64 Audio tia Trió IEM Review

[Blorg]

Hi

If I am to read the provided graphs, the Knowles curve starts deviating from Harman around 2.5 KHz, that is "brightness" territory; from that I can also infer that I would not tolerate such a curve in an IEM. I seem to be quite Harman-compliant .
Subjectively, I find the TCZ, shouty and peaky. Tolerable, but a bit of EQ to reduce the treble, even a gentle (negative) high shelf around 3 KHz , say -1.5 dB can help... I don't care enough to EQ but have tried a few, in particular those from @Maiky76 and @Chromatischism and they address some of those issues.

Peace.

This is because they started with the Harman curve below 10kHz and then looked at modifications 10-20kHz. They didn't look at changing Harman below 10kHz for this specific research, they took it as given. The consequent increases below 10kHz are thus there from their equalizations to Harman above 10kHz and the existing Harman curve, any equalization is a curve and spills over into other frequencies.

If you are that sensitive to even tiny sub-1dB increases at 2.5kHz and find the TCZ shouty and peaky, it sounds like even Harman is probably too shouty for you. This is very common, many people find Harman in-ear too shouty and fatiguing in the upper mids.

I also find this, this is why I prefer IEF neutral through the upper mids, like the Blessing 2 Dusk or Monarch Mk2. According to Jaakko of AutoEQ, the Dusk is the closest wired IEM in existence to Harman, but even it stays well below in the upper mids. Or the Variations, which although hits the initial Harman rise to 3kHz then drops below it, and is the better IEM for it.

Both of these IEMs, among the closest to Harman of anything, are actually below Harman through the upper mids, but then maintain much better treble extension. That's really what I'd be after, too, I don't want more in the 3-8kHz region. But I'll take it over 10kHz.

 

[Docmoggy]

I'm just thinking what you could buy instead of these IEM's. I mean you could buy 33 Truthear Hexas! It would be great to see the results of a blind test between a Hexa and the aforementioned! The test format could possibly involve a random selection of individuals and subject them to alternating IEMs. The format could even be expanded to comparing crazy premium (crazy price) to low budget (cheap) IEMs. Wouldn't that be interesting data? However, looking at the H plot, Hexa for me!

 

[Blorg]

I'm just thinking what you could buy instead of these IEM's. I mean you could buy 33 Truthear Hexas! It would be great to see the results of a blind test between a Hexa and the aforementioned! The test format could possibly involve a random selection of individuals and subject them to alternating IEMs. The format could even be expanded to comparing crazy premium (crazy price) to low budget (cheap) IEMs. Wouldn't that be interesting data? However, looking at the H plot, Hexa for me!

This IEM sounds totally different from the Hexa, it has gobs more bass. You wouldn't confuse them.

 

[staticV3]

This IEM sounds totally different from the Hexa, it has gobs more bass. You wouldn't confuse them.
View attachment 268224

Yeah but which one would be preferred by the majority of people in a blind test? I think that's what Docmoggy would like to know.

 

[markanini]

Yeah but which one would be preferred by the majority of people in a blind test? I think that's what Docmoggy would like to know.

A blind test would be interesting because the compensation used by Amir hasn't been tested in this way.

 

[Blorg]

Would be interesting. For me, the Trio has far more "wow" factor, it has fantastic bass and much larger sense of soundstage.

I actually tend to find the dryer more neutral stuff is more something that grows on me over time, stuff like the Monarch Mk2, or in headphones, the HD650, it's not an initial "wow" but a more long term, "this just sounds right" that I come to appreciate over time.

If I could only keep a single IEM I'd pick something more neutral. But I'm not sure this would do badly at all in a blind test against the Hexa, which although very well tuned can come over quite dry, thin. anemic. It would depend on people's preferences I guess but for a random sample of the public I'd put my money on this one.

 

[Docmoggy]

Would be interesting. For me, the Trio has far more "wow" factor, it has fantastic bass and much larger sense of soundstage.

I actually tend to find the dryer more neutral stuff is more something that grows on me over time, stuff like the Monarch Mk2, or in headphones, the HD650, it's not an initial "wow" but a more long term, "this just sounds right" that I come to appreciate over time.

If I could only keep a single IEM I'd pick something more neutral. But I'm not sure this would do badly at all in a blind test against the Hexa, which although very well tuned can come over quite dry, thin. anemic. It would depend on people's preferences I guess but for a random sample of the public I'd put my money on this one.

