compares.

The Utopia 2022s’ impedance is mostly flat at the rated 80 ohms for about 6.5 octaves, but then it takes a huge swing up to 770 ohms at 52Hz—which shows there’s been a change in the voice coil, because the old model rose only to 655 ohms. There’s also a big phase swing in the bass, corresponding with the impedance peak. This is why these headphones will sound different with high-impedance sources.

Sensitivity of the Utopia 2022s, measured between 300Hz and 3kHz, using a 1mW signal calculated for 80 ohms rated impedance, is 105.6dB—exactly the same as the original model, and no problem for any source device to drive.

Bottom line: There are a couple of unusual frequency-response characteristics with the Utopia 2022s, and definitely not a strong upper-treble response, but the engineering otherwise seems solid, with particular praise for the sensitivity, which is way high for high-end headphones.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 10 January 2023

Reviewed on: SoundStage! Solo, January 2023

I measured the Akoustyx S-6 earphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. The earphones were amplified using a Musical Fidelity V-CAN amplifier. Except as noted, I used the supplied medium-sized silicone tips for all measurements. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

This chart shows the S-6es’ frequency response. This is a very “smiley” response, with sharply boosted low bass and, especially, treble. I’d expect the midrange to sound recessed in comparison.

This chart shows how the S-6es’ tonal balance changes when they’re used with a high-impedance source, such as a cheap laptop, some tube amps, or some professional headphone amps. There’s just a small difference, a boost of about 1dB above 5kHz.

This chart shows the S-6es’ right-channel response compared with various earphones, including the AKG N5005s, which are said to be the passive earphones that come closest to the Harman curve. You can easily see how far outside the norm the S-6 earphones’ response is.

The S-6es’ spectral-decay plot looks mostly clean; there’s some bass resonance, but it’s well-damped and down by about 40dB after the first few milliseconds.

The S-6es’ harmonic distortion is mostly low; there’s a little bit of a rise between about 1 and 3kHz, but even at the crazy-loud level of 100dBA, it’s still below 2 percent.

In this chart, the external noise level is 85dB SPL, and numbers below that indicate the degree of attenuation of outside sounds. The lower the lines, the better the isolation. In the 43AG ear/cheek simulator, the S-6es offer a fairly modest amount of isolation with just the silicone tips, but if you add the silicone wings and use the foam tips, the isolation is truly outstanding.

The impedance curve of the S-6 earphones runs about 16 ohms up to about 2kHz, but above 5kHz, it starts to dive, ending up at 10 ohms at 20kHz; that’s why the response at these frequencies is down about 1dB with a high-impedance source. The phase takes a similar dive.

Sensitivity, measured in the right channel between 300Hz and 3kHz, using a 1mW signal calculated for 18 ohms rated impedance, is 102.9dB. That’s high enough that no portable device should have problems getting the S-6es to play as loud as you want.

Bottom line: The basic engineering of these earphones looks solid, but the frequency response shows extreme treble and low-bass boost, which I have to think will sound unnatural, even though decades of audio history suggests some listeners will probably enjoy it.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 January 2023

Reviewed on: SoundStage! Solo, January 2023

I measured the Moondrop Quarks earphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 QC audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. The earphones were amplified using a Musical Fidelity V-CAN amplifier. Except as noted, I used the supplied medium-sized silicone tips for all measurements because they fit best in the ear simulator. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

This chart shows the Quarks’ frequency response. This is within the norm for this type of product, although the peak at 3kHz is much higher-Q (i.e., narrower) than normal, and there’s a little less bass than usual, plus a little less oomph in the top two octaves (5 to 10kHz and 10 to 20kHz) of treble. The left channel is also consistently more sensitive than the right, with about 1.5dB more output for the same input voltage. I generally don’t notice mild imbalances like these, but some people might on certain recordings.

This chart shows how the Quarks’ tonal balance changes when they’re used with a high-impedance source, such as a cheap laptop, some tube amps, or some professional headphone amps. There’s no difference, so the Quarks will perform consistently with different source devices.

This chart shows the Quarks’ right-channel response compared with various competitors, including the AKG N5005 earphones, which are said to be the passive earphones that come closest to the Harman curve. The Quarks’ quirks are mild, as I described above.

The Quarks’ spectral-decay plot looks very clean, with no audible resonances.

The Quarks’ harmonic distortion is low, even at the mega-cranked level of 100dBA (measured with pink noise).

