CHANGSTAR: Audiophile Headphone Reviews and Early 90s Style BBS
Lobby => IEM Measurements => Topic started by: ultrabike on July 07, 2014, 08:42:06 AM
-
Sound
Definitively not bassy, and IMO slanted down from the lower mids down. Regarding the upper mids and lower treble I'm definitively missing detail and air... unless I pump up the volume in which case things get indeed substantially better across the board perception wise. This IMO is an analytic signature with a lot of clarity and detail. Reminds me a bit of the Audeos which I had to equalize by almost 10 dB (or more)...
http://www.innerfidelity.com/comment/482739#comment-482739
^--- Equalizer setting included before I ever made a single singing can or cutetip measurement over 2 years ago.
Comfort
Horrible. Triple flange-ing my brain or else SQ is uber-schitty.
Presentation
Doesn't matter, whole thing went inside my ear-hole.
Price
I think like $300...
Overall
OK if you like your music loud and EQ up the bass.
Measurements (uncompensated)
Frequency Response
(http://www.changstar.com/index.php?action=dlattach;topic=1665.0;attach=6934;image)
Distortion Right
(http://www.changstar.com/index.php?action=dlattach;topic=1665.0;attach=6936;image)
Distortion Left
(http://www.changstar.com/index.php?action=dlattach;topic=1665.0;attach=6938;image)
Impedance (w adapter)
(http://www.changstar.com/index.php?action=dlattach;topic=1665.0;attach=6940;image)
Comparo
(http://www.changstar.com/index.php?action=dlattach;topic=1665.0;attach=6942;image)
Note regarding compensation or not:
A lot of measurements I have seen w/o compensation do show substantial roll off in the treble area. Consider the following measurements from IF before compensation (gray)
http://www.innerfidelity.com/images/EtymoticER4PT.pdf
http://www.innerfidelity.com/images/AudeoPFE121GrayFilter.pdf
http://www.innerfidelity.com/images/DunuDN1000.pdf
I could go ahead and compensate, but I'm not certain that applying standard compensation will yield the absolute response I hear. Specially with IEMs considering differences in fit and ear shapes. IMO one may gain more by looking at these plots in relative terms, and discuss honest impressions using the measurements as just another data point. Specially with IEMs.
I do tend to measure less bass than some, but then again I felt the Dunus where pretty nice in the bass area and measured like so... again, I would look at these data points in relative terms. Some may find the Dunus bassy depending on their own baseline.
-
I think that given that the ER4 is one of the most commonly measured earphones around, you can get started on that compensation curve! :)p1
You can do it a few ways --- first compensate your raw data to fit raw ear simulator response (IEC60711/60318-4), and then compensate for the common DF, ID, etc. equalizations, and then you can do it according to your own hearing, and see how they differ.
-
Will see what I can do... :money:
-
I could go ahead and compensate, but I'm not certain that applying standard compensation will yield the absolute response I hear. Specially with IEMs considering differences in fit and ear shapes. IMO one may gain more by looking at these plots in relative terms, and discuss honest impressions using the measurements as just another data point. Specially with IEMs.
Fair enough. That can work well going forward.
But I'd like to add that most IEMs that don't even get close to following the Ety curve above 1k simply don't sound accurate to me. Too sucked out; it's like my brain is reaching out for extra midrange information but getting nothing. It really is unlistenable for me. These plots also provide some evidence to my Tenore impressions, in that once I was able to tame the bass on the Tenore, I felt it was a very good IEM (tonally). Again, it's because it's getting closer to that Ety curve above 1k.
We can argue about bass compensation all day I guess. But we should all be able to agree that the ER4S is not that rolled off (in the bass) in terms of raw response. It should be flat out to 100Hz or so...with a slight slope going toward 20Hz. Maybe 3dB if I'm going to use my ears as judge.
But so long as we can use the ER4S as a relative landmark for other IEMs, that provides a very good baseline for ultra's measurements.
