64 Audio U18t Measurement and Analysis

Discussion in 'IEM Measurements' started by purr1n, Jan 31, 2021.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    @rhythmdevils has covered these already from a subjective point of view, quite comprehensively, comparing to other TOTL IEMS from CFA, even applying EQ to the U18t to match frequency response to properly assess "technicalities": https://www.superbestaudiofriends.o...ding-andromeda-ara-og-solaris-and-more.10349/

    I've skimmed over @rhythmdevils posts, but did not read them in detail, or to the point where I remember exactly what he may have said. This was intentional because I didn't want someone else's observations to totally skew mine. This thread is intended to be more of an objective analysis, to see if any measurements taken and presented correlate to perception. Per my usual procedure, I listened to them using various tracks, and jotted down some notes before I took the measurements. There's some very interesting stuff here to be sure. FWIW, I use the CFA Ara IEM on a regular basis and as such consider this to be my reference. (A reference is just that, a reference. It does not mean awesome in every way known to humans).

    The U18t's tonal response is bassy (below mid-bass), slightly laid-back in the upper mids, and gentle in the highs. All of the CFA IEMs (Andro, Solaris, Ara) have more top end air, where the shimmer of cymbals and ambient cues are more evident. I would not say that one or more resolving than the other however as these are simply tricks of frequency response. The U18t particularly excels in timbre and fluidity from the midrange on up. In many respects, this aspect reminds me of the Soranik Mystery IEMs which I recently covered.

    If there is any fault in the upper registers, it may be that the U18t tends to impart sort of its own timbre or overtones, in the upper mids and lower highs. It's hard to explain because all gears impart their signature, but do I think this homogenization is a bit stronger with the U18t than with most other high-end IEMs I have heard. However, this is a super nitpick. The plus side is that highs are smooth and less grainy and prickly, which the CFA BA drivers can be at times. As such, I feel the U18t is much less picky with respect to source and amps. (I have preferred sources such as the Sony ZX2 and now Shanling M6 with the Ara and Andromeda IEMS). One thing that I do want to emphasize is that the midrange of the U18t is darn good if not fantastic.

    Unfortunately, now is where the fun stops. The bass sucks. I am a closet basshead, so I do not mind strong bass. The bass emphasis is just below the mid-bass and doesn't bleed into the lower mids, maybe it doesn't bleed enough as male tenor voices seem a bit on the cool side. The problem is that the bass isn't of good quality. The best I can describe it is this:

    Imagine that you have small wide-banders that presented the mids beautifully, for example, the 4" Fostex Sigma or the 8" Voxactivs (OK, just pretend there is no rising response or listen to it off-axis). Then pair these with a cheap Hsu VTF series subwoofer with a port tuned too high and sound level cranked up slightly too much.

    The problem is not impact, it's not grunt. It's indistinct one-note bass. The upper bass isn't go bad. The problem seems to get worse as we get lower in frequency that well before we get to the low E on a bass, the pitch differentiation is already garbage. I've probably gone over most of the songs in Daft Punk's Random Access Memories learning the basslines. There are a couple of tracks that I use to assess aspects bass quality, and the U18t failed miserably.

    I had an upright long ago and today I still play fretless (electric), so one of the things I always want to hear is pitch. The U18t fails so badly here that I can barely hear the pitch differentiation among the G#, B, D# (38Hz) bass drops in Doin' It Right. The progression pretty much goes fart fart fart. Same thing with Beyond, at least the notes on the E string. Get Lucky is only a bit better, but only because most of the bassline is on the A string. I simply cannot express how disappointing and such a let down it is to not hear the actual bass notes in such funky songs. But really, it's not just Daft Punk. Eva Cassidy's cover of Son of Preacher Man from the Nightbird album exhibits the same behaviors. It's such a shame because that bassist she has in her band is killer, but he may as well been playing the same E string over and over, blob blob blob. And let's not talk even about the drums. I've love to read what @Psalmanazar would say about this if he ever got this hands on the U18t.

    TBH, after continued listening for a second day running, the shitty bass didn't bother me as much as I got used to it. And really, the issue seems to be most evident below 50Hz (anything lower than the open G string on a bass). However, when we assess TOTL gear, we must compare to other TOTL gear. It's way too easy to not know or forget what we are missing. In a nutshell, upper bass to mids are definitely TOTL high-fi, highs mid/high-fi (lacks top end extension), and sub/low bass is low/mid-fi.

