Details

Parent Category: Products

Category: Headphone Measurements

Created: 15 December 2017

I measured the WH-1000XM2s using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, and a Musical Fidelity V-CAN amp. On the Model 43AG I used the original KB0065 simulated pinna for most measurements, as well as the new KB5000 pinna for some measurements, as noted. For Bluetooth-sourced measurements I used a Sony HWS-BTA2W Bluetooth transmitter to send signals from the Clio 10 FW to the headphones. These are “flat” measurements; no diffuse-field or free-field compensation curve was applied.

The WH-1000XM2s’ frequency response (shown here with a Bluetooth signal with NC on and with a cabled connection and power off) looks fairly standard, with the usual peak in the 2.5kHz range and another in the 6-8kHz range. What’s unusual is that, despite my best attempts, I got a little more left/right variance than I usually do. Also, it’s obvious that Sony didn’t put a ton of work into getting these headphones to sound good in passive mode. Although I don’t show it here, switching from a 5- to a 75-ohm source in cabled/power-off mode produced a boost of typically 4dB in the bass (depending on frequency), which means that these headphones will take on a different tone if the battery runs down and you plug them into a cheap PC laptop -- admittedly, probably a very minor concern.

This chart shows the WH-1000XM2s’ measured right-channel frequency response in Bluetooth mode with NC on, measured with the old KB0065 pinna (which I’ve used for years) and G.R.A.S.’s new KB5000 pinna. (I’ll eventually switch to the new pinna, because it more accurately reflects the structure and pliability of the human ear, and include it here mostly for future reference rather than as something you should draw conclusions from. I intend to show both measurements in every review for at least the next year before I begin using only the new pinna.)

This chart, measured with pink noise and Clio’s FFT real-time analyzer function (necessary to capture an accurate comparison of the WH-1000XM2s’ various modes, as the Bluetooth mode introduces a latency of about 200ms), shows that the WH-1000XM2 can sound quite different depending on which mode it’s set to. Incidentally, Ambient mode, which lets in outside sounds, measures effectively the same as NC Off mode, whether the headphone is cabled or connected through Bluetooth.

This chart shows the WH-1000XM2s’ measured right-channel frequency response compared with those of three other NC headphones: the Bose QC35 (original model), the PSB M4U 2 (generally considered among the best-sounding NC headphones), and the Sennheiser HD 4.50 BTNC. The WH-1000XM2 seems pretty much in the ballpark when it comes to responses typical of NC headphones -- or headphones in general, for that matter.

The spectral decay (waterfall) chart, measured with the WH-1000XM2s in Bluetooth/NC mode, shows a spectrum of plentiful but narrow resonances between 1 and 6kHz -- a strange result for dynamic headphones, but something I often see even in the best planar-magnetic models. It doesn’t concern me.

The total harmonic distortion (THD) of the WH-1000XM2s, measured with a wired connection because the Clio 10 FW’s sine sweeps can’t accommodate Bluetooth’s latency, is somewhat on the high side in passive cabled mode, but there are unusually large peaks in distortion at 70Hz and 2kHz when NC is switched on and the level rises to the extremely high 100dBA standard I use (and which most headphones pass pretty easily). I didn’t notice the distortion when I was listening, and considering that it’s limited to two narrow bands, you probably wouldn’t, either. I speculate that DSP-based EQ inside the headphones is pushing the internal amp past its limits.

In this chart the level of external noise is 85dB SPL; the numbers below that indicate the degree of attenuation of outside sounds. (Note that I recently switched to measuring at a level of 85dB instead of 75dB; this doesn’t change the way the isolation curves look, but an 85dB level lets me get better measurements of NC headphones, which demand a lower noise floor.) The isolation of the WH-1000XM2s is pretty good for active NC headphones; they can’t beat the industry-leading Bose, but they easily beat the AKG and Sennheiser models I compare them with here.

The WH-1000XM2s’ impedance magnitude is dead flat in powered mode with a cabled connection. With power off, there’s an impedance swing in the bass from the specified 14 ohms up to 39 ohms, as well as some phase shift in the bass, which is why in passive cabled mode you’ll be able to hear a difference in bass response if you plug the Sonys into a low-quality headphone amp such as the ones built into typical laptop PCs.

