CHANGSTAR: Audiophile Headphone Reviews and Early 90s Style BBS
Lobby => IEM Measurements => Topic started by: ultrabike on May 20, 2014, 04:52:15 AM
-
Sound
I liked the mids on these ones, but missed the bass. In fact, I got the JVCs and Philips 'cuz I needed more IEM references given what I was hearing. It was perhaps that I was not getting good seal, but the tips of the JVC fitted these, and while things seemed to improve, the bass still was somewhat less than what I liked.
That said, I felt these had very good detail. The bass deal might be a relative thing. Dunno. Maybe these had an issue.
I also tried the Noble 10's, and I liked those a lot.
Note these have some wacky impedance and proly best paired with low impedance amps... used my under-reference Sansa Cheap and a FiiO 6 for measurements.
Comfort
No issues.
Presentation
Cool. I think Noble has unique look options which I thought look great.
Price
If one likes what one hears, worth it.
Overall
I know some folks like these a lot, but I prefer moar bass.
Measurements
Again, like previous IEM plots, these are not compensated (will work something out later and update). Best to look at these relative to other similar IEM measurements.
Frequency Response
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6482;image)
Distortion Right
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6484;image)
Distortion Left
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6486;image)
Impedance (20 Hz to 20 kHz)
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6492;image)
Comparo w SHE3590 and Ostry
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6490;image)
-
I remember you mentioned at the mini-meet that you thought these were rolled-off on the top. Looks like you were right.
I'm surprised to see them appear so smooth and well-balanced, although maybe I'm misinterpreting these measurements. Next to the ER4S, I felt they weren't as smooth. I thought the measurements might reveal some treble peaks TBH. I didn't have any problems with bass quantity FWIW.
If you'd like to demo/measure my ER4 after Anax, let me know. I think having plots for the Etys would be a very useful reference.
popcorn
-
I don't trust anything past 9 kHz on these, but maybe after compensation something might show up (planning to come up with compensation and update the plots sometime).
I think optimal bass quantity is sort of relative... For sure I can say that in terms of bass quantity JVC > SHE3590 > Ostrys > Noble 4... but don't have an absolute reference point yet. So far w IEMs I sort of like the SHE3590s bass quantity. Same with treble. IMO IEMs measurements seem harder to qualify in absolute terms... at least for me at this time.
I'm all for measuring the ER4s... LOL! good cute-tips, though I also felt those were bass lite for me. Dunno. Bring 'em :)p1
-
I don't think they were as smooth either, the 6-7khz peak seems to support that a bit.
-
I heard them at the Fujiya Avic festival and didn't spend much time with them since they basically had no bass, not my type of sound.
-
Holy crap, that's a ton of roll-off... I'm trying to think of how that's possible, given what's used in the Noble 4.
I was also hoping to see more of the 2.5-3k mountain, but the mountain might be obscured by a big lower midrange lift that I know Noble likes to do.
The funky impedance is expected from BAs; single driver BAs usually exhibit an exponential curve impedance that starts really low and ends really high. Usually, the first resonance peak is induced artificially, so you'll see a weird sawtooth pattern to the impedance where the resonance peak is induced. When the drivers are joined together in a low/high configuration like the TWFK is for the midrange and tweeter portion here, the midrange is actually run in reverse polarity, which is why you see the impedance drop from the crossover point with the DTEC woofers, and then rise again for the tweeter portion. Running the midrange in reverse phase also helps avoid the upper midrange chasm that showed up earlier in the 4.Ai model.
My guess is that you'd like the Noble 4 better if they chose to use the vented DTEC (called the HODVTEC), which is unfortunately new and expensive. Currently, I only know of the UE900/900s using that dual woofer from Knowles. Drivers from Sonion are cheaper, but Noble seems to like to stick with Knowles. In some senses, Knowles is the technologically superior company. Their drivers exhibit lower distortion profiles, and their treble extends better, but their drivers are expensive (nearly 2x more than Sonion).
Even so, I would not have expected the Noble 4 to roll off this much. Maybe starting from 40 Hz, but not from 200 Hz. They'd have to add a cap to the woofer to do that (either that, or the driver is broken, but how can both sides be broken?) Did you check whether the coupling was secure and sealed?
