1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Etymotic ER2SE - A Reference Headphone for your Ears and Your Couplers

1 2 3 4 5 6
  1. twiceboss
    anybody here know cheap 2.5mm mmcx cable that is angled for ety's iems? anything around $20
  2. castleofargh Contributor
    @ClieOS is all knowing when it comes to ety cables and parts.
    twiceboss likes this.
  3. twiceboss
    Help me @ClieOS ! my er2se sounds more open with 2.5 balance on es100+xcan pairing. That's why im considering a 90 degree angled of mmcx, if not i will look like an alien wearing etys with straight cable lol
  4. ClieOS Contributor
    $20 might be too little a budget consider that the a pair of Ety MMCX connectors already cost $10 on their own. You will need to double your budget to $40 to find one. IIRC, Lunashops might have some. I can also build you one, if you are willing to pay for the material and shipping.
    twiceboss likes this.
  5. twiceboss
    yeh, cuz im finding something like yinyoo $20 cable :)
  6. csglinux
    With Ety's angled mmcx connectors???

    I suspect there isn't really the demand for Etymotic to mass-produce balanced cables of their own. (It might seem like there is on the main ER4 thread, but half a dozen people asking isn't really a demand.) Mass production and cheap wire is the only way you'd get down to prices like $20. Even if @ClieOS were kind enough to donate his time and labor for free, I'd bet he'd effectively end up paying you to take a set of custom cables from him at that price.

    BTW, on another topic... The tour headphones are now on their way to @yuriv. Some disappointing news here. We're not, at this stage, going to be able to get any measurements from @jude's rig :frowning2: They've currently got their measurement rig setup in some super-secret configuration for one specific customer, so the timing for our little ER2SE's measurement wasn't good. So, for now, we'll have to figure all this stuff out for ourselves :wink: Current plan is (I believe) for the European leg of the tour after @yuriv...
  7. twiceboss
    Yes, but he is in Johor. But yea, ill be going back to Malaysia for good soon too. Idk yet, im still in IEMs rolling. No endgame yet lol...
  8. ClieOS Contributor
    There has been talk of an official balanced cable from Ety for over a year now. Even if it is coming, it won't going to be cheaper than $50 since that's already the price for their single-ended cable.

    On the other hand, my cable is fairly cheap because all the no-thrill material I used. You can always make it more expensive by coating it with a layer of snake oil. I heard the effect is enormous. :grin:
  9. hakuzen
    twiceboss likes this.
  10. yuriv
    Here’s what I have so far. I measured with my portable rig with the Apple USB-C dongle instead of a “proper” interface like the Focusrite Forte I have parked at my desk. The results were close enough when I set the mic’s level in Windows to 60 or higher. It’s a bit strange; with the level any lower, the measurements show a roll-off in the bass.

    Most of the sine sweep measurements were made with 1M points and 8 averages. For stepped sine, it's 12 points per octave, 8 averages.

    Here's a sampling of images that can be made from the measurement data. I can upload only 25 pictures, so I'm showing the left channel only for most of the measurements. The graphs for the right channel look similar.

    Etymotic ER2SE frequency response.png
    Etymotic ER2SE, frequency response with ER38-15 small frost tai-flange tips

    Etymotic ER2SE impulse response, ER38-15SM.png
    Etymotic ER2SE, impulse response from sine sweep measurement

    Etymotic ER2SE step response, ER38-15SM.png
    Etymotic ER2SE, step response from sine sweep measuremnt​

    Etymotic ER2SE at various insertion depths, left channel.png
    Etymotic ER2SE, frequency response with ER38-15 small frost tri-flange tips at various insertion depths.​

    Etymotic ER2SE harmonic distortion, left channel.png
    Etymotic ER2SE, harmonic distortion from sine sweep measurement: 94 dB SPL at 500 Hz.​

    Etymotic ER2SE THD stepped sine L.png
    Etymotic ER2SE, harmonic distortion from stepped sine measurement: 94 dB SPL at 500 Hz.​

    The results are close, but the cheap electret mic in my coupler adds its own distortion to the measurement, mostly the second harmonic. The information isn’t completely useless. For example, it gives us an upper bound for the IEM’s distortion.

