These drivers have been anticipated by an even denser than usual storm of hype-y marketing content. All sorts of information have been released before and together Kbear BELIEVE launch, and the first “reviewers” spent nothing short of bold words about these.
Alleged magnificent features apart, a major disclaimer is one way or the other present on all articles : “BELIEVE require huge amping”. Of course I checked on that. Long story short: BELIEVE do require an uncommon source in a sense, but this is not about power figures as (milli)Watts written on the amp spec sheet, rather about quality of power delivery. You’ll find a detail of my analysis further down below.
This is very significant to our case today as – to me, at least – BELIEVE’s sound output when an inappropriate amp is being used is a total deal breaker. I pinpointed where and what I hear in such cases in my analysis down here.
BELIEVE is also sensible to cable and tips selections. Not at the same “dramatic” levels applicable to amping, but much more than other IEMs nonetheless.
Let me pass right away to my notes now, I’ll add more considerations and my conclusions at the bottom of the article.
Ah no, another dutiful thing first: I got this pair of KBEAR BELIEVE as a review unit from my friends at KEEPHIFI (www.keephifi.com) entrusting me to an unbiased analysis and openhearted subjective evaluation, which is what I’m reporting here below. You can purchase BELIEVE at their store.
Ok, off to the notes now.
At-a-glance Card
| PROs | CONs |
| Soundstage and imaging. | Prentetious about amp pairing. Sound “cheap” on inappropriate sources. |
| Snappy bodied midbass. | Thin, artificial timbre on high mids and low trebles. |
| Lively trebles. | Not cheap asking price. |
| Support quite a few different musical genres. | Cable and tip sensitive. |
Full Device Card
Test setup
Questyle QP1R / Sony NW-A55 mrWalkman DMP-A50R / Apogee Groove / Auglamour GR1 / ifi ZEN CAN / Hiby R3Pro / Hidizs AP80Pro / Fiio X3 mk-III / Lotoo Paw Gold Touch – Final E-clear tips – Nicehck 16core High Purity Copper cable – lossless 16-24/44.1-192 FLAC tracks.
Signature analysis
| Tonality | Timbre is almost natural with a light wam tint on top, with the exception of the highmids-lowtreble section sounding a bit artificial. Tonality is quite nicely calibrated, mildly warm, clearly U shaped, credible. |
| Sub-Bass | Rolled down and slower compared to midbass, averagely textured. Sub-bass and low-mid-bass are one of the two areas where inappropriate amping quality is more evident (they turn sloppy and very bleeding in such case – in the worst cases the effect extends way into the midbass) |
| Mid Bass | Mid bass is definitely pleasant, relatively fast decay (attack is not snappy though), with good texturing. |
| Mids | Mids are flatly recessed but not for their entire extension, their high part goes up quite quickly. Low and central mids are lean; within that I find them nice. Mids in the higher part of the segment are instead thin and deliver a somewhat artificial timbre. |
| Male Vocals | Male vocals are quite nice, too lean to aspire to a “vocal rec” score for my taste, yet still natural, good. |
| Female Vocals | Females are also nice, and also lean tendent to thin. They often get close to sibilance although very rarely I heard them pass the treshold (depending on tracks and source ofc). |
| Highs | Trebles are forward and lively, quite energetic too. I find them “cold” if I take them per se, thus somewhat incoherent with the general warm-ish tonality. They also would do with a bit more air. I dont hear any dreaded metallic zings around which is very good. Lowtrebles are too thin and share a somewhat artificial (“plasticky”) timbre with highmids. |
Technicalities
| Soundstage | Soundstage is quite wide and height is beyond decent. It takes a higher end source to widen it further and especially to hear some depth which is just hinted or missing on lower tier sources. |
| Imaging | Imaging is nice although it tends to a L-R distribution in lieu of a more natural X-Y casting. Inappropriate amping will severely screw imaging and instrument separation in the high mids, making them nothing short of “a mess” on crowded and/or loud orchestral passages. |
