Etymotic ER2 vs balanced armature

[Chyżwar]

If we compare them to ER2 they are simply inferior.

My ER4XR has a THD of 0.70% (1kHz 100dB SPL) which is inaudible to me and I have not seen any scientific evidence that anyone can hear such small distortion while listening to music.
I don't think the ER2 are better headphones, they're just a little different.

 

[MCH]

Not knowing (and not really caring) about technical details, this was my experience as a consumer with this family of headphones:
Wanted to buy the etys because of the supposedly isolation and good quality for the price. However i also felt intrigued about the BA vs dynamic drivers and of course if the huge (for me) price difference is justified.
After reading here and there, specially how much @JohnYang1997 likes the ER2 and after visiting websites of BA producers (from where i got the feeling that their real advantage is less power needs and smaller size so that there is more space for batteries - who cares, here we have cables), i started to think that maybe for the casual user like me all the marketing stuff, the individually matched in the USA story was more marketing BS than a real 200 euros worth difference... maybe not, but anyways, it wasn't convincing enough for me. So eventually ordered my ER2 to amazon US for 90 euros delivered to europe with taxes included and i am a happy man since then. I really doubt i could ever distinguish them from the ER3 or ER4.

 

[HRTF_Enthusiast]

Every comparison I see with ER2SE and ER4SR inserted deep enough has ER2SE with better extension

Crinacle does not insert deep enough. The quarter wave resonance will be past 10 kHz when inserted deep enough
(34300 cm/s)/(x Hz) *1/4 = (y cm).
View attachment 169889
ER2SE is <0.1% THD at 80 dB while ER4SR is close to 1% and dominated by 3rd order harmonics.
View attachment 169891

It becomes closer at 94 dB, but ER2SE is dominated by 2nd order harmonics.
View attachment 169893

ER2SE vs ER4SR (https://www.hypethesonics.com/iemdbc/). Also, this isn't even inserted as deep as banbeu's graph.

 

[Fregly]

I can only add my subjective impression. I have both and prefer the E2. It does seem to sound different. Certainly to my ears it does not sound less clear or not as refined, however you would put it. Cheap and great.

 

[Graph Feppar]

My ER4XR has a THD of 0.70% (1kHz 100dB SPL) which is inaudible to me and I have not seen any scientific evidence that anyone can hear such small distortion while listening to music.
I don't think the ER2 are better headphones, they're just a little different.

Can you post your measurements here? What microphone and coupler do you have?

 

[2M2B]

Ok, here is all the measurements I found.

Solder Dude measurements of ER2SE, ER2XR and ER4XR with aftermarket ear tips Comply Ts 100 and Spin Fit CP800.
View attachment 170321View attachment 170322View attachment 170323

Notice that the first graph, the ER2SE has smaller size font top and botton and it has wider gaps between lines so it looks different at first look but if you carefully compare you will notice how very similiar all these measurements are. The differenced between cheap ER2 and expensive ER4 are tiny, easily nulified by simple EQ and the dynamic driver ER2 has superior high frequency extension.

View attachment 170326
Impulse response with stock tips, again, very similiar.

View attachment 170329
THD at low SPL of only 80 dB. Notice how the expensive ER4 has much higher distortion and whats worse, unline most classic dynamic drivers like Sennheiser HD600 where distortion is only problems at low frequencies, the balanced armature driver has peak distortion in mid range around 800Hz, this means the distortion overtone will fall on the most ear sensitive frequency band around 3Khz.

Notice, this is also at very low level, I listen at 87-90 dB peak for compressed music and dynamic less compressed music easily goes above 100db. Now if it has 0.6% THD at 80 dB, imagine what 100 dB peaks from orchestra will do to this poor little driver.

View attachment 170337
The ER2SE has less than 0.5% peak THD at 94 dB vs ER4XR which has over 0.6% at just 80 dB

View attachment 170338
ER2SE distortion spectrum at 94 dB, notice most of the distortion is of the less offensive second harmonic even order type, I will mention it later, remember it for now.

View attachment 170339
SilverEars frequency response measurement of ER2SE and ER4XR, the differences bellow 8KHz are again very small, if this was apples to apples comparsion of ER2SE vs ER4SR they would be almost identical. In highs we can see ER2SE is flatter and more extended.

View attachment 170340
yuriv Distortion spectrum of older ER4PT model at 94 dB. Please note his microphone sucks and has high distortion so this graph in isolation doesnt mean much.