    I also made measurements with other tips:

    Etymotic ER2SE with ER38-18 L.png
    Etymotic ER2SE, frequency response with ER38-18 large gray tri-flange tips at various insertion depths.​

    Etymotic ER2SE with Comply TX100 L.png
    Etymotic ER2SE, frequency response with Comply TX100 tips at various insertion depths.​

    My impressions of the ER2SE’s sound: It sounds like the other Ety IEM’s that I’ve heard. I’m hearing slightly too much emphasis between 1k Hz and 2.5 kHz. The response there goes up too fast, compared to what I hear from calibrated speakers. Etymotic’s target has the peak at 2.7 kHz, and that’s exactly where I hear it in sweeps. With calibrated speakers, I hear it around 3.3. kHz. All of this makes me hear a slight coloration, which puts vocals a little too forward. Using an equalizer to shelve up the bass makes it less noticeable. It’s not as bad as how I hear it on an ER4S. I prefer the way the Sony and Samsung tour samples handle 1k - 3k. The treble on the ER2SE sounds a little shelved down, just like on an ER4PT.

    On a whim, I tried the so called P-to-XR adapter that I built for the ER4PT (25 ohms DC resistance), and it somewhat works for the ER2SE (32 ohms):

    Etymotic ER2SE with ER38-18 tips, P2XR.png
    ER2SE frequency response with yuriv’s so-called P-to-XR adapter:​

    I give up 13 dB of sensitivity, but it’s no problem for the Apple dongle. I prefer the resulting sound. The graph reminds me of the Harman on-ear or over ear targets.

    Sony MH755 tour sample L.png
    Sony MH755 tour sample, frequency response at various insertion depths.​

    I also had a problem fitting the medium silicone tip in the coupler, so I used a small tip. The measured distortion is a bit lower than on the Ety. My coupler mic adds its own distortion to the measurement. But it does establish an upper bound for the MH755's THD. Its second harmonic distortion could be much lower than shown below:

    Sony MH755 THD from sweep L.png
    Sony MH755 tour sample, harmonic distortion from sine sweep measurement: 94 dB SPL at 500 Hz.​

    Sony MH755 THD from stepped sine L.png
    Sony MH755 tour sample, harmonic distortion from stepped sine measurement: 94 dB SPL at 500 Hz.​

    The tour sample sounds more like my MH750 than my white MH755 samples. Here’s one channel from one of six samples from two batches:

    Sony MH755, different sample.png
    Another Sony MH755​

    Here it is compared to the tour sample:

    Sony MH755 tour sample vs. typical sample.png
    Sony MH755 tour sample (black) compared to (typical?) white sample. Right channels shown.​

    I prefer the sound from my samples. The response at 200 Hz is lower. Some people like the extra warmth with an elevated response there. I think it sound a bit muddy, so I’m more in agreement with the Harman target’s low point there. The peak from the ear canal resonance is more pronounced in the tour sample. On the white samples, the response is smoother; the peak is a smaller bump because it’s being covered up by another resonance, it seems. When I put dampers in front of the nozzles, the response around 6 kHz drops and the peak from the canal resonance becomes more pronounced, like on the tour sample.

    For me, the response below 80 Hz is already too elevated on my MH755, let alone the tour sample. Here’s the effect of sealing the rear cavity on the tour sample:

    Sony MH755 tour sample, rear volume sealed.png
    Sony MH755 tour sample, effect of sealing the rear cup​

    There’s some glue that fastens the rear cup to the front. On some samples, just blocking the top vent will result in a few dB less bass. But on the tour sample, that didn't make much difference in the measured response. There must be a leak somewhere in the seam. Even sealing all around and leaving just the bottom of the strain relief open had little effect. I had to plug all the leaks to get the response shown above. It sounds better IMO, but the effect isn’t as nice as on the other samples.

    BTW, the tour package also comes with two junk IEMs. One of them is a broken Philips SHE7000. I might have damaged it even more. It’s definitely unusable. The other is a fake Sony MH750. I can confirm that it is, indeed, a fake. At first, I thought that it could have been a malfunctioning unit, judging from the frequency response @SilverEars provided earlier. But I opened it up and discovered that the driver unit isn’t even glued on. The thing rattles when you shake it, and there’s nothing separating the front of the IEM from the rear cavity:

    Fake Sony MH750.jpeg
    Fake Sony MH750​

    I’ll keep the fake in the tour package because it might be of interest. I’ll replace the damaged Philips with something nicer.