| Details | When sub-bass bleed is controlled by appropriate amping bass detailing is quite nice. Treble details are average for the asking price. |
| Instrument separation | Separation and layering are in general well done, see “Imaging” for exceptions. |
| Driveability | Extremely capricious. Long story short: they sound inacceptably bad on lowend sources, so much that I strongly recommend assessing them on existing gear before purchase. Don’t ever take “reviews”, including this article of mine, as closely representative of your possible feelings about audio gear – and if you want to make exceptions don’t start with BELIEVE (pun intended)! They do get sensibly better with appropriate amping quality. |
Physicals
| Build | Very nice housings both in terms of construction, sturdyness and style. |
| Fit | Housing shapes are just a tad too big to seemlessly fit my concha which is a true pity but hey, we are all different. Nozzle length is ok for me, and insertion is easy. Although generous, stock tips’ supply is not adequate to my ears; after repeated tedious rolling sessions I settled on a pair of final E clear tips, which in spite of their realtively narrow bore don’t impact negatively on trebles here, and contribute to further control the very low end “sloppy” tendence. |
| Comfort | Due to the housing being just a whiff too big and not lightweight (not too heavy either, luckily) I can’t say they are supercomfortable to me. |
| Isolation | Good passive isolation once housings are properly fitted |
| Cable | Stock cable is not bad per se. Oddly enough I find cheap Nicehck 16core High Purity Copper having some positive impact on the highmids weakness I mentioned above. |
Specifications (declared)
| Housing | Five-axis-CNC “aviation grade” aluminum shells |
| Driver(s) | Pure beryllium diaphragm dynamic driver (imported from Japan) (* I’m unable to assess such alleged “purity”) |
| Connector | 2pin 0.78mm |
| Cable | 1,2m 4 strands 6N single crystal copper Litz cable, with 3.5 single ended termination |
| Sensitivity | 98 dB |
| Impedance | 17 Ω |
| Frequency Range | 20Hz-20KHz |
| Accessories and package | Elegant & practical leatherette carry case, 1 set (S/M/L) wide bore soft umbrella black silicon tips, 1 set (S/M/L) wide bore stiffer umbrella white transparent silicon tips, 2 pairs of foam tips |
| MSRP at this post time | $360,00 ($180,00 apparently “permanent” deal price) |
How bad do Kbear BELIEVE really need “a good source”?
Simply put, it’s very serious.
As I mentioned at the beginning, I found so many saying it’s a matter of sheer amping power. Of course I knew it very likely was not that, or more precisly not only that. In facts you do not need a source with a spec sheet mentioning “multiple watts” of output power to take the best out of Kbear BELIEVE (or any other IEM for that matter, even those featuring much lower sensitivity compared to them).
What you do need is a source delivering an adequate level of current, that’s sure, and capable of delivering it in all critical sections of the spectrum exactly when needed.
Let me offer a synthetic view on my excursus on a total of 9 different amping sources, ranging in price from $50 to $3500, approximately, and from 8.8mW to +-2W of output power face value.
For reference, as a test track I’ve purposefully taken “Limehouse blues” from what many consider the best jazz master of the XX century being “Jazz at the Pawn Show” recording, which I happen to have available in 24b/192K flac format. The song’s critical elements which are significant to Kbear BELIEVE’s powering capriciousness are the constant contrabbasso’s line, which of course be delivered as punchy, detailed and not bleeding, and the “treble-layered-crescendo” starting at +-7:00, which should offer good instrument separation, and no metallic zings or other artifacts.
For Single Ended sources the stock cable has been used. For Balanced Ended a NiceHCK 4N 23AWG High Purity OFC has been adopted.
