• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Etymotic ER2 vs balanced armature

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
I searched whole internet for measurements and everything seems to point to the ER2 being superior to the ER3 and ER4. The frequency response between them are very small and it would be trivial easy to EQ ER2 to have frequency response as ER4. The distortion is far superior on the dynamic driver of ER2 and it seems to even have better extension past 17KHz.

So why pay more for inferior headphone? Like I said before, the small subjective differences of " resolution, smearing, clarity, speed, bla bla bla" should be easily erased with EQ where you set the Q factor and precise frequency of the bands. The supreme matching of drivers of ER4 seems not worth the price increase since the ER2 and 3 seem to be perfectly matched so only advantage seems to be bigger variety of ear tips in the box.

I dont believe in magic, I believe in distortion, frequency response and phase response. I found nothing that would explain what advantage balanced armature ER3 and 4 has over the ER2 and why would anybody pay so much more for it other than slight out of the box frequency curve differences. Am I missing something?
 
Last edited:

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
Do you have data to back each of those points up? Here are Crinacle's measurements:

graph.png
 
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
Do you have data to back each of those points up? Here are Crinacle's measurements:

View attachment 169847
The data is on the internet. Do you want me to collect all measurements I found and dump them all here? I can do that if you like, we can sperg together about the graphs which is activity I enjoy alot.

The graph you posted proves everything I wrote. Nearly identical frequency response + the cheaper ER2 has better treble extension.
 
Last edited:

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
I'm curious about whose FR/distortion measurements convinced you specifically.
 

HRTF_Enthusiast

Active Member
Joined
Oct 22, 2021
Messages
178
Likes
80
Do you have data to back each of those points up? Here are Crinacle's measurements:

View attachment 169847
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).
graph - 2021-12-03T100816.970.png

ER2SE is <0.1% THD at 80 dB while ER4SR is close to 1% and dominated by 3rd order harmonics.
Screen Shot 2021-12-03 at 10.09.17 AM.png


It becomes closer at 94 dB, but ER2SE is dominated by 2nd order harmonics.
Screen Shot 2021-12-03 at 10.12.47 AM.png
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
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
Interesting. Is this your own data?

Edit: Also, at higher output the nonlinearity seems to stay the same or even decrease for the ER4SR. Why is that?
 
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
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?
 

HRTF_Enthusiast

Active Member
Joined
Oct 22, 2021
Messages
178
Likes
80
Interesting. Is this your own data?

Edit: Also, at higher output the nonlinearity seems to stay the same or even decrease for the ER4SR. Why is that?
I forgot to link the THD site:

As shown in the image, banbeu.com is where i got the SPL graphs. Neither websites have my own data.

For ER4SR nonlinearity, this is due to the behavior of the BA driver. When signal level increases, distortion doesn't increase much in comparison, so the percentage doesn't increase. Distortion typically increases exponentially, so you usually see it increase with SPL. Here's an example with LCD-24 where THD is lower at 114 dB than at 94 dB:
1638565616703.png
 
Last edited:
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
I forgot to link the THD site:

As shown in the image, banbeu.com is where i got the SPL graphs. Neither websites have my own data.

For ER4SR nonlinearity, this is due to the behavior of the BA driver. When signal level increases, distortion doesn't increase much in comparison, so the percentage doesn't increase. Distortion typically increases exponentially, so you usually see it increase with SPL. Here's and example with LCD-24 where THD is lower at 114 dB than at 94 dB:
View attachment 169974
The phenomenom of distortion decreasing at higher SPL doesnt exist. Its error caused by noise floor. If "THD" decreases at higher spl, its not THD, its noisefloor, the real distortion is so low its hidden beneath the noise floor. It is physically impossible for driver with moving mass to drop distortion at higher spl, THD always increases as the driver gets closer to its maximum excursion limit.
 
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
I never said distortion decreases at higher SPL. The distortion is still there. It's just that the ratio of signal to distortion is greater at higher SPL.

View attachment 170028
By distortion, I mean THD. The ratio never increases, its error of measurement. Its becose of background noise, noise is in reality just superposition of many sine waves that cover wide spectrum. The measuring microphone has no way to tell if the 300Hz sinewave it picked up is distortion tone of the driver being measured or part of broadband impulse emited as result of gaseus eruption induced by previous day bean stew.
 

HRTF_Enthusiast

Active Member
Joined
Oct 22, 2021
Messages
178
Likes
80
The phenomenom of distortion decreasing at higher SPL doesnt exist. Its error caused by noise floor. If "THD" decreases at higher spl, its not THD, its noisefloor, the real distortion is so low its hidden beneath the noise floor. It is physically impossible for driver with moving mass to drop distortion at higher spl, THD always increases as the driver gets closer to its maximum excursion limit.
Ah ok. I assumed V1 could increase at a faster rate than V2, V3, ...
Screen_Shot_2021-12-03_at_11.52.09_PM.png

Thank you for the correction!
 
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
Ah ok. I assumed V1 could increase at a faster rate than V2, V3, ...
Screen_Shot_2021-12-03_at_11.52.09_PM.png

Thank you for the correction!
No problem mate! People fall for this trap very often. For example the people at RTINGS still didnt figured this out yet, you would thing they are experienced professionals but no, they fell for it too. So its not just the "noobs", I wish this was common knowledge in the community so we have better graphs to compare and not be left to guess.

Measuring true THD of low distortion headphones is hard, or more accurately, expensive. It requires low noise and distortion amplifiers, A/D converters, microphones and robust sound isolation. The cheap electret mics home made rigs in our living rooms just doesnt cut it more often than not. This needs thick double walls covered in acoustic absorbers, preferably underground and far away from traffic, vibration dampening like sorbothane, B&K or similiar pro grade active measurement microphones and alot of averaging to push the noise down.
 

