Truthear x Crinacle Zero:RED IEM Review

[MacClintock]

And as I am repeatedly trying to explain to you, as only 1/3 variables is limited to >40 Hz, this only gives a modest priority to frequencies above this, which is perfectly reasonable as this is where most musical content lies.

Nope. A quick Excel calculation shows the average absolute slope value is within an order of magnitude of both the corresponding std and mean values (~0.2 and ~0.3 times respectively). How many more false claims not backed by evidence are you going to throw around?

I did show you the calculation without the slope (but std and mean maintaned). What more evidence could you possibly want?

Are you serious? RAA?? The only one of those sources that will allow for an accurate EQ to Harman is Oratory's measurements using a genuine GRAS RA0045 coupler, as that's the exact same one Harman used (with the same acoustic impedance) when developing and testing their target. Just take a look at the mean error and standard deviation as calculated by AutoEQ's Jaakko Pasanen of some of those sources' measurements compared to Oratory's below.

Spoiler

View attachment 291997
View attachment 291998
View attachment 292004

Crinacle's fake RA0045 has the least error but that's saying nothing really considering how awful the 'competition' is in this department. It would be like a grown(?) man hanging a medal around his own neck for winning a kids' egg and spoon race. Notably his measurements on average underestimate both the bass and ~3 to 7 kHz region, which would result in attempts to EQ to the Harman target from them in fact overshooting it and incorrectly EQing to the target with the bass and ear gain in the mid-treble boosted, so it's no wonder those trying to EQ to Harman using these measurements might erroneously think it's a v-shaped and 'shouty' target; they're not EQing to the Harman target. There's evidence of this in the significantly lower predicted rating for the Dusk from Crinacle's compared to Oratory's measurements as seen in AutoEQ's ranking table, which can be explained by the former underrepresenting the bass and ear gain and so misrepresenting the response as further from Harman. And yes, I'm fully aware that there are other variables at play in these comparisons like unit variance and measurer incompetence, but Oratory being a trained professional (who actually started out in physics I might add) mitigates the latter factor when using his measurements, and the fact is none of the other measurement sources above have proven accuracy to the rig Harman used when devising their target; a genuine GRAS RA0045 does however, with less than 0.1 dB error (an inaudible difference when EQing):
View attachment 292014
So, which IEMs with what exact EQ have you used to equalize to Harman?

You don't seem to have proper reading skills. I did used the EQ of holy and sacred Sr. Davy. With many IEMs. Like Chu, Salnotes Zero, 7Hz Timesless, Blon BL-03, Etymotic ER-2SE and Ikko OH-10. Is this sufficient for you? Did I pass the test? All sound shouty to me.

Perfection is an unobtainable ideal. I mostly use my Sony MH755/750s or CIEM specifically measured by Oratory and equalized to Harman using his EQ, or occasionally the original Truthear Zero from my phone or with a modified Maiky EQ, but the latter are so uncomfortable with their stupidly huge bore size I don't spend much time with them.

You are aware that even after EQ most IEMs do not smoothly follow the Harman curve? For example the Sony Mh750/755 has too much sub-bass and resonances in the treble. Maybe you never listened to the pure Harman curve ever?

 

[MacClintock]

Why are there still users placing so much emphasis over Harman-IE-adherence when it's already established that the IE research is fundamentally flawed relative to the OE research?

Let's put aside the fact that rankings based on adherence doesn't make sense considering that Harman research was a statistical exercise of user preference based on high & low shelf adjustments, and deviating from what is objectively an averaged result is not an indication of poor performance. Let's put aside the fact the the IE research only used data from an internal panel of 10 employees from Harman. Let's even put aside the fact that what I and many others labelled as "Harman IE 2019" is simply smoothed over version of Harman IE 2017 made for Listen Inc. and does not have the same statistical backing as the OE research or even the 2017 study.