The sensitivity of the WH-1000XM2s in wired mode with power off, measured between 300Hz and 3kHz with a 1mW signal and calculated for the specified 14 ohms impedance, was 96.3dB. Wired with power on, it’s 101.7dB for the specified 46-ohm impedance in that mode. This means that the WH-1000XM2s won’t play very loud if the batteries run down, but will work just fine if you’re on an airplane with NC on and are using a cabled connection to the airplane’s in-seat entertainment system.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 December 2017

I measured the Monolith M300s using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, and a Musical Fidelity V-CAN headphone amp. With the G.R.A.S. I used the original KB0065 simulated pinna for most measurements, as well as the new KB5000 pinna for certain measurements, as noted. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed.

The M300s’ frequency response is quite unusual, even for such an unusual set of earphones. Basically, if you moved the whole curve 1kHz higher, it would look reasonably normal. Most earphones and headphones have a big response peak between 2 and 3kHz, but the M300s’ peak is centered at 1.2kHz. The deep dip at 1kHz represents a cancellation, I would guess due to the internal acoustics of the eartube, but considering that it’s a dip, and a very narrow one, I doubt it would be readily audible.

This chart shows the M300s’ right-channel frequency response measured with the old KB0065 pinna (which I’ve used for years) and G.R.A.S.’s new KB5000 pinna, which more accurately reflects the structure and pliability of the human ear. I plan to show both measurements in every review for at least the next year before I begin to use only the new pinna, and include this measurement here mostly for future reference rather than as something you should draw conclusions from. Because of the KB5000’s more realistic construction, I’d consider the measurements taken with it more representative of the Monolith M300s’ performance, but unfortunately, at the moment I have only the right-ear model of the new pinna.

This chart shows the results of adding 70 ohms output impedance to the V-CAN’s 5-ohm output impedance to simulate the effects of using a typical low-quality headphone amp. The difference is zero; those little squiggles in the bass are due to noise caused by the much lower recorded level of the signal with the higher-impedance output. (I scaled up the 75-ohm result by 10.3dB in the chart for comparison purposes.)

This chart shows the M300s’ right-channel frequency response compared with that of the very similar Audeze iSine10s, the well-regarded PSB M4U 4 hybrid dynamic/balanced-armature earphones, and the Focal Sphears (rather “normal” single-driver dynamic earphones). The PSBs and Focals have the broad bass hump and strong upper-mid/lower-treble peaks typical of most earphones; the Audezes’ response looks much like that of a typical over-ear planar-magnetic model; but the M300s occupy a world of their own.

The spectral-decay (waterfall) chart looks unusual: While the resonances are well damped and die out quickly, they’re higher in magnitude than the norm, with an especially strong resonance centered at 1kHz.

The M300s’ total harmonic distortion (THD) is effectively nonexistent, even at loud listening levels, with just a narrow blip of distortion centered at 1kHz -- the same frequency as the dip in the frequency response, and the strongest resonance in the spectral-decay measurement. In a rather extraordinary display, the measured distortion at 100dBA is no higher than at 90dBA -- there’s actually an orange trace representing the distortion at 100dBA, but it’s completely obscured by the green trace at 90dBA.

In this chart, the external noise level is 85dB SPL; the numbers below that indicate the degree of attenuation of outside sounds. (Note that I recently switched from measuring at a level of 75dB to a level of 85dB. This doesn’t change the way the isolation curves look, but 85dB lets me get better measurements of noise-canceling headphones, which demand a lower noise floor.) Like the Audeze iSine10s, the M300s offer very little isolation from outside sounds; voices and other sounds will leak right in. However, the M300s do offer a little more isolation than typical over-ear, open-back, planar-magnetic headphones, such as the Monoprice M1060s included in this chart.

The M300s’ impedance magnitude and phase are about as flat as they could be, respectively at 26 ohms and a maximum phase shift of about +6° at 20kHz.

The sensitivity of the Monolith M300s, measured between 300Hz and 3kHz with a 1mW signal at the M300s’ specified impedance of 22 ohms, is 109.5dB. That’s high -- any conceivable source device should be able to drive the M300s to very loud levels.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 October 2017

I measured the HD 4.50 BTNCs using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, and a Musical Fidelity V-CAN amp. I used the Model 43AG’s original KB0065 simulated pinna for most measurements, as well as the new KB5000 pinna for certain measurements, as noted. For Bluetooth-sourced measurements, I used a Sony HWS-BTA2W Bluetooth transmitter to send signals from the Clio 10 FW to the headphones. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed.