Statistically speaking, I would expect a product that is a numbing 33 times the cost of a Hexa should be so easy to differentiate that 100% of people would make it obvious their preference. However, if we are comparing a H plot to be the main source of product differentiation, then it is perhaps likely that there would be some kind of split precisely because, as you point out, a reasonable quartile of people may actually not appreciate the output from the Trio. If there is then it makes the cost premium of the Trio highly controversial indeed. Therefore, in a subjective arena it is highly probable that, like Amir concluded, the epitome of sound may not indeed be a question of cost, but application and complexity is not a guarantee of sonic brilliance. I mean sonic brilliance will always be an unknown to me as I do not have brilliant hearing. What I can rationalise is excellent sound and cost that is concomitant with the Hexa. In conclusion, I would like to see a survey of IEMs, perhaps bundled into an analysis of variance format to illicit the unknown - what are the factors that contribute towards the ultimate IEM.

 

[Maiky76]

Hi, I don't appreciate the sarcasm thanks.....rather than aggressive belittling sarcasm you could instead have just replied that you'd included it in your post and then posted up a seperate graph showing the difference between the two in your reply so that others that had missed it could be enlightened too - but no you didn't do that. In your EQ post it wasn't overly clear in your wording preceding the graph, you could have praps said simply "Let's have a look at how the Knowles Target is different from the Harman Target" and then included a seperate graph after that, rather than lumping it in with your other graphs. Yes, I'm making excuses, but it wasn't overly clear in your post to me if you're trying to read it quickly, for those that missed it here's the Knowles Target vs the Harman Target, and the following graph is taken from a link to an article that you included in your post:

To be honest, with my experience with IEM's that is limited to Truthear Crinacle X Zero, no way on earth would I want to make that IEM brighter, which is what that Knowles Target is doing - for me I'd say that Knowles Target is dead on arrival thanks!

I'd rather think you did not read the post attentively than missed the the link I think you followed to get the graph you subsequently attached in your post:

I have been itching to make some EQ with the Knowles target that can be found here as I am intrigued by what people think about this target.
Here is the perfect opportunity...

And this graph showing 2 times the Harman/Knowles comparison, 3 times the actual difference between the two curves as well as, for it's worth, the computed score of the Knowles target as seen from the Harman point of view.
Feel free to be offended, but having asked my fair share of unneeded question in my time (i.e. RTFM), in this case I personally wouldn't, but that is just me...
So here is reproduced the graph:

IMO the end game for EQ could be reduced to the following:

Individualized_Target = ( Normal_Hearing_Target x Hearing_Correction x Taste )Adapted_to_real_time_fit

- Provided that Taste and Hearing_Correction are completely orthogonal
- Hearing_Correction is probably not 1:1 with regards to hearing losses but should not be ignored completely.
If one thinks about it, completely ignoring losses because the brain adapts (which is true to a certain extend) would actually mean
a. the negation of the need for hearing aids, which is obviously not true
b. suppose that everyone's brain is hardwired to a certain curve and the brain "applies an EQ" to revert to this curve, taste might already disprove that
- One could argue about a per content EQ specifically aimed at compensating the shortcomings of the media but we'll have to wait for AI to recognize artistic choices from actual technical issues and be able to enact modifications...

Some audiometric/otoacoustic based audio personalization solutions that might make things much easier:
- https://www.nurasound.com/
- https://mimi.io/
- https://www.audiodo.com/
- https://jacoti.com/

 

[PeteL]

I'd rather think you did not read the post attentively than missed the the link I think you followed to get the graph you subsequently attached in your post:

And this graph showing 2 times the Harman/Knowles comparison, 3 times the actual difference between the two curves as well as, for it's worth, the computed score of the Knowles target as seen from the Harman point of view.
Feel free to be offended, but having asked my fair share of unneeded question in my time (i.e. RTFM), in this case I personally wouldn't, but that is just me...
So here is reproduced the graph:
View attachment 268390


IMO the end game for EQ could be reduced to the following:

Individualized_Target = ( Normal_Hearing_Target x Hearing_Correction x Taste )Adapted_to_real_time_fit

- Provided that Taste and Hearing_Correction are completely orthogonal
- Hearing_Correction is probably not 1:1 with regards to hearing losses but should not be ignored completely.
If one thinks about it, completely ignoring losses because the brain adapts (which is true to a certain extend) would actually mean
a. the negation of the need for hearing aids, which is obviously not true
b. suppose that everyone's brain is hardwired to a certain curve and the brain "applies an EQ" to revert to this curve, taste might already disprove that
- One could argue about a per content EQ specifically aimed at compensating the shortcomings of the media but we'll have to wait for AI to recognize artistic choices from actual technical issues and be able to enact modifications...