In this chart, the external noise level is 85dB SPL, and numbers below that indicate the degree of attenuation of outside sounds. The lower the lines, the better the isolation. In the 43AG ear/cheek simulator, the Quarks offer a typical level of isolation for compact earphones, although less isolation above 2kHz than a couple of (much more expensive) competitors offer.

The impedance curve of the Quarks is almost entirely flat at 15 ohms, and the phase response is also flat.

Sensitivity, measured in the right channel between 300Hz and 3kHz, using a 1mW signal calculated for 16 ohms rated impedance, is 93.1dB. That’s very low for earphones. So some portable source devices might have problems getting the Quarks to play as loud as you want with some recordings.

Bottom line: The measurements show some compromises in the Quarks’ performance, which has to be expected in a set of earphones this inexpensive. But other than the very low sensitivity, there’s nothing here that would likely pose a problem. For the price, the safe’n’sane frequency-response tuning is surprising and impressive.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 20 December 2022

Reviewed on: SoundStage! Solo, December 2022

I measured the Crosszone CZ-10 headphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 QC audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. For most measurements, the headphones were amplified using a Musical Fidelity V-CAN amplifier; I used a Schiit Magnius amplifier for distortion measurements. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

This chart shows the frequency response of the CZ-10s’ right- and left-channel drivers only, without the contribution of the crossfeed drivers. (This is what I think will show what the tonal balance of these headphones looks like; adding in the contribution of the crossfeed drivers might sound good to the ear, but the cancellation effects will just look bizarre to a microphone.) This isn’t too crazy a response, considering how unusual the design is. The response is definitely weak in the top two octaves of treble. The bass, while it looks adequate, certainly doesn’t appear to be what you’d call bumping—but more on this later. The peak at around 2.8kHz is common, although its high Q means it covers about 1/4 octave, where usually this peak would cover about a full octave. So this probably wouldn’t sound too unusual, but it definitely wouldn’t sound airy.

This chart shows the right-channel response from above, with the response of the left-channel crossfeed driver in the same earcup of the headphone. The contribution of the crossfeed driver peaks between 1 and 2kHz, but at higher frequencies, the contribution (and, I assume, the spatial effects) won’t be strong.

Here we can see the difference in the CZ-10s’ response when a high-impedance (75 ohms) source is substituted for a typical low-impedance source (5 ohms). The difference is negligible, so as long as the amp has enough power to drive the CZ-10s, it shouldn’t affect their sound.

This chart shows the CZ-10s’ right-channel-only (no crossfeed driver) response compared with two closed-back models (the Beyerdynamic T5 3rd Generation and AKG K371 headphones) and an open-back model (the HiFiMan Sundara headphhones). All are normalized to 94dB at 500Hz. In this chart, the CZ-10s certainly seem bass-deficient, but they didn’t sound that way; my guess is that the left-channel crossfeed driver is in phase with the right-channel drivers at low frequencies, and thus elevates the overall bass level.

The CZ-10s’ spectral decay—i.e., resonance plot—shows a pretty strong resonance centered at 2.8kHz, the same frequency as the peak in the frequency response. But it’s well-damped, and totally imperceptible after just 10ms.

Here’s the THD vs. frequency chart, measured at 90dBA and 100dBA, both levels set with pink noise, using just the right-channel drivers with no crossfeed driver. There’s some observable distortion in the bass, which isn’t unusual for dynamic drivers, but considering that it’s below 4% THD even at the insanely loud level of 100dBA, I’d be very surprised if it’s audible.

In this chart, the external noise level is 85dB SPL (the red trace), and numbers below that indicate the degree of attenuation of outside sounds. The CZ-10s’ isolation is almost the same as all the other models included here.

The impedance of the CZ-10s takes a big plunge, running about 85 ohms in the bass but plunging to below 30 ohms in the treble. There’s a corresponding phase swing (apologies for exaggerating the phase shift; my scale is usually +180 to -180 degrees). I’m surprised, given this result, that there wasn’t more of a change when I went from the 5-ohm source above to the 75-ohm source.

Sensitivity of the CZ-10s, calculated for 75 ohms rated impedance, using all of the drivers in the right earpiece and averaged from 300Hz to 3kHz, is 100.9dB with a 1mW signal. That’s not terribly low, but low enough that you may not get enough volume if you plug these straight into a smartphone or tablet.