The HD 4.50 BTNCs’ frequency response, taken with a Bluetooth signal with noise canceling (NC) on, may look a little weird due to its two prominent peaks centered at roughly 50Hz and 2.2kHz. Actually, it’s not far from what’s considered a standard “flat” headphone response: a broad bass peak, a midrange dip, a prominent response peak around 2.5kHz, and a lesser peak around 6kHz.

This chart shows the HD 4.50 BTNCs’ measured right-channel frequency response measured with the old KB0065 pinna (which I’ve used for years) and G.R.A.S.’s new KB5000 pinna, which I’ll be switching to because it more accurately reflects the structure and pliability of the human ear. I include this mostly for future reference rather than as something you should draw conclusions from; I intend to show both measurements in every review for at least the next year before I begin to use only the new pinna.

This chart shows the right-channel frequency response of the HD 4.50 BTNCs measured with Bluetooth and NC on, with Bluetooth on and NC off, and with a wired connection. Obviously, switching NC off substantially changes the sound of these headphones. Although I don’t show it in the chart here, I also measured the response of a wired connection, adding 70 ohms output impedance to the V-CAN amp’s 5-ohm native output impedance, but saw no notable change in response.

This chart shows the HD 4.50 BTNCs’ measured right-channel frequency response compared with those of three other NC headphones: the AKG N60 NC Wireless, the Bose QuietComfort 35, and the PSB M4U 2 (the last generally considered to rank among the best-sounding NC headphones). The HD 4.50 BTNCs’ response is similar to that of the AKG.

The spectral-decay (waterfall) chart shows a few very narrow, low-magnitude resonances between 1.5 and 2.1kHz; these are not likely to be audible.

The total harmonic distortion (THD) of the HD 4.50 BTNCs, measured with a wired connection because the Clio 10 FW’s sine sweeps can’t accommodate Bluetooth’s latency, is negligible at 90dBA. At 100dBA, it’s about average for dynamic over-ear headphones; it rises to 3% at 60Hz, and spikes from 3% at 30Hz to 13% at 20Hz -- but note that 100dBA is an extremely loud listening level, and that few music recordings have significant content below 30Hz.

In this chart, the external noise level is 85dB SPL; the numbers below that indicate the degree of attenuation of outside sounds. (I recently switched to measuring at a level of 85dB instead of 75dB; this doesn’t change the way the isolation curves look, but a level of 85dB allows me to get better measurements of NC headphones, which demand a lower noise floor.) In this measurement, the isolation of the HD 4.50 BTNCs looks somewhat below average for NC headphones, which surprises me because my subjective tests showed it to be better, and this measurement usually corresponds closely with subjective impressions.

The HD 4.50 BTNCs’ impedance magnitude and phase in wired mode are nearly flat, averaging about 23 ohms and with negligible phase shift.

The sensitivity of the HD 4.50 BTNCs in wired mode, measured between 300Hz and 3kHz with a 1mW signal and calculated for the specified 18 ohms impedance, is 100.9dB. This means that if you have to use a wired connection, the Sennheisers will still play plenty loud.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 15 September 2017

I measured the E55BT Quincy Editions using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, and a Musical Fidelity V-CAN headphone amplifier. On the G.R.A.S., I used the original KB0065 simulated pinna for most measurements, as well as the new KB5000 pinna for certain measurements, as noted. For Bluetooth-sourced measurements, I used a Sony HWS-BTA2W Bluetooth transmitter to send signals from the Clio 10 FW to the headphones. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed.

I had trouble getting consistent frequency-response measurements from the E55BT Quincy Editions. Their relatively small earcups (which don’t swivel very freely) and relatively small headband made it difficult to get a consistent fit and seal on the ear/cheek simulator. The measurements you see here are the results of many curves taken to find the best seal (indicated by the level of bass) and the most characteristic average response. Also, measurements using Bluetooth signals had to be gated to compensate for Bluetooth’s latency; this gating can affect the measurement curve.