Ultra: Since you feel the ER4S is too lean, I'd be interested in what sort of EQ settings you'd use to compensate.
-
As far as EQ settings I would probably end up with something close to what I used for the Phonaks (see IF linky on first post), but will try again tonite.
I really don't hear these as flat down to 100 Hz. They are smooth though. I've seen the Ety (and Phonak and so on) plots a fair amount of times and that's not what I feel I get with speakers that I measured relatively flat down to 100 Hz. I know there has been a fair amount of research on this IEM comp stuff, but I'm not getting this perceived flatness I keep reading about. Furthermore, I don't hear the Etys uber neutral above 1 kHz. I hear it analytical.
I think Tomscy2k brought up that when the bass is brought up, some sacrifices are made in resolution due to masking... But while I do feel that bringing the bass down may bring clarity and information forward, that IMO does not mean making such a move gets things to sound closer to neutral. Furthermore, adding further emphasis on certain frequency ranges (say 300 to 3400 Hz for vocals and other ranges for other instruments) may bring detail in some areas at the expense of others.
I don't think these sound like HD6x0s... Perhaps closer to a stock HD800, but in dire need of bass equalization... That and/or my ears are weird.
-
But I'd like to add that most IEMs that don't even get close to following the Ety curve above 1k simply don't sound accurate to me. Too sucked out; it's like my brain is reaching out for extra midrange information but getting nothing. It really is unlistenable for me. These plots also provide some evidence to my Tenore impressions, in that once I was able to tame the bass on the Tenore, I felt it was a very good IEM (tonally). Again, it's because it's getting closer to that Ety curve above 1k.
We can argue about bass compensation all day I guess. But we should all be able to agree that the ER4S is not that rolled off (in the bass) in terms of raw response. It should be flat out to 100Hz or so...with a slight slope going toward 20Hz. Maybe 3dB if I'm going to use my ears as judge.
I agree about >1khz tuning, BA IEMs have typically been rolled off up until the past few years. Way back when the ER4 first hit Ety went to great lengths to get as much usable bandwidth out of the BA as they could.
And yeah, maybe 6db-9db down at most in the bass going by ear.
-
As far as EQ settings I would probably end up with something close to what I used for the Phonaks (see IF linky on first post), but will try again tonite.
I really don't hear these as flat down to 100 Hz. They are smooth though. I've seen the Ety (and Phonak and so on) plots a fair amount of times and that's not what I feel I get with speakers that I measured relatively flat down to 100 Hz. I know there has been a fair amount of research on this IEM comp stuff, but I'm not getting this perceived flatness I keep reading about. Furthermore, I don't hear the Etys uber neutral above 1 kHz. I hear it analytical.
I think Tomscy2k brought up that when the bass is brought up, some sacrifices are made in resolution due to masking... But while I do feel that bringing the bass down may bring clarity and information forward, that IMO does not mean making such a move gets things to sound closer to neutral. Furthermore, adding further emphasis on certain frequency ranges (say 300 to 3400 Hz for vocals and other ranges for other instruments) may bring detail in some areas at the expense of others.
I don't think these sound like HD6x0s... Perhaps closer to a stock HD800, but in dire need of bass equalization... That and/or my ears are weird.
Can't say I agree. These are a hell of a lot closer to HD600 than stock HD800 in terms of the upper mids and treble. Which shouldn't be surprising as the HD600 are tracing a similar DF target curve as the ER4S. If the ER4S was as bright as stock HD800, I simply wouldn't be able to listen to it for hours on end. Enough said. The ER4S is certainly analytical/ruthless in nature, but not bright.
Regardless, you can look at just about every other raw IEM measurement of the ER4S and it'll show flat bass. Whether that's the ideal curve or not for a human listener we can argue, but it should have flat response. The bass roll-off/shelving shown in your plots (it's in all of them) is clearly a function of your coupler or technique.
But I don't think your ears are weird ultra. I think you're just used to uber bassy IEMs.
-
Dunno mang. I can tell you that my first IEMs were the Phonaks... But I'm more than willing to concede that my brain may be playing games with me.