Some audiometric/otoacoustic based audio personalization solutions that might make things much easier:
- https://www.nurasound.com/
- https://mimi.io/
- https://www.audiodo.com/
- https://jacoti.com/

I am not fully sure if it is the hearing loss, but I did notice that many headphone that I found really lacking in the Highs to be judged somehow fine by many. Last example were Rode NTH-100, well they sounded just horrible to me, extremely muted and dull. They somehow got some enthusiastic reviews... Not the first time I noticed that I like more Highs than many. Looking at some other sites measurments now, might not be only the lack of Upper treble but they are also quite strong in the upper bass. There's that. And the sub bass roll off, there's that too. Not a good look objectively to me but still the fact that some friends thought it was quite neutral was beyond any understandable factor to me, they just sounded terrible... But even with all these flaws that was still the lack of highs that was killing it for me. I am 48 years old and I have been listening loud all my life for professional reasons.

 

[GaryH]

most measurements including Amir's do show the Trio with quite a lot more bass particularly sub-bass than the Zero

Depends what you get in the 64 Audio unit frequency-response lottery:

I think what it's going for with the 1-2kHz recess is for soundstage

Great, so we get a soundstage lottery too, what a thrill, not knowing how much of that sweet, sweet soundstage you're gonna get, so exciting!

The recess at ~5-7kHz I imagine is to eliminate sibilance

Or it's a phase cancellation issue due to poor design, which seems likely from its absolute mess of a step response that looks to show out of phase drivers (not to mention wrong polarity, that $2300 just keeps on giving!):

c.f. a $10 IEM:

I honestly don't notice a huge difference EQing this flatter

Which is just what you'd expect with a non-minimum phase cancellation that cannot be effectively EQed out.

The Zero a very good IEM, but I don't find Harman quite ideal as a target for IEMs.

Truthear Zero ≠ Harman target

I get there is a lot of adherence to Harman on ASR

You know why? Because this is a science forum, and the Harman target (both in- and over-ear) has been shown to be preferred by the majority in controlled, level-matched, double-blind listening studies by esteemed acoustic scientists. Your, or anyone else's, anecdotal opinion based on uncontrolled sighted listening subject to innumerable cognitive biases, groupthink, and even vested interests (e.g. for most 'reviewers' bankrolled by ads, sponsorships, 'collabs' and affiliated online stores) in contrast have zero valid scientific import. Unlike Head-fi, headphones.com, Reddit and Discord servers, provable claims backed by valid evidence actually matter here.

Even Harman themselves diverge from it quite a bit in the actual physical products they put out

Nope. What you conveniently failed to mention is the actual highest-scoring (most Harman compliant) in-ears in AutoEQ's ranking are the AKG (Harman-owned) N400 with a score of 89% (90% as measured by Sean Olive himself). Then there's the JBL (also Harman owned) Club Pro+ at 88%, and several other JBLs that score highly like the Endurance Run at 85% for just $20.

The Moondrop Blessing2 Dusk is another of my favourite IEMs and is also very close to Harman, according to AutoEQ's Harman list, the closest wired IEM in existence to Harman. But even it fixes basically all the issues I have with Harman, it has a smoother transition from the sub-bass, it doesn't do the dip, it is flat rather than upsloping to 1kHz, it is below Harman from 3-7kHz and it doesn't totally murder the treble (although this is the Dusk's weak point, there are IEMs with much better treble extension, such as the Variations and Monarch Mk2 - or indeed, the Trio).

Sounds like you're looking at Crinacle's measurements (using his fake knock-off coupler of course), which score a predicted preference rating of 81% in AutoEQ's ranking, not the 87% Oratory's measurement of the Dusk scores using his genuine GRAS coupler (the same Harman used for the data the preference rating formula is based off), which shows closer adherence to Harman (than the Truthear Zero too with its score of 81%).