-
Yup. Made a lot of measurements to make sure. I used the tips from the JVC to do the measurements (got better seal with those for some reason).
As far as the treble, I have a feeling that after compensation the treble will come up relative to the 1 to 5 kHz range.
-
On second examination, it's probably your tubing that exaggerates the roll-off, then. I suspect that under an expensive ear simulator, the Philips and JVC would have upward rising bass (usually the case for consumer-grade earphones), while the Noble has some minor roll-off, as well as a small mid-bass hump (even though people are saying it has no bass, surely it has some, because relative to the midrange, the entire shelf is elevated).
The ER4 would probably resemble the Philips curve, except from 800 on down, it'll keep sliding down to 20 Hz.
Yeah, certainly with compensation, it'll look a lot more relatable, though I've been trying to translate these measurements to raw ear simulator measurements.
-
It's possible.
I'm not sure, but some believe the SHE3580 and the SHE3590 are essentially the same. Tyll's measurements show the SHE3580 does have a rising bass response:
http://www.innerfidelity.com/images/PhilipsSME3580.pdf
However, his ER4PT measurements seem a bit more flat:
http://www.innerfidelity.com/images/EtymoticER4PT.pdf
That said, these guys show something closer to what I measured for the Philips (perhaps even closer after compensation which might bring the upper mids lower):
http://personalaudio.ru/raa/otchety/naushniki/philips-3590-gn/
-
Yes the Noble 4 uni has a small mid bass lift compared to the ER-4S, then it starts to roll off in deep bass similarly to the ety. I also heard a large 6.5k peak in the Noble. It was large enough to affect timbre and be an annoyance. The ER-4S was much smoother overall.
-
It's interesting when people bring up bass with IEMs. It seems that most manufacturers (and users) like to see boosted bass, in order to make up for all the energy that your head/body absorb when listening to speakers or full-sized headphones. To my ears, this isn't accurate because it also messes things up tonally. Intuitively, it doesn't make sense that you can just boost bass in order to replicate that phenomena. That's why so many IEMs on the market are intolerable for me; they muddy up the bass for the sake of giving you a more "realistic" feeling of it. Maybe this is a discussion for another thread.
I think the Noble 4 did the bass pretty well. It was definitely boosted next to the ER4S, but well balanced and not excessive to my ears.
-
Perhaps. Note the plots so far have no compensation though. I made no attempt to compensate for excess of lack of perceived bass so far.
Would love to take a more critical listen and measure your ER4S OJ. Possible that by that time I will have a compensation curve ready and maybe can provide both pre and post compensation measurements.
-
I'm not saying they lack bass because of the graph. I heard then, I'm not bass head either, I think the HD800s and sr-009s have perfect bass, the 4s were definitely lacking bass and in particular any deep bass in comparison
-
Soup: For reference, which IEMs do you feel have ideal bass levels? In terms of quantity at least.
-
I'm not sure I've ever heard an iem with ideal bass response. I do like my JH-13 FPs but I know they aren't reference in the bass area. Maybe the Fitear MH334s? Still haven't had a chance to hear the UERMs
-
Perhaps. Note the plots so far have no compensation though. I made no attempt to compensate for excess of lack of perceived bass so far.
Would love to take a more critical listen and measure your ER4S OJ. Possible that by that time I will have a compensation curve ready and maybe can provide both pre and post compensation measurements.
Have you managed to develop some kind of RECD (http://www.etymotic.com/publications/erl-0091-1993.pdf) curve yet? Just curious.
-
I have an idea of what to do using some full size headphones and the ear coupler, but haven't done it yet. Thanks for the paper link. Will read.
-
FWIW, I seem to recall reading in one of Sean Olive's papers that subjective tests indicate that people are more sensitive to too much bass in IEMs than regular sized headphones. Seems bassakwards to me...that's why I remember it. But if he said something I take note.
-
Shut... I'll do some comparos in that area after I get a compensation curve. Many of the dynamic IEMs, even the more "neutral", do sound a little bass heavy (with proper seal). That said, many of the BA's sound lean to me, including the Audeos which seemed pretty analytical. Only the uncomfortable as hell TF10s of the BA type IEMs I heard were a little bassy ... Well, the Noble 10s maybe, but didn't spend to much time w them.
We'll see.