    Samsung EO-IG955 tour sample L.png
    Samsung EO-IG955, frequency response at various insertion depths​

    The treble response is very sensitive to differences in insertion depth. Compare it with the Sony MH755 measurements, especially my sample, where the bump from the ear canal resonance just slides up and down the main response. If I were to make a guess, I'd say that the typical EO-IG955 response isn't the one with the peak at 9 kHz, like the measurements at zerodecibel or The Ear-fi Blog, but rather this one:

    Samsung EO-IG955 tour sample.png
    Samsung EO-IG955, frequency response with shallow insertion.​

    The EO-IG955 is a shallow-insertion IEM with short tips. I'm actually hearing the peak lower than shown above—at 6.3 kHz, which is really obvious out when I verify it with a sine sweep (like this one: https://www.szynalski.com/tone-generator/ ) or a program like Sinegen. I'm hearing a huge slide in the response immediately above that, and 7 kHz sound a lot softer. The ideal response places the peak closer to 8 kHz, but that's almost 6mm beyond what's possible for my ears. I wrote about this experience in my review of the EO-IG955: link. Perhaps a really long silicone sleeve propped up on the stem will fix the response for me. I haven't tried it yet.

    Here’s what the frequency response looks like with a foam tip:

    Samsung EO-IG955 tour sample with Comply TX400 tip L.png
    Samsung EO-IG955, frequency response with Comply TX400 tip at various insertion depths.​

    For the measurements with silicone tips, there's a little wiggle in the response between 100 Hz and 200 Hz. It's the resonance of a mechanical system with a spring, mass, and damper, just like the "pad bounce" in Tyll's headphone measurements at IF. I didn't use Blu-Tack to mass load the earpiece in the coupler for most of the measurements. When I did, the frequency of the wiggle got lower, which is to be expected.

    We included this IEM in the tour so that we can make stepped-sine distortion tests at 94 and even 100 dB SPL without caring too much if it gets damaged (it shouldn’t):

    Samsung EO-IG955 tour sample THD at 94dB from stepped sine L.png
    Samsung EO-IG955, harmonic distortion from sine sweep measurement: 94 dB SPL at 500 Hz, left channel​

    Samsung EO-IG955 tour sample THD at 100dB from stepped sine L.png
    Samsung EO-IG955, harmonic distortion from stepped sine measurement: 100 dB SPL at 500 Hz, left channel​

    The corresponding measurements from the sine sweeps are similar. Like before, the measurement also shows the distortion of the coupler's mic. The IEM's THD should be below 1% across most of the spectrum even at 100 dB. The MH755 got the lowest distortion of the three, judging from the measurements at 94 dB.

    The tour sample has a channel imbalance,:

    Samsung-AKG EO-IG955 frequency response showing channel imbalance.png
    Samsung EO-IG955 frequency response showing channel imbalance.​

    After my chat with @csglinux, I still wasn't sure if he set the level for each channel so that each one was at 94 dB at 500 Hz. If that's the case, then it's why the measurement in a previous post shows better channel balance, with more treble for the right channel. @SilverEars: Can you confirm that the tour sample had this imbalance before it shipped?

    I should have finished all of the measurements earlier, but I got sidetracked by preparations for Hurricane Dorian last weekend. That and a new audio toy I got yesterday. Lol. I expect to finish this week, then I’ll send the package on its way to Europe. If anyone in the U.S. wants in, now is the time to speak up.

    I’m thinking of making another set of measurements with the ER2SE. I just saw that there’s a pair of ER38-14F foam tips in the package. I would have use them instead of the Comply TX100 had I noticed earlier. Maybe I’ll make a set of impedance vs. frequency measurements, but only the Samsung EO-IG955 will have something much different from a flat line with a tiny blip in the middle and perhaps a little rise up top. The other two basically look like resistors to a headphone amp.

    If any of you want me to measure anything else for comparison to future measurements you plan to make, please let me know soon.

    @McMadface Please grant me access to the shared drive so I can upload my .mdat files. It’ll be good to compare notes. My frequency response measurements look similar to the ones made with the RA0045. The ones for harmonic distortion, obviously not. I’d love to see what everyone else got.
    Last edited: Sep 8, 2019
    hakuzen, castleofargh and Slater like this.
  11. csglinux
    Wow, epic post @yuriv! That's a lot of work you've done there :) This will be an inspiration to those that follow you on the tour, and a mark of shame to those of us lazy folks that preceded you and only did one measurement at one insertion depth and then went off to play pubg :wink:

    Your coupler FR looks really good. One day we've got to try and track down that manufacturer/seller! I really like the idea of showing FR results as a spread, i.e., as a function of insertion depth. This nicely identifies the driver and canal resonances and gives users an idea of what they'd hear, depending on insertion. Obviously that's a lot of extra work, but very informative.

    I can confirm I saw a channel imbalance with the tour AKG headphones. I presume that imbalance was there from the get go, but I think @SilverEars only measured one channel, so we've no record of any imbalance before I received them. @SilverEars - did you notice this? I'm 100% certain I would simply have aligned the L and R data by adding an offset after measuring. I typically do that anyway (either at 1 kHz, or these days at 500 Hz) with all my measurements. So unfortunately I don't have a record of the differing sensitivities of the left and right buds, but your measurements and observations would certainly line up with what I saw.