    Measurements to be presented...
     
    Last edited: Jan 31, 2021
    Bozoskeleton, Lyander, 9suns and 11 others like this.
  2. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    64 Audio U18T
    Frequency Response
    upload_2021-1-31_12-8-27.png

    64 Audio U18T vs CFA Ara (GRY)
    Frequency Response
    upload_2021-1-31_12-11-11.png

    64 Audio U18T vs GAudio Nair (GRY)
    Frequency Response
    upload_2021-1-31_12-12-58.png
     
    Lyander, 9suns, vtrinidad87 and 6 others like this.
  3. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The distortion plots below are plotted differently than what we usually see. The distortion is plotted at the frequency where the harmonic "activates", that is where we hear it manifest, as opposed to the fundamental frequency of the harmonic.

    64 Audio U18T Distortion Behaviors
    upload_2021-1-31_12-14-53.png

    These may explain a lot of what I'm hearing. Note I've provided Ara results at the bottom for reference and comparison.
    • INTERESTING: The bunching of the third harmonic together regardless of SPL without distortion much being below -70db indicates use of BA drivers.
    • GOOD: A good amount of second harmonic in relation to the third harmonic. Usually with BA drivers, third harmonic will dominate from the mids on down. Note that dynamic and ortho headphone drivers will tend to have higher D2 than D3, until pushed hard where the D3 catches up. With BA drivers, the opposite tends to happen. The U18T distortion behavior is somewhere in between. This probably explain why I love the U18T mids so much.
    • GOOD: The distortion pattern, especially the second harmonic, is fairly smooth no radical ups and downs. Again, possibly another reason I felt the mids sounded so right and less grainy (also note drop of third harmonic past 5kHz).
    • INTERESTING: the lower midrange dip of the third harmonic: could this explain why at times I felt the lower mids to sound cool despite no indications otherwise in the FR?
    • NOT SO GOOD: Very high levels of distortion in the lows, particularly in D2 as we crank it up. Very high levels of D3, even when we don't crank it up. Could these, particularly the rapidly rising D3, explain why I felt bass pitch differentiation was so subpar with the U18T?
    • CAUTION: I plan on doing more of these in-depth distortion analysis in the future. The more data we have, the better. The data is interesting, but let's not rush to hard conclusions like objectivist audio fanatics.
    • FUNNY: As an American, I don't know shit about IEMs because we don't have IEM stores on every street corner where we can demo 813 IEMs, 213 of them with TOTL pricing, and of these, 207 sounding NOT-TOTL.
    For reference:
    Campfire Audio Ara Distortion Behaviors.
    upload_2021-1-31_12-13-57.png
     
    Last edited: Jan 31, 2021
    Lyander, 9suns, vtrinidad87 and 5 others like this.
  4. limesoft

    limesoft Facebook Friend

    Joined:
    Oct 4, 2016
    Likes Received:
    189
    Dislikes Received:
    0
    Trophy Points:
    43
    Location:
    London
    I really was not a huge fan of these when I heard them, there was something off about the treble - too much of it, a bit etched and uneven at it's highest registers (I think 10khz+), exacerbated by the weird one note bass which you mentioned. I found the U12t a bit more balanced/cohesive, but still bit sharp/uneven with the top treble. I ditched the U12t and kept my Andromeda which I find more cohesive and more worth the money.
     
    Baten likes this.
  5. rhythmdevils

    rhythmdevils Best SBAF member of all time

    Staff Member Friend
    Joined:
    Apr 15, 2020
    Likes Received:
    2,827
    Dislikes Received:
    49
    Trophy Points:
    113
    Location:
    Bay Area
    Love your analysis of these distortion measurements. Very helpful, interesting and insightful.

    I Also found the, well at least the upper midrange, to be excellent in the U18t. In my review it was the only thing I thought outperformed the CA iem’s. I don’t extend this appreciation to the entire midrange though. Just the upper midrange.
     
  6. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Animated GIF, flipping between Ara and U18T
    Webp.net-gifmaker.gif

    I am not sure what the designer was trying to do with the 8 mid drivers and 8 low drivers. Perhaps the goal was to get a distortion profile more similar to that of ortho and dynamic drivers that we with headphones are more used to. Unfortunately, the resultant distortion is quite a bit on the high side.
     