Bottom line: It’s tough to judge headphones like this based on conventional measurements, because they’re about how the contributions of the different drivers add up at the eardrum and are interpreted by the brain. I can confidently say that the Crosszone CZ-10 headphones are lacking in high-frequency response, but that’s the only clear flaw I can see in the measurements.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 10 December 2022

Reviewed on: SoundStage! Solo, December 2022

I measured the Sivga Oriole headphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 QC audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. For most measurements, the headphones were amplified using a Musical Fidelity V-CAN amplifier; I used a Schiit Magnius amplifier for distortion measurements. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

This chart shows the Orioles’ frequency response, which shows some weird stuff. That dip centered near 200Hz looks scary, especially considering that it’s about an octave wide, although I should note that response dips are less audible than peaks, if the magnitude and Q are comparable. However, above about 500Hz, they’re not very far off the Harman curve—but there’s a lot less energy below 500Hz than the Harman curve recommends.

Here we can see the difference in the Orioles’ response when a high-impedance (75 ohms) source is substituted for a typical low-impedance source (5 ohms). It’s just a 1dB bump in the bass with the high-impedance source, which would be at most barely noticeable.

This chart shows the Orioles’ right-channel response compared with the open-back Sivga P-IIs, the HiFiMan Sundara Closed-Backs, and the AKG K371s (one of the headphones said to come closest to the Harman curve). All are normalized to 94dB at 500Hz. Clearly, the Orioles are outside the norm in some ways.

The Orioles’ spectral decay—i.e., resonance plot—shows a pretty strong resonance centered at about 500Hz, but it’s got a very high Q (i.e., narrow bandwidth), so I doubt it’ll be very noticeable. There’s another resonance centered at about 4kHz; it’s better-damped, but it might result in a slightly bright sound at times.

Here’s the THD vs. frequency chart, measured at 90dBA and 100dBA, both levels set with pink noise. There’s some observable distortion in the bass at very high levels, and certainly more than you’d likely get with a planar-magnetic driver. But considering that it’s only about 2% THD at the loud level of 90dBA, I doubt it’d be audible.

In this chart, the external noise level is 85dB SPL (the red trace), and numbers below that indicate the degree of attenuation of outside sounds. The Orioles’ isolation is excellent, delivering what should be an audibly quieter experience than the other two closed-back models included in this chart. I threw in the Sivga P-II headphones so you could see how an open-back design compares.

The impedance of the Orioles is nearly flat at about 35 ohms through most of the audio range, with a rise and corresponding phase-response wrinkle centered at 45Hz, which is typical of a dynamic driver. That rise is the cause of the difference in frequency response when the Orioles are used with a high-impedance source, but considering that it’s only about a 1dB bump in the bass, it’s nothing to be concerned about.

Sensitivity of the Orioles, calculated for 32 ohms rated impedance and averaged from 300Hz to 3kHz, is 106.8dB with a 1mW signal, which is plenty high enough that any source device should be able to drive these headphones to loud levels.

Bottom line: The Oriole headphones’ engineering looks solid, and you can certainly drive them with any conceivable source device. However, the frequency response is idiosyncratic. Whether you’ll dig it, I can’t say, but I’d recommend you hear them before you buy, or buy them from a merchant that accepts returns.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 December 2022

Reviewed on: SoundStage! Solo, December 2022

I measured the Linsoul 7Hz Timeless earphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. The earphones were amplified using a Musical Fidelity V-CAN amplifier. Except as noted, I used the supplied medium-sized silicone tips for all measurements because they fit best in the ear simulator. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

This chart shows the 7Hz Timelesses’ frequency response. This is fairly normal for earphones, with a few exceptions. There seems to be a bit of excess upper bass, and the 3kHz peak, while common, is unusually narrow in bandwidth. Also, there’s less energy in the mid-treble (around 7kHz) than we might normally see, but more energy above 10kHz than is normal for earphones. I can’t really speculate as to what this combination will sound like—maybe a little “smiley” (i.e., bass- and treble-boosted)?

This chart shows how the 7Hz Timelesses’ tonal balance changes when they’re used with a high-impedance source, such as a cheap laptop, some tube amps, or some professional headphone amps. There’s a very tiny, and inaudible, bass boost below 30Hz.

This chart shows the 7Hz Timelesses’ right-channel response compared with various earphones, including the AKG N5005s, which are said to be the passive earphones that come closest to the Harman curve. Basically, the variations from what’s more or less “normal” for earphones are as I described above.

The 7Hz Timelesses’ spectral-decay plot looks clean; that peak at around 3kHz corresponds with the earphones’ response peak at that frequency, and it’s very well-damped and won’t be audible as a resonance.