Here are his other measurements of LG Quadbeats 3, a headphone with super low distortion and Sony MH755, a low distortion headphone too.
View attachment 170341
View attachment 170342
The distortion you see here is the distortion of the microphone, thats why these two low distortion headphones look the same. In contrast the high distortion ER4PT reaches far above the distortion floor of this microphone and we can see that most of its distortion is the nasty third harmonic odd order type, think hard clipping of transistors vs soft clipping of vacuum tubes.

View attachment 170343
yuriv Distortion spectrum of ER2SE, surprisingly it doesnt measure much lower THD as ER4PT but I take this result with grain of salt due to the shit microphone used. Notice, the dynamic driver of ER2SE has mostly second harmonic even order distortion.

View attachment 170344View attachment 170345
Left and right channel distortion spectrum of ER2SE at 94 dB @ 500 Hz by hakuzen. Just to hammer the point of the ER2 having mostly even order distortion.

View attachment 170350Crinacle ER2SE vs ER4SR, a apples to apples comparsion. These two are so similiar below 3 KHz that you will find bigger differences between different units of same model of headphone from other manufacturers due to random variability. Above 3 KHz the differences are still small, no night and day difference as some claim, cheaper ER2SE is smoother and more extended.

View attachment 170351
Crinacle ER2XR vs ER4XR, same as above, just that ER2XR has 6 dB more of sub bass.

View attachment 170353
View attachment 170354
Left and right channel distortion of ER4SR from purr1n at SBAF. SPL is little unclear but this should be somewhere around 85 dB at 500Hz. This is better result that the Solder Dudes measurements, third harmonic dominates to 1.6 KHz then its mostly second harmonic.

View attachment 170356View attachment 170357
ClarityFidelity distortion spectrum of ER4SR at 94 dB @ 500 Hz, left and right channel. We can see its mostly third harmonic up to 1.5 KHz after which it changes with second harmonic back and forth on right channel while on left the second harmonic dominates always past the switch point.

View attachment 170359
View attachment 170360
Same as above just ER4XR, it has much higher distortion and its mostly concentrated around 800 Hz, same second and third harmonic behaviour as ER4SR.

View attachment 170362
View attachment 170363
ClarityFidelity again, this time with older model of dynamic driver Etymotic headphone, the MC5. This one has the Kids max SPL cock blocking filter so little Timmy doesnt obliterate his hair cells while headbanging to Slayer at 110 dB. Notice the much lower distortion compared to the balanced armature ER4SR and XR. There is high distortion peak in left channel at 4 KHz but that can be damaged headphone or measurement error, the distortion is strongly dominated by even order second harmonic.



View attachment 170385
View attachment 170386
View attachment 170387

ER4SR, XR and MC5 waterfall plots. I read reports from subjectivists saying the balanced armature Etymotics have more "speed" than the "slow" dynamic driver, what does "speed" or "fast" even mean? Becose looking at impulse responses and waterfall plots shows that the balanced armature doesnt perform better in the speed that sound decays and if by "speed" is meant literal velocity of the moving membrane, then the dynamic driver is faster as it goes past 19 KHz while the BA drops off the cliff past 16 KHz. Please note the above is the old kids MC5, it has even poorer HF extension that the ER4, it goes to like 15 KHz. I dont have waterfall for ER2SE but I expect it to perform very similiary to the above in terms of decay speed but with extension that goes past 19 KHz as seen in the plots showed earlier.


And best for last, the good old Innerfidelity measurements from times when our lord and saviour, Tyll Hertsens, was still active and Innerfidelity was actually worth visiting.
View attachment 170367
View attachment 170368
ER4SR and ER4XR, the midrange distortion is bad in SR but pretty tolerable outside the 8 KHz peak, the XR in comparsion has not only higher midrange distortion peak, but the area under the peak, the lower midrange and bass region is significantly elevated too. Above the peak around the treble they are similiar. If you care about distortion I think at this point its safe to say the SR is strongly superior to the XR.

View attachment 170369

MC3, a old dynamic driver Etymotic, a predecessor of the modern day ER2, look at the distortion, it completly shits on the much more expensive ER4 series. Also notice how clean the square waves and impulse response are.

TLDR: The balanced armature Etymotic headphones are extremly similiar to the dynamic driver ones in terms of frequency response, decay and bass extension but have higher distortion, especially in the worst place, the mids, and their distortion is the nasty odd harmonic type while the dynamic driver ones have the warm and smooth even order. Also for younger users, the BA Etys turn all music into sort of MP3 due to their high frequency extension which falls into black hole past 16 K. There is not a single measurable advantage of the BA Etys and they cost alot more. The ER2SE and XR are best measuring headphones Etymotic ever made.