So I think the IEC 60318-4 (previously IEC 60711) comp curve is this one:
http://rinchoi.blogspot.com/2012/08/introducing-my-60318-4-occluded-ear.html
another alternative reference:
http://www.aco-japan.co.jp/eng/sound/type2128/type2128.html
I also think that that's sort of an average, things get pretty dude(tte) dependent on the HFs. Frequency range on the comp curve seems to go from 100 Hz to 10 kHz... Will see what I can make of that.
-
To me, the ER4S has bass linearity until about 45 Hz, when it starts to roll off pretty badly. I think the issue is that the ER4S underdamps the driver with a pair of resistors, and that affects the way the bass responds to any amplifier. Even though I appreciate the ER4's smoothness, its bass doesn't ever feel convincing to me.
At the same time, the FitEar F111, which uses no resistor underdamping and chooses instead to use horn acoustics to inflate the highs and smooth out the high frequencies by matching acoustic impedances, the bass response actually sometimes sounds a little "muddy" to me for some reason. It's the same driver, just without passive resistance equalization and a titanium horn.
On the other hand, custom-molded versions of the ER4S tend to be more peaky in the upper midrange perhaps because of tubing and the acrylic interface.
Then there's the Noble FR/PR in "R" configuration, which is a facsimile of the ER4P except with a miniscule amount more resistance applied.
http://www.innerfidelity.com/images/NoblePRRTuning.pdf
http://www.innerfidelity.com/images/EtymoticER4PT.pdf
-
Very interesting to see the Noble PR's response next to the ER4. Didn't realize he had them up. Looks like Tylls plots do reveal a bit of a bump near 5k, which I did feel was audible on demoing. Other than that, very similar response. It's all starting to make sense to me.
-
Dunno mang. I can tell you that my first IEMs were the Phonaks... But I'm more than willing to concede that my brain may be playing games with me.
So I think the IEC 60318-4 (previously IEC 60711) comp curve is this one:
http://rinchoi.blogspot.com/2012/08/introducing-my-60318-4-occluded-ear.html (http://rinchoi.blogspot.com/2012/08/introducing-my-60318-4-occluded-ear.html)
another alternative reference:
http://www.aco-japan.co.jp/eng/sound/type2128/type2128.html (http://www.aco-japan.co.jp/eng/sound/type2128/type2128.html)
I also think that that's sort of an average, things get pretty dude(tte) dependent on the HFs. Frequency range on the comp curve seems to go from 100 Hz to 10 kHz... Will see what I can make of that.
- You cannot just apply "random" IEC, Dr. Zhivago, ANSI, Olive-Welti, Mr. Fantastic, ad inf compensation curves because the couplers have different responses, e.g. you have to be using the the exact same coupler / setup / rig as Sean Olive to use the Olive-Welt curves, etc.
- DF, FF, HRTF, quantum fields, black holes, singularities, whatever. At the end of the day, it's all subjective bullshit / incomplete research based on sh1t which has not been established or is very difficult to establish.
- Use your ears. I know you have good ears. Making a comp curve based on that will yield better results that whatever BS AES paper. Use the Sada-Chen-Flores curve. It's really no less valid than any other fuckers'. They only difference is that you haven't written a fricking AES paper and you may actually have better ears than scientists. It's OK. This is changstar. Not AES where you need to write 15 pages to justify something.
- IMO, the ER-4S overdid the "DF" thing. Slightly too much 2-3k. It also has a minor treble spike which causes me fatigue in 45 minutes. This is based on my comparisons to speakers tuned similarly to B&K ideal response: http://techtalk.parts-express.com/attachment.php?attachmentid=6329&stc=1&d=1262996968 (http://techtalk.parts-express.com/attachment.php?attachmentid=6329&stc=1&d=1262996968)
- What's important is to be consistent and establish your own framework so your results will be comparable to each other - or at least have batches of measurements be comparable to each other. You'll never find absolutes only relatives.