    I'm intrigued about your use of that Apple dongle. Were you driving the whole setup from REW (if so, how?!?), or were you playing back REW sweeps from file on the iDevice? I think I may have already asked you this, but 1) I forgot the answer and 2) the answer would be worth sharing with everybody anyway :wink:
    hakuzen likes this.
  12. SilverEars
    I don't have records of the measurements of the two channel differences. I recall measuring two channels and then realized I shouldn't set the SPL to 94dB for both if I'm checking for channel imbalance. What is important though is making sure the both sides are precisely at the same insertion position when checking for channel imbalance as I notice significant change in loudness the SPL meter reads just from a small nudge of the iem on the coupler.

    I have couple of other sets of AKG. I'll see if those have channel imbalance. By the way, I recently got couple other Philips iems and both of them had pretty significant channel imblance (this is not mere channel imblance, responses on both sides were different) and at this point, I question their QC. That's 4 Philips iems I've measured channel imblance. I did included one of them in the tour kit.

    I also noticed that the black MH755 provided on the tour unit has a different sig from the white one I have. The black one does roll-off quicker in the highs, and subs are at a much greater level.
    Last edited: Sep 8, 2019
  13. yuriv
    It doesn’t seem much of a chore if it’s the background task. Push the IEM a little farther in, then take another measurement while you’re being productive with something else. Little and often gets the job done too. Believe me, I’m just as lazy as anyone here.

    Showing the effect of the insertion depth on the resonance frequency is useful information. What you see on that graph isn’t too far from the truth, probably for most people. It’s basic physics. I actually verified that I’m hearing the peak at the same frequencies. It’s pretty obvious when a test tone is so much louder when it hits resonance. What’s the rule of thumb, 3 dB for a noticeable increase, and 10 dB for a perceived doubling of loudness? That doesn’t happen when I’m listening to calibrated speakers. At least, not to the same degree. It’s eye opening how many dB needs to be cut with the parametric EQ to flatten some of them by ear until the sweep makes it sound only very slightly louder than the surrounding frequencies. Almost without fail, I hear an improvement whenever I do this. Of course, YMMV, yada, yada, yada.

    The RA401/RA402 high resolution couplers don’t capture this effect, which is very real and happens in real human ear canals. Because physics. If I’m reading the G.R.A.S. whitepapers and brochures correctly, it’s not really the point of the hi-res coupler. They make no claim that its accuracyis better than the RA0045’s below 10 kHz. What they claim is better precision. The results from measurements made with the RA402 are more repeatable. This is useful for the folks designing hearing aids and IEMs. If they see a peak on a measurement made with the RA402, they know it’s a tube or driver resonance and not the effect of the ear canal. Additionally, their distortion plots won’t overestimate the THD; the graphs won’t show spikes at fractions of the ear canal’s resonance frequency, like you see on our data. Ideally, they’d have both kinds of couplers. The really serious outfits probably do by now.

    It’s good to know about the channel imbalance. True, the SPL goes up when you push the IEM a little farther in, but the difference I got between channels is still large enough after taking that into account. Look: each one of the following traces has the same signal voltage level at the IEM’s terminals:

    Left Channel (shown in my previous post)​

    Samsung EO-IG955 tour sample R.png
    Right channel (not shown until now) showing different SPL from left channel even after accounting for differences in insertion depth.​

    Yes, I got the same thing, as you can see from my previous post:

    Sony MH755 tour sample (black) compared to (typical?) white sample. Right channels shown.​

    Yes, I was using the Apple USB-C dongle with REW. It’s similar to the following setup, but instead of using my cheapo in-ear mic for measuring headphones, it’s connected to the coupler. The IEM goes into the other 3.5mm jack coming out of the splitter:

    yuriv’s cheap-ass headphone measuring rig: link

    Both Windows and MacOS recognize the dongle with no additional driver needed. But it doesn’t show up in the list of devices unless something is plugged into the 3.5mm jack. A mic or headphone plugged into the splitter like the one in the photo above will do this. I’ve observed that the results have some roll-off in the bass unless you set the Windows mic level to 60 or higher.

    In our chat we proved that we can build an even more portable measurement rig that can be used with REW, albeit indirectly. We exported the sine sweep from REW as a wav file and put it into the iPhone or iPad. This signal is called the stimulus. The setup looks like this:

    iPhone --> headphone/mic splitter --> IEM --> coupler -->headphone/mic splitter --> iPhone​

    We played the file straight from the Dropbox app, but we could have used iTunes. We used the AudioTools app’s recorder to record the output of the coupler’s mic. It’s convenient that the Dropbox app can play audio in the background. The resulting file is called the response to the stimulus and can be imported back into REW. AudioTools has a built-in web server to make the file transfer more convenient.