The E55BT Quincy Editions’ frequency response, taken with a Bluetooth signal, looks basically textbook above 1kHz, with a strong peak at about 2.6kHz and a weaker one at about 6kHz; this is the response generally considered to best mimic the sound of real speakers in a real room. The dip in the lower midrange at around 300Hz is a little unusual.

This chart shows the E55BT Quincy Editions’ right-channel frequency response measured with the old KB0065 pinna (which I’ve used for years) and G.R.A.S.’s new KB5000 pinna, which I’ll be switching to because it more accurately reflects the structure and pliability of the human ear. I include this mostly for future reference, rather than as something you should draw conclusions from; I intend to show both measurements in every review for at least the next year before I begin using only the new pinna.

This chart shows the right-channel frequency response of the Quincy Editions, measured using pink noise fed via a Bluetooth signal and using a wired connection from a Samsung Galaxy S8 phone. The wired-connection response is almost identical, with just a couple dB more low-bass output. This is an admirable and, sadly, rare result for active headphones; most show substantially different response when used in passive, wired mode. Although I didn’t include the chart here, the wired connection produced no notable difference in response when I added 70 ohms additional output impedance to the V-CAN amp’s 5-ohm native output impedance.

This chart shows the right-channel frequency response of the E55BT Quincy Editions vs. the standard E55BTs, measured using pink noise fed via Bluetooth. There seem to be slight but consistent differences in the responses of the two models.

This chart shows the E55BT Quincy Editions’ measured right-channel frequency response compared with those of three closed-back headphones: NAD Viso HP50, Bose QC35, and Sony MDR-7506. Except for the Quincy Editions’ deep midrange dip at about 300Hz, their response seems well within the norm.

The spectral-decay (waterfall) chart shows a very clean response free of troublesome resonances.

The total harmonic distortion (THD) of the E55BT Quincy Editions -- measured with a wired connection because the Clio 10 FW’s sine sweeps can’t accommodate Bluetooth’s latency -- is higher than average below 100Hz, rising to maximums of 4.7% at 90dBA and 13.5% at 100dBA. However, these levels, especially 100dBA, are much louder than most listeners will want to -- or should -- use, and I heard no distortion when listening to the headphones.

In this chart, the external noise level is 85dB SPL; the numbers below that indicate the degree of attenuation of outside sounds. (Note that I recently switched to measuring at a level of 85dB instead of 75dB; this doesn’t change the way the isolation curves look, but an 85dB level lets me get better measurements of noise-canceling headphones, which demand a lower noise floor.) In my measurements, the isolation of the E55BT Quincy Editions is relatively poor. This is surely due in part to the difficulty I had in getting a good seal on the ear/cheek simulator, but it also corresponds with my subjective listening impressions.

The E55BT Quincy Editions’ impedance magnitude and phase in wired mode are almost flat, with an average of 36 ohms through most of the audioband, and a low maximum phase shift of about +20°.

The sensitivity of the E55BT Quincy Editions in wired mode, measured between 300Hz and 3kHz with a 1mW signal, is 103.8dB. If the battery runs down and you have to switch to a wired connection, the JBLs will still play plenty loud.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 15 August 2017

I measured the Wave 5s using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, a Musical Fidelity V-CAN amp for most measurements, and an Audio-gd NFB-1AMP amplifier for distortion measurements. On the G.R.A.S. Model 43AG, I used the original KB0066 simulated pinna for most measurements, as well as the new KB5000 pinna for certain measurements (as noted). These are “flat” measurements; no diffuse-field or free-field compensation curve was employed.

The Wave 5s’ frequency response is a little unusual, but not crazy unusual. Normally we see a pronounced peak around 2.5kHz, but here it’s a gradual peak rising from about 500Hz. Note also the strong peak at 6kHz. It’s common to see a second treble peak, but usually it’s about 6dB lower than the peak at around 2.5kHz. This is probably the cause of the occasional mid-treble brightness I heard. Note also the hashiness between 500Hz and 2kHz; I’ve seen this in some other planar magnetics, but I can’t recall seeing it so pronounced.

This chart shows the Wave 5s’ measured right-channel frequency response, measured with the old KB0066 pinna (which I’ve used for years) as well as with G.R.A.S.’s new KB5000 pinna, which I’ll be switching to because it more accurately reflects the structure and pliability of the human ear. I’m including this mostly for future reference rather than as something you should draw conclusions from; I intend to show both measurements in every review for at least the next year, before beginning to use only the new pinna.