-
OK. So I tried creating a RECD curve. Comp curve (what I would have to add to IEM measurements) looks a bit too wacky to me:
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6759;image)
(dB in the y-axis, and Hz in the x-axis)
I will not currently apply it because I'm not convinced its correct, and because I have a feeling that different ear models might yield different results. I feel it's proly best to use FR results relative to other FR results from a similar rig and methodology (specially for IEMs).
-
I will not currently apply it because I'm not convinced its correct, and because I have a feeling that different ear models might yield different results. I feel it's proly best to use FR results relative to other FR results from a similar rig and methodology (specially for IEMs).
Agreed on that count.
Although it would still be interesting to see what the curves look like once compensated.
-
Sure, this is what I get for the Noble 4 (20Hz to 10kHz):
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=6777;image)
However, I don't trust that compensation. It makes it seem as if the Noble 4 was more loss in the upper midrange and lower treble than I think it should (if at all). Will have to work some more on it...
-
Will be very interesting to see how your plots look w/ compensation for a more known quantity, like the ER4S or UERM.
-
(http://i.imgur.com/kWWTJUNl.png)
(http://i.imgur.com/i34KxIgl.png)
(http://i.imgur.com/JAX7xirl.png)
(http://i.imgur.com/wDFu0TFl.png)
Not my measurements; the guy who did them is a manufacturer so he doesn't want any conflict of interests. I have the original PIR files, so I can output them any way I want.
Looks okay, pretty much what I expected; for accuracy the 2.5k peak should be larger, at least 3 dB higher but this is a mid-centric IEM anyway so whatever, and the 6-7k peak is definitely there but isn't as big --- that peak in the TWFK always gets exaggerated in the IEC711. Bandwidth is decent for BA.
Anyway, a much better improvement over the 4.Ai.
-
Yeah, I definitely heard that peak. It ruined the treble tonality and timbre for me. And that bass looks awfully similar to the ER-4S. I heard them pretty similar there with the noble having a bit more mid bass than the ety. But definitely significantly better overall than the heir 4.ai
-
Bass roll-off is greater in the ER4. It probably just looks similar due to the way the graph is rendered, but in a direct overlay, the rolloff in the ER4S should be steeper. But yes, there should definitely be more mid-bass presence.
In the custom 4C, the treble peak is surprisingly tame, probably less than half of what you see on the graph.
-
I'll overlay the ER4S and Noble 4s FRs. I also have another set of Noble 4s and will take some measures tonite.
-
In the custom 4C, the treble peak is surprisingly tame, probably less than half of what you see on the graph.
Agreed, it's less of a peak on my custom 4C.
-
So... Noble 4 vs ER4S (previous uncompensated FRs). The plots are lined up ~1kHz
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=7331;image)
-
IME, it should probably be normalized to a point of "inflection" below the effect of canal gain starts at around 600 Hz (even though your rig doesn't account for canal gain), so here it should probably be 400 Hz or so. The ER4S rolls off slightly steeper even when normalized to 400 Hz. It doesn't look like much, but when the SPL at midrange frequencies exceeds threshold levels, the lows drop off below the realm of threshold of hearing for the sub-bass more easily for the ER4S.
-
Here is a second set of uncompensated Noble 4 measurements. These are Greed's. Thanks mate! :)p5
Frequency Response:
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=7493;image)
Distortion Right:
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=7495;image)
Distortion Left:
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=7497;image)
Comparo with previous Noble 4 measurements:
(http://www.changstar.com/index.php?action=dlattach;topic=1604.0;attach=7499;image)
The first Noble4s I heard and measured had regular tips. These were heard and measured with double flange tips.
-
The 2nd set of measurements look a little closer to what I hear on my 4C, i.e. larger 1k bump that brings mids/vocals close. The larger 6.5k peak is expected with the TWFK (WBFK half) tweeter whenever measured with silicone tips, though I think the peak feels relatively controlled in the custom 4C (corroborating the findings of Kunlun the cat). I've yet to come up with a plausible explanation of what certain resonance peaks sound less prominent when the acoustic output interface encompasses a large and solid surface (as with CIEMs), as I find the phenomenon has less to do with insertion depth and residual canal volume than it does with compression of quarter wavelength resonances when reflecting away from an irregular surface (soft silicone tip).