The distortion we see with the 7Hz Timeless earphones is much less than we normally see from earphones with dynamic drivers or balanced armatures—but comparable to what we usually see from planar-magnetic headphones.

In this chart, the external noise level is 85dB SPL, and numbers below that indicate the degree of attenuation of outside sounds. The lower the lines, the better the isolation. In the 43AG ear/cheek simulator, the 7Hz Timelesses offer less isolation than most earphones do; I imagine that’s because their large size prevents them from filling the ear as completely as most earphones can.

The impedance curve of the 7Hz Timelesses is dead flat at 14 ohms, which is typical for planar-magnetic drivers. The phase response is also flat.

Sensitivity, measured between 300Hz and 3kHz, using a 1mW signal calculated for 14.8 ohms rated impedance, is 99.2dB. That’s low for earphones, but still high enough that the 7Hz Timelesses should play reasonably loud from almost any source device.

Bottom line: Looks like pretty solid engineering here, with high-quality drivers. There are some frequency-response anomalies; I’m eager to find out what Geoff Morrison heard.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 20 November 2022

Reviewed on: SoundStage! Solo, November 2022

I measured the HiFiMan Sundara Closed-Back headphones using laboratory-grade equipment: a GRAS Model 43AG ear/cheek simulator/RA0402 ear simulator with KB5000/KB5001 simulated pinnae, and an Audiomatica Clio 12 QC audio analyzer. For isolation measurements, I used a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface. For most measurements, the headphones were amplified using a Musical Fidelity V-CAN amplifier; I used a Schiit Magnius amplifier for distortion measurements. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed. If you’d like to learn more about what our measurements mean, click here.

Well, this is weird. This chart shows the Sundara Closed-Backs’ frequency response, which looks a lot more idiosyncratic than I expected. There’s a strange bump around 300Hz (the source of the extra upper-bass energy I heard), and then an unusual midrange bump centered at about 1.3kHz. There’s a lot of treble energy, although probably not enough to make these headphones sound overtly bright.

Here we can see the difference in the headphones’ response when a high-impedance (75 ohms) source is substituted for a typical low-impedance source (5 ohms). As usual with planar-magnetic headphones, there’s no significant difference, so these headphones’ response will not change significantly when you switch to a different source device.

This chart shows the Sundara Closed-Backs’ right-channel response compared with the open-back Sundaras, the Focal Celestees, and AKG K371 closed-back headphones (the K371s are one of the headphones noted for coming closest to the Harman curve). All are normalized to 94dB at 500Hz. Clearly, the Sundara Closed-Backs have more lower- and mid-midrange energy than the other headphones, more energy around 5 to 7kHz, and less energy in the 3kHz range. Frankly, I’m surprised they sound as normal as they do.

The Sundara Closed-Backs’ spectral decay—i.e., resonance plot—has the usual “hash” we see with planar-magnetic models. This doesn’t present itself as a coloration, per se, but it does seem to enhance the sense of space. Normally I might expect to see a broad hash band between 2 and 5kHz, but here, we see narrower hash bands instead, one centered at about 3.5kHz, the other spanning the range between 6 and 9kHz.

Here’s the THD vs. frequency chart, measured at 90dBA and 100dBA, both levels set with pink noise. With most planar-magnetics, these lines show near-zero distortion, but with these, we see a distortion peak that corresponds with that frequency-response anomaly around 300Hz, and more distortion peaks that correspond with the high-frequency hash bands we saw in the previous measurement.

In this chart, the external noise level is 85dB SPL (the red trace), and numbers below that indicate the degree of attenuation of outside sounds. The Sundara Closed-Backs’ isolation is more or less in the range of other closed-back models. I also included the HiFiMan HE400se headphones, so you can see how an open-back model compares.

The impedance of the Sundara Closed-Backs is essentially flat at 19 ohms, with a correspondingly flat phase response—par for the course with planar-magnetics.

Sensitivity of the Sundara Closed-Backs, calculated for 20 ohms impedance and averaged from 300Hz to 3kHz, is 93.2dB with a 1mW signal, a little lower than rated, but I do this measurement differently than the industry standard, which measures only a 500Hz tone.

Bottom line: It’s hard to ignore the Sundara Closed-Back headphones’ idiosyncratic frequency response. I liked them a lot, but I’d suggest you give them a listen yourself before you buy, if possible.

. . . Brent Butterworth
brentb@soundstagenetwork.com