Are you implying that Etymotic are lying about there THD numbers on the data sheets, Since I've seen people with ER4XR's at <0.4% at 1KHz. At 94db the ER2XR/ER2SE reach 0.7% at 1KHz with some probably reaching 1.2% since QC not the same with the ER4SR/ER4XR.

You left quite a bit out on the Head Fi THD data for the Etymotic ER4PT where they were testing the Coupler mic since they noticed the high 2nd order on the ER4 at 94db could be the microphone driver clipping hard, It wasn't a actual ER4P test. They said if a TOTL mic was used instead the THD number would be like the 80db test or maybe lower than that.

I've noticed few of those ER4XR/ER4SR THD tests don't compare to the Etymotic data sheet that you get.

 

[Dealux]

There is no such a thing as "BA timbre" or any other driver specific timbre. Objectively speaking, the difference in sound between the ER2 and ER4 is somewhat subtle but...

The dynamic driver in the ER2XR is definitely by far the most capable at producing bass lol. Even the bassy ER3XR sounds thin by comparison. In the mids and highs, the ER2 (especially the warmer XR) comes across as the more correct sounding Etymotic. As far as I can tell this is down to two things. The ER4 seems to roll off from 3K a bit too strongly with some energy lacking at 5-8K vs the Etymotic target (which already leans a bit warm) but the treble response peaks again at 9K. This isn't the best thing ever because it adds some sibilance to the already slightly honky sounding mids at 1-2K. Couple that with the bass being limp and you have a fairly thin sounding IEM. Yes, even the ER4XR sounds thin vs the ER2XR.

The ER2XR has just enough 5-8K to my ears (percussion instruments don't sound as thin as the ER4) and the upper treble peak is at 10 or 11K instead. A bit higher and thus less annoying. It actually kinda adds some nice sparkle to overtones. There's still a bit of imbalance between the treble and upper treble which results in some thinness, but this is mitigated by the punchier and more extended bass on the ER2XR.

To my ears, the ER2XR is almost perfect tuning wise and has superb clarity and timbre quality. There is a slight lack of air (upper treble) because they technically still dip in the treble after 11K which makes the high upper treble sound a bit hollow and soft (especially noticeable on classical and orchestral). But honestly detail wise they're up there with the HD800 for me. Just not as wide obviously but tonally more clear and correct.

 

[2M2B]

I seriously love accounts pouring in telling people why to avoid balanced armature tech by hiding under poorly done THD data or subjectivist terms with zero backing.

But I'm the one who gets bashed as ER4XR fanboy?.


Cause the Bass levels in all Etymotic IEM's are:

ER2XR = 7db

Evo = 7.5db

ER3XR = 5db

ER4XR = 3.5db

ER3SE/ER4PT = 1db

ER4SR = 0db

ER4B = -1db

The ER3XR is the bassiest Single BA on the market yet all I see is how It thin, Either folk are lying about getting a true seal or just mad about BA drivers being super popular in IEM's for a cringy reason. Because I find the ER4SR when EQ'd to mimic 2XR bass blows the 2XR out the water.

Funny how no one even does a A/B with a ER2SE & ER4SR to see if it just bass quantity there hearing than a DD being magically better than a BA driver.

 

[Dealux]

Those BA drivers didn't respond well to EQ in my experience. I've also ran into some audible distortion whilst playing around with EQ. The ER3XR is definitely thinner in the bass by far (lacks bass extension) with the full 98% seal. It's also more sibilant. So is the ER4.

 

[solderdude]

Solder Dude measurements of ER2SE, ER2XR and ER4XR

Well.... not this solderdude

The plots seem to come from head-fier csglinux.

 

[Graph Feppar]

Well.... not this solderdude

The plots seem to come from head-fier csglinux.

Wait, you arent csglinux?

 

[solderdude]

No... but I do use Linux Mint so there is some Linux in my life.

 

[Graph Feppar]

No... but I do use Linux Mint so there is some Linux in my life.

He even has penguin in his avatar like you! How many high post count measurement posters are there in headphone forums with penguin avatar? I thought for sure it must be you

 

[solderdude]

He does but a different penguin... I have never been a head-fi member nor do I inspire to be

I can only be found on ASR and DIYAH and don't like, nor measure, IEMs. I hate showing things in my ear, regardless of sound quality.
Used to be on Rock Grotto, Hoved-Fi and Changstar (in the hope I could bring some objectisvm in there)... in that order.