-
To me, the ER4S has bass linearity until about 45 Hz, when it starts to roll off pretty badly. I think the issue is that the ER4S underdamps the driver with a pair of resistors, and that affects the way the bass responds to any amplifier. Even though I appreciate the ER4's smoothness, its bass doesn't ever feel convincing to me.
Tom: do you think it would be worthwhile to use a crossover network that adds more series resistance to the ER4. Obviously it would have to shunt away the extra resistance above 1k or so (like the ER4B did?)
On the other hand, custom-molded versions of the ER4S tend to be more peaky in the upper midrange perhaps because of tubing and the acrylic interface.
The custom molds I have for the Etys are soft silicone, not acrylic. But you're right about the added peakiness. Tri-flanges are not only smoother and more tonally correct, but also more detailed. Customs do have the ability to give a wider, less "in-your-head" image though which can be a nice effect.
-
Are AES papers really 15 pages long? If so, that's terrible. Whoever's doing the peer reviewing system needs to get fired.
DF is a bit of an oversimplification. It essentially assumes a static ear, with no active cochlear process, will register truthfully whatever signals its environment gives it.
More series resistance to the ER4? Probably not. Really marginal treble gains probably with no real improvement in linearity is my prediction. The 4B uses a really shallow high-pass filter, which is why it has a bit more roll-off than the 4P and 4S when the bass shelf is compared directly.
The peakiness most likely has to do with the tubing more than the substrate, though the material of the shell does have to do with the sound.
-
Alright, so I'm doing quick ear comparos between my lowly amp whore HD558 and these... No way these sound like the Senns to me... They have bass, but IMO these are more analytical and have obviously less bass (well, at least to me). That said, I can easily and clearly perceive tones all the way down to 30 Hz with the Etys, just the tonality I get is towards the analytic side of things.
As far as compensation, I fully understand what Marv is saying. I get pretty consistent results with my coupler all the way up to 9 or so kHz. However, this is not an official, standard, Pope blessed coupler so the gap shape and volume may necessitate a different compensation + whatever adjustments. I'll see what I can get, but in the mean time, relative results across measurements should be comparable.
So if one feels Etys are the neutral reference, other things like the Dunu, SHE cutetips, and most random dynamic and perhaps even 100 driver BA designs may sound uber-bassy.
Closest things I've heard to these Etys would be the Phonaks PFE112 (gray filters), followed by the Noble 4s and indeed the Noble FR/PR in R config.
I really was originally skewed towards thinking "objectively" that the PFE112s where neutral given the measurements and the standards, until I had them on my ears. For a long time I tried all kinds of tips (including some Comply ones that I gave away in the end), cuz I thought the problem was seal... to no avail and pretty much gave up on IEMs. Tried some "OK for the $" dynamic IEMs to get more measurement baselines, and bass absence problem went away using similar size and shaped tips, even "slightly" bassy ones according to some reviewers, which helped me rule out the seal problem theory.
You guyses know I felt the Noble 4s were a bit bass lite to me. Maybe with IEMs things get a bit more relative, Dunno...
-
Not that I seem to be able to get IEMs or similar to work with my ears, but could the HD558's extra harmonic distortion in the bass account for some of its subjectively greater (quantitatively) bass relative to this IEM?
-
Don't think so... HD6x0s have a bit less of that, and still sound close to the HD5x8 to me.
When Milos came over with the 009s we tried an uber-bassy track from youboobs. The 009 might have been a little attenuated relative to some of the dynamic cans I have (HD600s and such), but they still delivered skull crushing bass from that track. LOL! I remember the look on Milos when I was checking that track up... One could hear the stators going "tic-tic-tic" loudly from the outside... None of that bone crushing bass with these. Gets better if one pumps up the volume though. Could see there being a deficit in the ~45 Hz area, but who knows.
-
what was on those tracks to make the 009 dias hit the stators? 33Hz bass slams?
-
Kinda:
https://www.youtube.com/watch?v=7KRe8duwpwk