This chart shows the results of adding 70 ohms output impedance to the V-CAN’s 1 ohm, to simulate the effects of using a typical low-quality headphone amp. There’s no audible difference in the response, which means the Wave 5s’ tonal balance won’t change with a change in amp.

This chart shows the Wave 5s’ measured right-channel frequency response compared with some other high-end open-back headphones. The Wave 5s’ response looks, at first glance, flatter than the others above 100Hz, but in general, a headphone that sounds subjectively flat will have more of a bass hump, plus a strong peak at about 2.5kHz and a less-strong peak around 6kHz. The Tidal Forces also seem to have the most restrained bass response of all the models represented here.

The spectral-decay (waterfall) chart shows a strong resonance at about 400Hz, and another at that frequency’s second harmonic, 800Hz. Above 1kHz are some more very narrow, “hashy” resonances, but this is fairly common for open-back planar-magnetic models.

The Wave 5s’ total harmonic distortion (THD) is slightly high, at 1% to 3% below 1kHz, but I doubt that such a level of THD in a transducer would be readily audible. Unusually, raising the level to 100dBA -- which is extremely loud -- had a minimal effect on the distortion measurement.

In this chart, the external noise level is 75dB SPL; the numbers below that indicate the degree of attenuation of outside sounds. (Note: I took this measurement before I switched to measuring at a level of 85dB instead of 75dB. That doesn’t change the way the isolation curves look, but an 85dB level lets me get better measurements of noise-canceling headphones, which demand a lower noise floor.) The Wave 5s offer a tiny bit more isolation than typical open-backs, but not much compared to the closed-back and noise-canceling models also included in this chart.

As with most planar-magnetic headphones, the Wave 5s’ impedance magnitude and phase are dead flat, in this case at 31 ohms.

The sensitivity of the Tidal Force Wave 5s, measured between 300Hz and 3kHz with a 1mW signal, is 103.2dB. That’s excellent for planar-magnetic headphones, and enough to ensure that the Wave 5s will play loudly from any source device.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 15 August 2017

I measured the Monoprice M1060s using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, a Musical Fidelity V-CAN amp for most measurements, and an Audio-gd NFB-1AMP amplifier for the distortion measurements. On the G.R.A.S. 43AG I used the original KB0065 simulated pinna for most measurements, as well as the new KB5000 pinna for certain measurements (as noted). These are “flat” measurements; no diffuse-field or free-field compensation curve was employed.

The M1060s’ frequency response is pretty standard for an open-back planar-magnetic model. I noted two anomalies. First, the response between 8 and 16kHz is lower than I’m used to seeing. For reasons discussed here, above about 8kHz, headphone frequency-response measurements are less reliable, but this measurement does jibe with what I heard in my listening tests. Second, there’s an apparent mismatch between the right and left channels between 1 and 4kHz. This is the best match I was able to get after many attempts with both the Monoprices’ left and right earpieces.

This chart shows the M1060s’ measured right-channel frequency response measured with the old KB0065 pinna (which I’ve used for years) and G.R.A.S.’s new KB5000 pinna. I’ll soon be switching to the new pinna, because it more accurately reflects the structure and pliability of the human ear. I include this mostly for future reference rather than as something you should draw conclusions from; I intend to show both measurements in every review for at least the next year before I begin using only the new pinna.

This chart shows the results of adding 70 ohms output impedance to the V-CAN’s 1-ohm output impedance, to simulate the effects of using a typical low-quality headphone amp. The difference is negligible; the M1060s’ tonal balance won’t change depending on the amp used.

This chart shows the M1060s’ measured right-channel frequency response compared with two similar $299, open-back, planar-magnetic headphone models (the Tidal Force Wave 5 and HiFiMan HE400S), plus the NAD Viso HP50, a conventional closed-back model. The M1060s have a flatter measured response than the other planar-magnetics, with more bass -- an effect that could be due to a better seal of the M1060s’ plush faux-leather pads on the ear/cheek simulator. Again, you can see that the M1060s’ response between 8 and 16kHz is comparatively soft.