 

[Earfonia]

There is no such a thing as "BA timbre" or any other driver specific timbre. Objectively speaking, the difference in sound between the ER2 and ER4 is somewhat subtle but...

The dynamic driver in the ER2XR is definitely by far the most capable at producing bass lol. Even the bassy ER3XR sounds thin by comparison. In the mids and highs, the ER2 (especially the warmer XR) comes across as the more correct sounding Etymotic. As far as I can tell this is down to two things. The ER4 seems to roll off from 3K a bit too strongly with some energy lacking at 5-8K vs the Etymotic target (which already leans a bit warm) but the treble response peaks again at 9K. This isn't the best thing ever because it adds some sibilance to the already slightly honky sounding mids at 1-2K. Couple that with the bass being limp and you have a fairly thin sounding IEM. Yes, even the ER4XR sounds thin vs the ER2XR.

The ER2XR has just enough 5-8K to my ears (percussion instruments don't sound as thin as the ER4) and the upper treble peak is at 10 or 11K instead. A bit higher and thus less annoying. It actually kinda adds some nice sparkle to overtones. There's still a bit of imbalance between the treble and upper treble which results in some thinness, but this is mitigated by the punchier and more extended bass on the ER2XR.

To my ears, the ER2XR is almost perfect tuning wise and has superb clarity and timbre quality. There is a slight lack of air (upper treble) because they technically still dip in the treble after 11K which makes the high upper treble sound a bit hollow and soft (especially noticeable on classical and orchestral). But honestly detail wise they're up there with the HD800 for me. Just not as wide obviously but tonally more clear and correct.

ER2XR is my favorite Etymotic. I recently created an EQ profile for ER2XR to match my personal target curve (I called it EITC-2021). I use REW to create the EQ profile than exported it as text to be loaded into Equalizer APO:

Below is my measurement of my ER2XR compared to my EITC-2021 target curve, and the main differences between the 2:

If you use the Equalizer APO, kindly please give it a try, I would love to get some feedback on my EITC-2021 target curve. Please apply -2.5 to -3.0 dB gain compensation to avoid clipping. Thanks!
The equalizer text file is attached here. Or you can just create a new text file with the following lines:

Filter Settings file

Room EQ V5.20.4
Dated: 5 Jan, 2022 12:13:08 AM

Notes:Etymotic ER2XR v1.0

Equaliser: Generic
A over B
Filter 1: ON LS Fc 37.00 Hz Gain -0.50 dB
Filter 2: ON LS Fc 135.0 Hz Gain 2.00 dB
Filter 3: ON PK Fc 445.0 Hz Gain -0.90 dB Q 0.850
Filter 4: ON PK Fc 880.0 Hz Gain 0.30 dB Q 2.000
Filter 5: ON PK Fc 1315 Hz Gain -1.70 dB Q 1.150
Filter 6: ON PK Fc 2423 Hz Gain -1.70 dB Q 1.850
Filter 7: ON PK Fc 4585 Hz Gain 2.70 dB Q 3.150
Filter 8: ON HS Fc 8000 Hz Gain 1.50 dB

 

[Dealux]

I'm afraid that by using these IEMs with third party tips and the 1000 ohm filter, the tonality differs a lot from your stock ER2XR.

With the Shure foam tips and the filter I get more 5-8K for instance.

 

[Earfonia]

That 80db ER4SE seems like faulty measurement, it shows distortion decreasing at 94db vs at 80, that is physicaly impossible. Who did that test and how was it done?

You're right. The 80 dB SPL measurement seems was limited by the noise floor of the environment ambient noise. When measuring distortion it is crucial that the environment noise is kept as low as possible. At 80 dB SPL, measuring 0.1% THD means the harmonic level is 60 dB lower than the fundamental, or 20 dB SPL if the 80 dB SPL is set at 500 Hz. To get 20 dB SPL ambient noise at home or office is practically impossible without noise isolation container. When using REW, it will indicate if the harmonics are masked by the noise floor.

When looking at distortion measurement, I will ask 4 things:

1. What the sample rate that is used for measurement? I know HypeTheSonics do their measurement at 96kHz sample rate. So we get up to 4th harmonic at 10kHz, and up to 6th harmonic at 7.1 kHz. I use the same method as well and do my measurement at 96kHz ADC sample rate. When the distortion measurement is done at 44.1kHz or 48kHz sample, at 10kHz we only get up to the 2nd harmonic and at 7kHz only the 3rd harmonic. That makes the distortion measurement not very useful as the higher order harmonics are what usually the important factors in THD observation.