All I need to do is to cite is the title of research that went into developing IE 2017 2016:



Respondents in this study (if you can even call it that) did NOT have any high shelf or slope controls. Harman IE research only accounted for preferences in bass response (level and frequency of a 2nd order low shelf filter), which while admittedly better than nothing, is still a far far cry from the rigour that was established in OE research.

Anyone who claims to follow the scientific methodology should be able to acknowledge the limitations and shortcomings of the Harman IE research, and not put the same level of faith in it relative to the OE research.

Unfortunately we have here a Harman warrior who will not be convinced. There are religious people and then there are peole fighting for a supposedly scientific goal....

 

[MacClintock]

Yet you blindly believe a load of anecdotal uncontrolled sighted impressions of 'the IEM community' polluted by innumerable cognitive biases? Is this...satire?? I really can't tell anymore. I don't blindly believe anyone; my belief is based on whether scientific methodology and standards are followed, logical reasoning, and backing up claims with valid evidence, and unlike Sean and Oratory you're failing on all counts.

I assume you never worked as a scientist or published articles in peer-reviewed journals? It's funny, only non-scientist have this very romantic notion of how science works. But if you really participate, you will experience how much arbitrariness, jealeousness, envy, sloppyness, over-ambition, malice, greed, vanity, incompetence and so on is also involved. And I am talking about a hard science like theoretical physics. Applied acoustic research will have even more of all that.

 

[Blorg]

@MacClintock I find Oratory1990's Harman in-ear EQs to be too shouty and thin as well. So it's not just that Crinacle's IEC-711 measurements are a bit off, Oratory is both measuring and doing the EQ to Harman- and it's a bit shouty. I prefer his oratory1990/U-Sound IEM target. I do like his Harman over-ear EQs, they usually sound right from the get go.

 

[MacClintock]

@MacClintock I find Oratory1990's Harman in-ear EQs to be too shouty and thin as well. So it's not just that Crinacle's IEC-711 measurements are a bit off, Oratory is both measuring and doing the EQ to Harman- and it's a bit shouty. I prefer his oratory1990/U-Sound IEM target. I do like his Harman over-ear EQs, they usually sound right from the get go.

Same here, but not only Oratory's or crinacle's, but the HarmanIE2019v2 itself. But I like the Harman bass, which I usually take up to 1kHz and above oratory1990/U-Sound IEM target. Or the Red, which is amazingly close to "my target"

 

[MayaTlab]

Let's put aside the fact the the IE research only used data from an internal panel of 10 employees from Harman.

I think that subsequent studies did involve more participants. In particular A Statistical Model That Predicts Listeners’ Preference Ratings of In-Ear Headphones: Part 1 – Listening Test Results and Acoustic Measurements involved 71 individuals.

Respondents in this study (if you can even call it that) did NOT have any high shelf or slope controls. Harman IE research only accounted for preferences in bass response (level and frequency of a 2nd order low shelf filter), which while admittedly better than nothing, is still a far far cry from the rigour that was established in OE research.

In The Influence of Program Material on Sound Quality Ratings of In-Ear Headphones, it is mentioned that unpublished studies did involve both treble and bass adjustments :

I also think that the in-ear research was nowhere near as thorough as the over-ear one, but I don't think that this is because of the listening panel size or that because some studies only adjusted the bass or treble response.

 

[crinacle]

I think that subsequent studies did involve more participants. In particular A Statistical Model That Predicts Listeners’ Preference Ratings of In-Ear Headphones: Part 1 – Listening Test Results and Acoustic Measurements involved 71 individuals.



In The Influence of Program Material on Sound Quality Ratings of In-Ear Headphones, it is mentioned that unpublished studies did involve both treble and bass adjustments :

View attachment 292179
View attachment 292180

I also think that the in-ear research was nowhere near as thorough as the over-ear one, but I don't think that this is because of the listening panel size or that because some studies only adjusted the bass or treble response.

I think the 2016 paper was the only published one that I could find where respondent input both created and validated the eventual "target" much like the OE research. As you've mentioned, the parameters of the 2017 target is unpublished and the respondents themselves did not play a part in the creation of the target itself, only "validated" against 30 other in-ear headphones via a virtual headphone test method.