Here’s what I heard on each of the tested sources:
| Apogee Groove |
| Too soft on the x-lowend, and on the x-highend. The contrabbasso hits deeps but vibes in a relatively fastidious way. In the treble section clarinet and cymbals steal the scene a bit too much vs the vibrophone, some separation is lost and some microdetail is gone. Anything between (roughly) 100 and 8K is good (within Kbear BELIEVE limits) and spatial drawing is “wow” level, as always when Groove is involved. |
| Mixed feelings face. |
| Fiio X3 mk-III |
| Single Ended output : Loudness is there, bass is boomy, treble looses definition, stage is meh. Low or High gain, same stuff more or less. Balanced Ended output: (Low Gain) Contrabasso punches more softly and bleeds less. The crowded-treble section is not bad at all, yes separation is not masterful but way better than other entry-level sources (see below). (High Gain) waaay better on contrabbasso, “almost” nice actually. Trebles get a better control not much, separation also goes up a bit, but trebles also get dangerously near some thin-metallic nuances in the upper above frequencies. Finally, too bad for the range which is evidently compressed. |
| SE: semi-yuck face. BE: perplexity |
| Questyle QP1R |
| (Low Gain): sub-bass gets within nice limits, the contrabbasso hits “maybe” (I’m not sure) a tad less deep but its countour is definitely sharper, better than the one produce by Groove. What changes completely is the treble part where the vibrophone on second layer comes out much better, louder and detailed, and cymbals are apparently not “pasted”. The counter of this seems to be lower resolution in terms of “greys in the middle”: if Groove lets me hear (say) 1000 nuances, QP1R only delivers like 5-600. Spatiality is also evidently less “woah-level” compared to Groove but if I had to give a comprehensive score I’d prefer QP1R, under which Kbear BELIEVE sound “convincingly well” amplification-wise. Switching to Mid Gain doesnt change sound, just gives a bit more of loudness headroom (which I dont need). Switching to High Gain compresses sound a (slight but perceivable) bit. |
| Smile face |
| Auglamour GR1 (Groove or QP1R as source) |
| Xbass loudness and definition is on an in-between level vs amping from either Groove or QP1R directly. Treble section rendering is worse than Groove direct though. |
| Yuck face. |
| iFi ZEN CAN (Groove or QP1R as source) |
| Single Ended (0dB Gain) : X-bass levels are ok, haloing is similar to GR1, so above decent at least; treble section control is very good, layering is preserved properly. Switching gain up makes these 2 aspects even a tad better without reaching QP1R quality though. For the record the best compromise seems 12dB. X-bass switch screws up on this aspect of course (as expected). On the other hand ZEN CAN’s SE channel kills 80% of incoming dynamics so the overall sound output is dull and mushy, a shame compared with QP1R / Groove native outputs. Yuck face (at ZEN CAN) Balanced Ended (0dB Gain) : We can’t even start talking about this. ZEN CAN’s BE channel inherently makes subbass sloppier which is obviously the opposite of what Kbear BELIEVE need. Big yuck face (not their fault, but gets way worse on them than, say, on Oxygen or OH10). |
| Yuck face (mostly at ZEN CAN tbh) |
| Sony NW-A55 / MrWalkman DMP-A50R firmware |
| (Source on “DSEE AI”; DC Phase Linearizer on Type A High, High Gain) : Well this is… Sony. The contrabbasso is less punchy and less bleeding. Switching DSEE to “Percussion” the contrabasso becomes a razor but the rest is too edgy even for my tastes. DSEE AI remains the best compromise (as almost always), and also renders the treble part very well. Spatiality is roughly on QP1R level so way south of Groove. |
| Smile face |
| Lotoo Paw Gold Touch |
| The source is per se supremely smooth and nuanced. Recuping my previous analogy if Groove delivers “1000 greys” and QP1R only 5-600, LPGT has like… 2000. Some, me inclued, may consider them even… too many 🙂 Contrabbasso is evidently less punchy, not boomy but not much more detailed than on Groove. On the treble side, the vibrophone on second layer comes out much better not because (like QP1R does) the vibe itself is given more authority, but because the clarinet is way smoother, nuanced, silkier thus leaving more space for the background voice to come out. Stage and imaging are +- at Groove level, maybe a tad less but top of the rack anyhow |
| Smile face (…at this price, better be!!) |
| Hiby R3 Pro |