2. What is the noise floor and ambient noise level around the measurement setup? As mentioned earlier, very low ambient noise level is crucial for distortion measurement. We don't want those low level higher order harmonics to be masked by the noise floor that makes the measurement result useless. I placed all my measurement setup in noise isolation container for all measurement.

3. Is the measurement setup SPL reading calibrated? SPL calibration is crucial for distortion measurement.

4. What is the measurement methodology? Is it 'Continuous Sweep' or 'Stepped Frequency Sweep' ? Stepped is the more accurate method but lower in frequency resolution as it usually takes a lot longer than sweep. So if we use very high number of frequency points per octave the measurement will take a long time to be completed. I use the same settings as HypeTheSonics for distortion measurement, Stepped Frequency Sweep, 6 PPO. Takes about 5 minutes to complete a single measurement.

IMHO, the measurement methodology is as important as the measurement result. Only getting the measurement result alone without knowing the methodology can be misleading.

Here is my distortion measurement of the Etymotic ER2XR at 104, 94, and 80 dB SPL. I don't usually do at 80 dB SPL because there is no point measuring distortion at that level. We are more interested to see how low the distortion at higher SPL not at lower SPL. Both Left and Right channel of my ER2XR are showing similar THD profile, so below are the right channel measurement only. The Brown graph is the noise floor level so we can see the noise floor level in comparison to the higher order harmonics (> 5th). We can see that at 80 dB SPL most of higher order harmonics are already below the noise floor, making the measurement result at this level inaccurate.

Etymotic ER2XR - Right - Distortion measurement at 104 dB SPL @ 500Hz:

Etymotic ER2XR - Right - Distortion measurement at 94 dB SPL @ 500Hz:

Etymotic ER2XR - Right - Distortion measurement at 84 dB SPL @ 500Hz:

All measurement done inside noise isolation container. Here is the comparison of ambient noise inside the container when the container lid was opened (RED graph) and closed (BLACK graph). Measured using MiniDSP UMIK-2 measurement mic.

More info about my measurement methodology here:
Earfonia IEM Measurement Setup & Methodology

 

[2M2B]

Also he ignoring a core point I'm making you can't just say all Etymotic ER4's are 0.9 ~ 1.8% since Unit variation is a thing. I posted a photo way back here when people said the ER2XR was <0.2% at 1KHz but when people did 94db tests the 2XR/2SE were 0.7% ~ 1% like most ER4XR's. That my ER4SR was <0.3% at 1KHz when tested at 97db.

 

[notexactly]

I've seen it mentioned that each pair of ER4SRs are driver matched to ~1dB between 100Hz and 10kHz and that below 100Hz there can be driver imbalances, as seen in these measurements: https://clarityfidelity.blogspot.com/2016/09/etymotic-research-er4sr-iem.html

Here, ER2XRs appear very well matched: https://audiosciencereview.com/forum/index.php?threads/etymotic-er2xr-review-iem.29818/

Is this a reason to go with dynamic drivers? Or is the ER4SR measured at clarityfidelity just a poor example of the headphone?

I notice that Crinacle's measurements show the ER4SR with slightly better driver matching in the <100Hz region than the ER2SE and slightly worse in the treble.

As for whether any of these differences are audible...

 

[DivineCurrent]

Speaking of audibility of the BA driver distortion, I did an experiment with the Klipsch X10 a while back when I still had it, which is also a single BA driver (and actually has a bit less distortion than the ER4 series according to Innerfidelity graphs). The high third harmonic distortion is audible only when doing a sine sweep in the 1-3kHz area. You can hear some resonance not part of the sine tone (in both drivers so not broken), and you can actually replicate this in Audacity or SineGen by playing a sine wave at 1khz and then playing another 3kHz sine wave. It doesn't take much volume for the 3kHz sine wave to intrude on the 1kHz wave, I can start to hear it pretty easily at about -45 dB when the 1kHz tone is at full volume 0 db. -45 dB is 0.56% distortion according to the distortion calculation, which lines right up with the amount of distortion the ER4XR and ER4SR have at around 1kHz. I did not measure the decibel level of my test, but it was at a comfortable listening level, nothing too loud.

Now, I am very well aware that this is only third harmonic distortion, and I ignored THD with every other order of distortion that could potentially be audible. And I am also aware this is probably not audible with real world music. But at the very least we already know that at least third harmonic distortion can be audible under some circumstances even under 1% with sine sweeps. By the way, I own an ER2SE, and with the same sine sweep experiment, I do not hear any third order distortion at all throughout the whole frequency range.