  7. Degru

    Degru Facebook Friend

    Joined:
    Jul 22, 2017
    Likes Received:
    129
    Dislikes Received:
    4
    Trophy Points:
    33
    Location:
    Portland, OR
    What are the advantages of looking at distortion referenced to the harmonic frequency?
     
    purr1n likes this.
  8. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    It's merely a different way of looking at things.

    What is harmonic distortion? It's basically an extraneous signal, a multiple of the original signal, that should not be there.

    For example, the second harmonic distortion of a 50Hz tone would be 100Hz. The third harmonic is 150Hz. For single tones, harmonic distortion will be the dominate form of distortion, particularly the early orders, second and third.

    In the traditional way of plotting, when we see a third order distortion plot, we must multiply in our heads the frequency on the x-axis by two or three (respectively for second and third harmonics) to arrive where the distortion actually manifests. For example, third harmonic distortion at 200Hz does not mean there is an extraneous tone at 200Hz. It means that there is a tone at 600Hz that should not be there.

    What this alternate visualization does is do the homework for you. It shows us where in the audio band that the distortion, the signal that shouldn't be there, manifests.

    The downside of course is that we now need divide by the order of the harmonic to know the the frequency of where the original signal resided.
     
  9. Degru

    Degru Facebook Friend

    Joined:
    Jul 22, 2017
    Likes Received:
    129
    Dislikes Received:
    4
    Trophy Points:
    33
    Location:
    Portland, OR
    Hm, makes sense. Depends how you think about distortion I suppose. "If I play content in X frequency it will distort this much with these harmonics" vs "During full-range playback I will get extraneous tones at these frequencies by this much". Personally the former makes more sense to my brain tho.
     
    purr1n likes this.
  10. Biodegraded

    Biodegraded Friend

    Friend
    Joined:
    May 28, 2017
    Likes Received:
    4,804
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    Vancouver BC
    I like 'em both. Complementary.

    @purr1n I do like seeing the different harmonics at various SPLs as you've displayed above. And while I'm here, a request: could you measure the impedance profile for these?
     
    Lyander likes this.
  11. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I got that "etched" sense too, but in the end (subjectively) decided that is was less of a peak as opposed to a certain timbre. I did mention that there was a homogenization of timbre. There's nothing in the measurements that totally explains this though (not that there always should).

    Of course! IEM measurements, particularly BA drivers, should always come with impedance measurements. We all know how crucial this is based on our experience.

    Now this is interesting. Ugh. I know we've critical of CFA impedance curves, but this is not good, arguably worse even. The U18t is going to be very reactive to output impedance. Higher Zout sources will increase two spots in the upper mids and lower highs.

    U18t Impedance
    GRN = Left Ch
    YEL = RIght Ch
    upload_2021-2-8_10-32-28.png
     
    rhythmdevils, jexby and Biodegraded like this.
  12. Biodegraded

    Biodegraded Friend

    Friend
    Joined:
    May 28, 2017
    Likes Received:
    4,804
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    Vancouver BC
    Various sources relative to a flat nominal 0 Ohm output Z (red):
    Orange = 1.5 Ohms (various Sony DAPs)
    Green = 2.5 Ohms (Ifi iEMatch dongle)
    Blue = 3.5 Ohms (AK100II, iEMatch output on Ifi AOIs)
    Purple = 5.5 Ohms (Sony DMPZ1)

    [​IMG]

    Not recommended with yer desktop tube amps :eek:
     
    jexby and purr1n like this.
  13. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    72,364
    Dislikes Received:
    115
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Also not recommended with ZX2, M6 (balanced out), Pixel 4A 5G, etc. It's a reverse Andromeda situation. Higher Zout to a certain extent shaved off the increasing bass response of the Andro. In this case, higher Zout, just makes the U18t more U/V-shaped.
     
    Biodegraded likes this.
  14. rhythmdevils

    rhythmdevils Best SBAF member of all time

    Staff Member Friend
    Joined:
    Apr 15, 2020
    Likes Received:
    2,827
    Dislikes Received:
    49
    Trophy Points:
    113
    Location:
    Bay Area
    I only thought they were moderately listenable by EQ.
     

Share This Page