The spectral-decay (waterfall) chart shows low resonance in the bass, scattered narrow resonances in the midrange, and a strong series of resonances centered at 4.5kHz. The last, while scary looking, are not surprising for an open-back planar-magnetic model, most of which show a similar resonance pattern somewhere in the upper midrange/lower treble.

The M1060s’ total harmonic distortion (THD) is about as low as I’ve measured. Planar-magnetic headphones usually do well on this test, but this is excellent even when compared to most of the other planar-magnetics I’ve measured.

In this chart, the external noise level is 85dB SPL; the numbers below that indicate the attenuation of outside sounds. (Note that I recently switched to measuring at a level of 85dB instead of 75dB; this doesn’t change the way the isolation curves look, but a level of 85dB allows me to get better measurements of noise-canceling headphones, which demand a lower noise floor.) The M1060s offer about the same isolation as most open-back models: almost none.

The M1060s’ impedance magnitude and phase are almost perfectly flat right at the rated 50 ohms, except for a little bump to 57 ohms at 4.4kHz.

The sensitivity of the M1060s, measured between 300Hz and 3kHz with a 1mW signal, is 100.5dB. That’s fairly high for planar-magnetic headphones, and means that the M1060s should deliver adequate volume from any source device.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 July 2017

I measured the Susvaras using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, and an Audio-gd NFB-1AMP amplifier. On the Model 43AG, I used the original KB0066 simulated pinna for most measurements as well as the new KB5000 pinna for certain measurements, as noted. These are “flat” measurements; no diffuse-field or free-field compensation curve was employed.

I see nothing out of the ordinary in the Susvaras’ frequency response; it’s a lot like what I’ve measured from other open-back, planar-magnetic headphones. Although the response of the two channels doesn’t seem to match quite perfectly, headphone measurements are susceptible to slight changes in headphone position on the ear/cheek simulator. This was the best match I could get.

This chart shows the Susvaras’ measured right-channel frequency response, measured with the old KB0066 pinna (which I’ve used for years) and G.R.A.S.’s new KB5000 pinna, which I’ll be switching to because it more accurately reflects the structure and pliability of the human ear. I’m including this mostly for future reference rather than as something you should draw conclusions from; I intend to show both measurements in every review for at least the next year, before I switch to using only the new pinna.

This chart shows the results of adding 70 ohms output impedance to the NFB-1AMP’s 1-ohm output impedance, to simulate the effects of using a typical low-quality headphone amp. There’s no difference in the response. This doesn’t mean you can use a low-quality amp with the Susvaras -- they won’t play loud enough -- but it does mean you can expect good sound even with a tube headphone amp that has a high output impedance.

This chart shows the Susvaras’ measured right-channel frequency response compared with some other high-end, open-back headphones. The Susvaras’ response is largely similar to the others, although it appears to have the flattest response of the bunch.

The spectral-decay (waterfall) chart shows a lot of strong resonances across the audioband. Though this isn’t uncommon with open-back, planar-magnetic headphones, these might be the most I’ve ever measured. But there are enough of them, and they’re so well spread out, that it’s unlikely that any of them will stand out other than the multiple narrow resonances centered at 4kHz. However, those correspond with the Susvaras’ peak in response at that frequency, a characteristic observed in the measurements of most headphones.

The total harmonic distortion (THD) of the Susvaras is very low at 90dBA, and at 100dBA it’s only slightly higher, in the band of 1-2kHz. Interestingly, the tiny distortion peaks seem to correspond with some of the resonances in the spectral-decay plot.

In this chart, the external noise level is 85dB SPL; the numbers below that indicate the attenuation of outside sounds. (Note that I’ve switched to measuring at a level of 85dB instead of my previous 75dB; this doesn’t change the way the isolation curves look, but it does let me get better measurements of noise-canceling headphones, which demand a lower noise floor.) Like almost all other open-back, planar-magnetic headphones, the Susvaras offer little more isolation than draping a piece of thin fabric over your ear. But that’s OK -- if their isolation were any better, they couldn’t sound the way they do. For comparison, I’ve included the isolation charts of the NAD Viso HP50 closed-back and Bose QC25 noise-canceling headphones.

As with most planar-magnetic headphones, the Susvaras’ impedance magnitude and phase are effectively dead flat. The measured impedance is 62.5 ohms across almost the entire audioband.