| Single Ended: bass is worse than X3-III. Boomier, less defined, almost distorted, in a word crap. Same for the treble-dense part where separation goes nuts. Stage and imaging I just avoid describing… Balanced Ended: (Low Gain) xbass is a better than SE, but not decisevely. Possibly marginally better than X3-III SE but it’s probably even debateable, it’s certainly worse than X3-III BE. Layering and separation in the crowded-treble part is way better than R3Pro SE but still far from good. In general, I hear a sort of “grey veil” blanketing everything. Stage becomes “intimate”. |
| SE : big yuck face BE : yuck face |
| Hidizs AP80Pro |
| Balanced Ended: (Low Gain) contrabbasso deepest punches are still invasive but not all punches are deep enough too be. The crowded-treble part is better than R3Pro BE in terms of levels: back voices (vibe and piano) come out louder and stand better vs foreground ones, but separation is still meh so the result is a quite messy salad anyway, just louder. Part of the problem seems to be that – always in the “grey levels” analogy I made before – if LPGT has 2000 levels, Groove 1000 and QP1R 5-600, AP80Pro has just 200, and this shows big time on Kbear BELIEVE (and on other drivers too). Stage is averagely wide, and has at least “some” depth. Dynamics are “basic”, but existing. (High Gain) This makes contrabbasso situation worse. It goes up in loudness but also down in speed, increasing the bleed. The crowded treble phase, instead, is better, if not that much. Dynamics are I would say on par. |
| Yuck face |
Conclusions
When properly paired, Kbear BELIEVE are not bad at all. I’m not particularly fond of their timbre, especially that artificial-flavoured nuance in the high mids, and details are nothing more than average at their asking price, but the rest is more than OK and although not “inexpensive” they might be a sidegrade to assess for many.
I feeel there’s more to say though.
As I reported above, Kbear BELIEVE are extremely sensitive to source quality (DAC and AMP capacities).
No you do not need a multi-Watt amp to enjoy Kbear BELIEVE at the best of their abilities, and not even a mid-high powerful one. Even a quite low powered and definitely not kilobuck-priced DAP like Sony NW-A55 can get “correct” amping into BELIEVE and good sound quality out of them. Again, it’s not a matter of (big) amping power, it’s a matter of delivery quality.
However, you do need an “adequate” amp for Kbear BELIEVE not to sound bad. Bad as in “cheap”. “Deal-breaking” bad, if you ask me. Ymmv, needless to say, but I-myself would not buy them after auditioning them paired on an R3Pro, for example.
As I am a quite fortunate individual, I had the opportunity to assess Kbear BELIEVE on nine different sources ranging from a nice inexpensive portable amp to a multi-kilobuck TOTL DAP. Three of my sources could drive them properly – “BELIEVE bless…” – let’s talk about the rest though.
First: Kbear BELIEVE are not “specially well biased” by Apogee Groove. Given how spectacularly Groove feeds just about all other most capricious IEMs I own, I’m not going to book this situation as Groove’s limit rather as a testimony of how “unpredictable” (to be very mild) BELIEVE’s amping requirements are. I.o.w. the average BELIEVE user is called to be “un-rationally” careful about pairing.
Second: which parameters do I have to take to give a comprehensive opinion on Kbear BELIEVE’s “value”, meant as the audio pleasure they deliver vs their asked price & pairing effort ?
Should I give all the source pairing effort for granted, as I wrote above Kbear BELIEVE sound nicely on A55, and even “very nicely” on QP1R. I’m not in love, but others might.
If I think to a few other drivers I personally like – on the same or different timbres – in the same price bracket and I imagine to apply the same enormous patience I dedicated to finding Kbear BELIEVE a suitable source pair, well… their score will go down very sharply though.
Final E4000 on Groove or QP1R are not “just very nice”, they are “monumentally good“. Same for Ikko OH10, and for Shuoer Tape (these last on QP1R or A55 only as Groove can’t bias them by design).
Oppositely, if I pair E4000 / OH10 / Tape with the same sources which make Kbear BELIEVE sound downright “bad” – let’s say an R3Pro – I do get lower output quality of course, but not so dramatically worse result compared to what comes out of BELIEVE.
So my bottom line is Kbear BELIEVE are just OK if you apriori decide you want to love them. Their main “flaw” is their pretentiousness.