    What we did was a rough proof of concept, but it does work. If we ever write a how-to guide, we’ll have to come up with way to set the level and calibrate the output precisely. I’m sure there’s a way to come up with a calibration signal that will make it quick and convenient.
    Last edited: Sep 9, 2019
    csglinux likes this.
  14. csglinux
    Nice posts @yuriv! Maybe even too nice, because now you'll have intimidated and frightened away potential future contributors to this tour :wink: We've not yet had anybody brave enough to volunteer to measure these headphones with a vinyl tube coupler. It still think it would be interesting to see what the responses look like from a similar-length, flat transfer-impedance cylinder.

    Aha! Apple USB-C dongle - got it! Yep, that makes sense. I even have one of those lying around in a drawer somewhere. I should test it with REW...

    It looks like we're all in agreement on the hi-res (RA0401/2) coupler. What a shame. After seven pages of this thread, we've had zero arguments. Come on guys, this is headfi! Ok, let me put my GRAS hat on for a minute and play Devil's advocate in defense of the hi-res coupler...

    One of the main problems with existing 711 couplers is that the transfer impedance is only nominally defined up to 10 kHz. Here's the typical "standard" transfer impedance and existing "standard" spec, which stops at 10 kHz:
    Screenshot 2019-09-15 16.35.25.png

    Here's the typical hi-res coupler response and GRAS' own spec for the transfer impedance beyond 10 kHz:
    Screenshot 2019-09-15 16.35.39.png

    What the transfer impedance does beyond 10 kHz in a standard coupler is up in the air, beyond design specification, and could potentially vary significantly from coupler to coupler. That's a problem if we're wanting to compare measurements from two different 711 couplers. Also, bear in mind that the 10 kHz mark is actually a bit arbitrary for us, because this refers to IEMs or hearing aids that finish at the coupler reference plane. These couplers were never designed for shallow-insertion IEMs, which can cause the half-wave resonance to fall well below 10 kHz.

    The ingenious part of GRAS' hi-res coupler is that it separates out resonances caused by the headphone driver (something an IEM designer can change) from the ear canal resonances (something the IEM designer can't so easily anticipate or design around). Also, in practice, the eardrum isn't a hard, flat/normal termination, but rather angled or slanted to the axis of the ear canal, and should therefore be expected to diffuse canal-related resonance peaks somewhat. By completely removing the uncertainty of those canal resonances, the designer can get more consistent and repeatable FR measurements, and lower THD.

    Ok, let me take that GRAS hat off now. Phew. That's better. That was slightly uncomfortable :wink: Ok, here's what I really think. I actually mostly agree with @yuriv. Yes, it's a clever invention, and maybe(?) it has some use for a designer, but ultimately it doesn't seem to represent what you (or at least, me) would actually hear, because those ear canal resonances are real phenomena and existing 711 coupler transfer impedance(s?) seems to get sufficiently close and in practice don't appear to differ all that much between the various crop of 711 (or 711 clone) couplers. And whether the IEM is designed for deep or shallow insertion is, and should be, very relevant to the frequency response. A potential customer should want to see all resonance peaks, because they'd all contribute to the net experience of using that IEM. I'm not sure high repeatability is helpful if it's not that accurate. I joked to GRAS that they should make a DA0402 (Douglas Adams coupler) that would alway return 42 dB, regardless of frequency or IEM being used. That would be highly repeatable :)

    From the perspective of a designer, maybe(?) it makes sense to use a hi-res coupler, but from the perspective of somebody that wants to report what listeners might actually hear from a given headphone (e.g., reviewers, rtings.com, @crinacle, all of us, etc.), I don't think we'd want to be using an RA0401 or RA0402 coupler. This extends to distortion measurements too - because if those resonance peaks actually exist, so do any distortion components that are amplified by those resonance peaks. Hi-res coupler fans - if I've overlooked something important, feel free to push back...
    Last edited: Sep 16, 2019
    yuriv and hakuzen like this.
  15. JohnYang1997
    Not the hires coupler.... It's fundamentally wrong. This statement sounds too crude but that's the truth.
    The tamed impedance over 10khz is actually down under to 6khz when shallow insertion. Comparison on 0db shows that hires coupler just isn't usable anymore. The older 711 at least shows a peak at different frequency when the iem has a peak. Hires one doesn't show it at all.
1 2 3 4 5 6

Share This Page