2M2B

Active Member
Joined
Mar 9, 2021
Messages
118
Likes
73
I would never once compare the ER2 to the ER3/ER4 performance wise. The ER3XR I'm using right now blows away the ER2XR I had & I'm waiting for cash to get the ER4SR. Those people mainly on Reddit/Head Fi(mostly) rarely have any idea what they're talking about beyond subjectivist woo when It comes to BA headphones.


The actual bandwidth of the ER4SR is 20Hz ~ 18.5KHz with THD under 1%. They hype planar headphones for their bass decay but when it comes to the ER4 suddenly It's a problem. Heck I've seen them claim the sundara makes the HD650 sound like slow junk?. But then claim the ER2 has better timbre despite me finding the ER4/ER3 crystal clear and the ER2 sounds hazy/dirty, That DD win on decay despite trashing them when It comes to planar tech. The ER4SR can be easily EQ'd to have the same bass as the ER2XR without any issues since it a 7db boost.

Reddit been in cope mode ever since 2017 when the ER4XR was launched showing a bassy single BA IEM can be done. Same with Head Fi when Custom art did a Single BA CIEM but only say it was after the launch which got reviewers to rewrite their impressions on how the bass sucks after giving it a 5 stars on bass alone?.
 
Last edited:
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
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.
er2se Ts 100 vs Cp800.png
er2xr Ts 100 vs Cp800.png
er4xr Ts 100 vs Cp800.png


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.

er2 er4 impulse.png

Impulse response with stock tips, again, very similiar.

er2 er4 THD 80db.png

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.

er2se THD 80 94.png

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

er2se THD spectrum 94.png

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.

ER2SE vs ER4XR SilverEars.jpg

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.

yuriv ER4PT THD spectrum 94.png

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.
yuriv Quadbeats 3 94 THD.png

yuriv MH755 94 THD.png

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.

yuriv ER2SE THD spectrum 94.png

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.

Hakuzen even order ER2SE R.png
Hakuzen even order ER2SE.png

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.

Crinacle ER2SE vs ER4SR.png
Crinacle 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.

Crinacle ER2XE vs ER4XR.png

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

purr1n ER4SR THD spectrum.jpg

purr1n ER4SR THD spectrum R.jpg

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.

Clarity Fidelity ER4SR THD spectrum L.png
Clarity Fidelity ER4SR THD spectrum R.png

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.

Clarity Fidelity ER4XR THD spectrum L.png

Clarity Fidelity ER4XR THD spectrum R.png

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.

Clarity Fidelity MC5 THD spectrum L.png

Clarity Fidelity MC5 THD spectrum R.png

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.



Clarity Fidelity ER4SR Waterfall L.png

Clarity Fidelity ER4XR Waterfall L.png

Clarity Fidelity MC5 Waterfall L.png


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.
Innerfidelity ER4SR.png

Innerfidelity ER4XR.png

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.

Innerfidelity MC3.png


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.
 

Attachments

  • er2 er4 THD 80db.png
    er2 er4 THD 80db.png
    53.6 KB · Views: 128
Last edited:
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
Harmonic distortion below 1% are completely inaudible when listening to music...
The problem is, music isnt composed just from harmonic sine waves, and when two inharmonic tones get played back through nonlinear system, we get inharmonic intermodulation tones. Furthermore, distortion audibility depends on where in spectrum it happens and what order the distortion is.

The midrange odd order distortion of BA Etymotics is far more audible than for example even order bass distortion of HD800 dynamic driver headphone even through the percentage is similiar. Besides that, ER4 do have distortion higher than 1% at listening levels realistic for dynamic music.

I dont think its deal breaker and that BA Etys suck, but lets not pretend they sound like Stax or high end planar like HE6 or Audeze LCD 24 which have ultra low distortion. In my opinion a better target for practically inaudible distortion would be 0.1% which means 60 dB gap. And if one wants peace of mind to never ever hear any distortion even with test sounds specifically tailored to maximize distortion audibility, then 0.01%, or -80 dB is good target.
 
Last edited:

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
I think the extension difference in the older model is not significant. Distortion is better.

I don't know what their reasoning is. The BA they use is custom. Perhaps it will make sense once Etymotic describes the production process. Consistency and cost are important too.

I also think this is proof enough that there's nothing wrong with BA drivers per se, seeing these results.
 
OP
Graph Feppar

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
269
Likes
258
I think the extension difference in the older model is not significant. Distortion is better.

I don't know what their reasoning is. The BA they use is custom. Perhaps it will make sense once Etymotic describes the production process. Consistency and cost are important too.

I also think this is proof enough that there's nothing wrong with BA drivers per se, seeing these results.
What does "wrong" even mean? Its vague term that can mean so many different things its almost meaningless.

I think the BA Etymotics are excellent headphones when judged against other headphones from other companies. If we compare them to ER2 they are simply inferior. Inferior doesnt necessarily mean bad or wrong, it means there is something better out there. Mercedes E class is inferior to the S class, but its still one of the best cars money can buy.
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
What does "wrong" even mean? Its vague term that can mean so many different things its almost meaningless.

I think the BA Etymotics are excellent headphones when judged against other headphones from other companies. If we compare them to ER2 they are simply inferior. Inferior doesnt necessarily mean bad or wrong, it means there is something better out there. Mercedes E class is inferior to the S class, but its still one of the best cars money can buy.
I was making a general comment. There's lots of stuff online that unfairly disparages BA drivers.
 
Top Bottom