As to whether it's somehow "objectively" worth it, of course it's not, it's not 50 times better. There are massive diminishing returns.

It's not a case of diminishing returns. It's a case of less return for massively more money. The Trio is an objectively mediocre IEM. For $2300. No amount of convoluted excuses is going to change that.

 

[IAtaman]

The Trio is an objectively mediocre IEM.

I think that is a very fair assessment.

You know why? Because this is a science forum, and the Harman target (both in- and over-ear) has been shown to be preferred by the majority in controlled, level-matched, double-blind listening studies by esteemed acoustic scientists. Your, or anyone else's, anecdotal opinion based on uncontrolled sighted listening subject to innumerable cognitive biases, groupthink, and even vested interests (e.g. for most 'reviewers' bankrolled by ads, sponsorships, 'collabs' and affiliated online stores) in contrast have zero valid scientific import. Unlike Head-fi, headphones.com, Reddit and Discord servers, provable claims backed by valid evidence actually matter here.

This part I find a bit confusing. Target is all scientific,it represents preference of people, amazing work and very very useful If you are a manufacturer. As a consumer it is important to understand as well to make sure we make the right choices, at least from tuning point of view. So far so good. But beyond some point, why do I care if a product adheres to a target that is supposed to represent general preference - I most certainly don't care about that in my choice of music, why should I in choice of equipment?

 

[Blorg]

@GaryH - I know Harman have that one TWS that comes quite close to the target. But they also have many that diverge and when they do diverge they very often diverge in exactly the ways I am saying they should. Harman and its associated companies have a lot more IEMs (and headphones) diverging more from Harman that that one N400NC TWS (which is often criticized for being a bit shouty). I own several of them personally.

My own favorite IEMs- The Dusk, Variations, Monarch Mk2, are all among the very closest IEMs to Harman. But they still diverge from Harman in these key ways I consider positive.

Oratory's Dusk measurement that you cite there as one of the most compliant Harman IEMs (and also just about my own favourite tuning) diverges from Harman in exactly the ways I am saying I want an IEM to diverge from Harman. So that's a great example. Because I'm not saying Harman in-ear is radically wrong and I want something totally different. My favourite tuning is something very close to Harman, but not quite. And the IEM you are holding up as a perfect example does actually diverge in these ways. It's slight, but Harman only needs very slight tweaks to be right for me.

I could take more high treble extension on the Dusk, that's the one thing it is weak on. The Monarch Mk2 and Variations are better on this. I think Knowles is on to something with this, that most people want substantially more treble in the 10-20kHz region than is in Harman (which, my understanding, is not based on any concrete data or science, but is just a "fill in the blank" up to 20khz). What I don't think they want is more treble lower down, and the high contrast between elevated upper mids/lower treble and then diving off a cliff is one of the issues with Harman in-ear.

 

[markanini]

The elephant in the room is that the target used by Amir wasn't gone through blind testing.

 

[GaryH]

This part I find a bit confusing. Target is all scientific,it represents preference of people, amazing work and very very useful If you are a manufacturer. As a consumer it is important to understand as well to make sure we make the right choices, at least from tuning point of view. So far so good. But beyond some point, why do I care if a product adheres to a target that is supposed to represent general preference - I most certainly don't care about that in my choice of music, why should I in choice of equipment?

Because sound reproduction preference was used as a proxy for perceived neutrality, and you (and anyone else) are most likely to fall into the majority-preference group. Comparing to music taste is a false equivalence. It would be like comparing TV display preferences with taste in movies.

 

[GaryH]

I agree with everything, but you have to understand that for most people, after having spent several thousands on an audio gear unit, the subjective bias is huge. Remember the Naim Atom thread? So they will never admit that it is just mediocre, even if it measures average or lousy, they will attribute some magical properties to it. Otherwise it woudn't be very flattering for the own rationality.

Oh I understand perfectly, pricing bias and post-purchase rationalization are rife in this hobby. But that's just why we need to keep calling out these companies and expose their products for the rip-offs they are using objective data, to prevent other unsuspecting potential customers from falling into the same fallacious higher price = better audio quality trap perpetuated by marketers, dealers and reviewers.

From which source are the impulse response measurements?