The sensitivity of the Susvaras, measured between 300Hz and 3kHz with a 1mW signal, is 80.6dB. Compare this to 86.9dB for the HE1000 V2s, 79.6dB for the classic HE6es, and about 100dB for a set of typical headphones.

. . . Brent Butterworth
brentb@soundstagenetwork.com

Details

Parent Category: Products

Category: Headphone Measurements

Created: 01 June 2017

I measured the Libratone Q Adapts using a G.R.A.S. Model 43AG ear/cheek simulator, a Clio 10 FW audio analyzer, a laptop computer running TrueRTA software with an M-Audio MobilePre USB audio interface, a Musical Fidelity V-Can headphone amplifier, an Audio-gd NFB-1AMP amplifier for distortion measurements, and a Sony HWS-BTA2W Bluetooth transmitter/receiver. In some cases, the inherent latency of Bluetooth necessitated a change in measurement technique, to using pink noise as a stimulus rather than the swept logarithmic “chirp” tone I normally use. This is a “flat” measurement; no diffuse-field or free-field compensation curve was employed.

The response curve seen here, measured with the Q Adapts in Bluetooth mode with noise canceling on, isn’t quite as anomalous as it might seem at first glance. The large peak in the 2.5kHz frequency is fairly typical of most of the headphones and earphones I measure. There are a few dB more bass relative to that peak than I’m used to measuring, and a little less energy between 3 and 10kHz. There does seem to be a bit of a mismatch in the bass response between the left and right channels, probably due to different internal acoustics in the two earpieces (which is fairly common in active headphones), but it’s unlikely to be audible.

This chart shows the response in Bluetooth mode with NC set to Level 4, compared with a passive connection using the included cable. It’s not a huge difference compared with what I’ve measured in some other active headphones, but it indicates that the sound will be a bit different: more bass kick in Bluetooth, more upper bass (and perhaps a bit of a bloated sound) in wired mode.

This chart shows the response in Bluetooth mode with NC off (Level 2) and on (Level 4). This is a subtler difference than I usually measure in this test; many headphones sound completely different with NC on, but the Q Adapts will sound pretty much the same either way. The measured response with Level 1 (open mike mode) was essentially the same as with Level 2, and the response with Level 3 was essentially the same as with Level 4.

This chart shows the Q Adapts’ measured right-channel frequency response compared with Bluetooth NC headphones (Bose QC35s), wired NC ’phones (AKG N60 NCs), and conventional closed-back ’phones (NAD Viso HP50). The Q Adapts have a more pronounced response in both the bass and treble, which suggests that they’ll sound fairly balanced yet rather vivid, with perhaps some deemphasis of the midrange.

The spectral-decay (waterfall) chart shows a little bit more resonance below 500Hz than I’m used to seeing, but all the resonances are extremely well damped and die out within a few milliseconds.

The total harmonic distortion (THD) of the Q Adapts (which I had to measure using a wired connection because of Bluetooth’s latency) is low for compact headphones such as these, hitting about 1% at 20Hz at 90dBA, and 3% at the extremely loud listening level of 100dBA.

In this chart, the external noise level is 75dB SPL; the numbers below that indicate the degree of attenuation of outside sounds. For comparison, I’ve included the isolation plots of two other noise-canceling headphones: the Bose QC25s and AKG N60 NCs (which, like the Q Adapts, are both on-ear headphones, which in my experience can’t match the isolation of comparable over-ear models). The Q Adapts beat out the N60 NCs in the midrange around 1kHz, but don’t cancel bass frequencies as well. They might do a little better at rejecting the noise of conversations or an airliner’s air-circulation system, but probably not quite as well at reducing the noise of jet engines.

Using a wired connection, the electrical impedance of the Q Adapts is basically flat at 25 ohms, and the electrical phase is similarly flat.

The Q Adapts’ sensitivity, measured between 300Hz and 3kHz with a 1mW signal, and calculated for 32 ohms impedance (my “unless otherwise specified” test impedance for active models with passive input), is 100.2dB. That means that, unlike some active headphones, the Q Adapts will play plenty loud enough when connected with a cable to the average smartphone.

. . . Brent Butterworth
brentb@soundstagenetwork.com