Reference Audio Analyzer

reference-audio-analyzer.pro

 

[markanini]

Oh I understand perfectly, pricing bias and post-purchase rationalization are rife in this hobby. But that's just why we need to keep calling out these companies and expose their products for the rip-offs they are using objective data, to prevent other unsuspecting potential customers from falling into the same fallacious higher price = better audio quality trap perpetuated by marketers, dealers and reviewers.

Reference Audio Analyzer

reference-audio-analyzer.pro

I disagree with the second part, that's is naive white knight thinking.

 

[IAtaman]

Because sound reproduction preference was used as a proxy for perceived neutrality, and you (and anyone else) are most likely to fall into the majority-preference group. Comparing to music taste is a false equivalence. It would be like comparing TV display preferences with taste in movies.

Right, so target's correlation to preference is due to "preference of flatness". Makes sense. But I am still not sure what you said is 100% accurate. The reason I say that is this CanJam keynote from Axell Grell in which he was showing in-ear FR graphs of various people and talking about of how different people perceive frequencies between 1k-10k differently, and that best you can do as a headphone designer is to develop a headphone that a lot of people will like. I actually tried to check my own HRTF using Earful to test my hearing threshold at various frequencies after that video but never made any sensible progress. Dealing with cheap IEMs recently, I do know however I do like the 200hz tuck of Harman curve, without it, things sound a bit muddy, but I most certainly find the fully compliant 1k-5k region too bright. What is your take on this if you don't mind me asking?

 

[GaryH]

Nothing Axel Grell said contradicts what I wrote.

Dealing with cheap IEMs recently, I do know however I do like the 200hz tuck of Harman curve, without it, things sound a bit muddy, but I most certainly find the fully compliant 1k-5k region too bright. What is your take on this if you don't mind me asking?

What IEMs with what EQ if used led you to that conclusion?

 

[IAtaman]

Nothing Axel Grell said contradicts what I wrote.

In that case I am misunderstanding you or him, or both. If I understand you correctly you are saying compliance to target is the ultimate measure of flatness and what he is saying is that flatness might mean different FR to different people. Where do I go wrong?

What IEMs with what EQ if used led you to that conclusion?

Hola, Hexa, TruthEar Zero x Crinicle, Dunu Titan S and Tangzu Wa'ner are the most recent ones.. I tried EQs from Oratory and Crinicle, both via AutoEQ in EQ APO and entered them manually from the PDF files. I tried some EQs also from ASR forums. All sounded better when I pulled back something around 3k a bit back after being EQ'ed to target. Those who has a peak at around 10K, that for sure bothers me as well but I think target also is lower there.

 

[Blorg]

@IAtaman I find Harman in-ear too intense in the gain region as well. A lot of people do. And even stuff like the Dusk, Variations, Monarch Mk2, which are among the most Harman compliant IEMs in existence, pull it back a bit in that region. I don't find Harman too intense there at all with over-ears, it's fine there.

With in-ears though it is initially impressive what it does for clarity, but with longer listening often tends to get a bit too much and become fatiguing. Not "totally unbearable", but just a bit too much, too shouty. This is a very common sentiment with Harman IE, much more than with the over-ear target which has far wider acceptance without modification. I wonder if this is part of the issue so many people have with Harman's research for the IE curve, because I like the curve at first listen too, it's a sort of Coke vs Pepsi Challenge situation, it sounds better initially but only becomes fatiguing with extended listening.

Two example IEMs I find a bit too shouty, that are highly compliant with Harman in this region, are the Tripowin Olina and Samsung Galaxy Buds+. Again, not intolerable, I still think both of these are good IEMs. But a bit more than my ideal. Tripowin revised this with the Olina SE which reduces that region and it's much better for it. A very popular mod with the OG Olina was to put a different filter on it (from the Tanchjim Tanya), this also reduced that region. There seems to have been this very widespread feeling from the community that it was too much, and I agree with that personally. Samsung also reduced the region in their Buds series after the Buds+, which I also found positive in the Buds Pro and Buds 2. Buds 2 Pro has similar above 3kHz but less 1-3kHz, I don't think I find this particularly problematic so far but I suspect I could take a little less.

Jaakko of AutoEQ has the Olina as his seventh most Harman compliant IEM in existence with 84, he has the Buds Plus way down with only 76 but I have no idea why as it is extremely close to Harman.