I mean, it's great that IE 2017 is validated in some way, but my bigger issue is with how it was created in the first place (hopefully with some input from the 2016 research) which is kind of a black box at the moment.

 

[markanini]

The way Red diverges from Harman IE adherence, with a focus on generalist performance, has been a long time coming. Whether it represents a takedown of Harman IE is debatable. I could care less. OTOH, If Red will influence future IEMs I'm all here for it, it's a great sounding IEM.

 

[kairo.de]

To me, Crinacle's tuning sounds a lot more natural than the Harman approach. For me, it's not even close. I've had a few IEMs. In all price ranges. Couldn't listen to any of them for a long time. None have clicked with me like the Zero Red. Of course I am only one. I'm really curious what the others will report. @crinacle Congratulations on the Red. Fantastic tuning.

 

[F1308]

Unfortunately we have here a Harman warrior who will not be convinced. There are religious people and then there are peole fighting for a supposedly scientific goal....

I think perhaps the trick behind the Harman curve is exactly that it is here, ready, available.

Because, what else is there ?

As soon as you go wondering if what you are hearing is right or wrong, real or unreal, you land on what is available...where else ?

 

[Lbstyling]

Can anyone suggest a good strong cable with mic that fits this IEM?

 

[Matias]

Can anyone suggest a good strong cable with mic that fits this IEM?

I am using Moondrop MKI.

I would have ordered my Red with mic cable for a couple of USD more if there was the option... @crinacle can you do something about it? Not that I am going to buy the IEM again but a spare 3.5 mm mic cable from Truthear would be nice.

 

[MacClintock]

I think perhaps the trick behind the Harman curve is exactly that it is here, ready, available.

Because, what else is there ?

As soon as you go wondering if what you are hearing is right or wrong, real or unreal, you land on what is available...where else ?

Yes, it is a good basis (for over-ear headphones already the solution), for IEMs a starting point, that has to be taken with a grain of salt and not a rigorous compulsion that has to be followed by all means.

 

[markanini]

https://twitter.com/i/web/status/1661253505833902080

Slight OT, Tanchjim released an app for their DSP powered IEMs. Within there's an option to apply correction to various targets, including their internal target:

 

[julian_hughes]

Why are there still users placing so much emphasis over Harman-IE-adherence when it's already established that the IE research is fundamentally flawed relative to the OE research?

Established how?

You have a huge database of measurements. Very impressive and indeed useful for many people. But it seems they were performed with tools of unknown provenance so cannot reliably be compared with the data of people who used tools with known, reliable parameters. You offer your own preference curve - that's fine, I guess we can all make one we like. But why would that be better than a curve determined in blind tests with trained listeners and with data gathered by tools which can be relied upon? And then, after years and years of offering your measurements as bona fide you buy new tools and announce that the old data was wrong and we should all now trust your new data, which is *substantially* different from your old data!

 

[julian_hughes]

Can anyone suggest a good strong cable with mic that fits this IEM?

https://www.aliexpress.com/i/33005210890.html These are OK. Choose the option on the far right (A3 0.78mm With Mic)

 

[crinacle]

Established how?

You have a huge database of measurements. Very impressive and indeed useful for many people. But it seems they were performed with tools of unknown provenance so cannot reliably be compared with the data of people who used tools with known, reliable parameters. You offer your own preference curve - that's fine, I guess we can all make one we like. But why would that be better than a curve determined in blind tests with trained listeners and with data gathered by tools which can be relied upon? And then, after years and years of offering your measurements as bona fide you buy new tools and announce that the old data was wrong and we should all now trust your new data, which is *substantially* different from your old data!

Funny how you'd quote me and ask "Established how?", and then completely ignore (and delete!) the rest of my post where I proceed to establish how.

Let's put aside the fact that rankings based on adherence doesn't make sense considering that Harman research was a statistical exercise of user preference based on high & low shelf adjustments, and deviating from what is objectively an averaged result is not an indication of poor performance. Let's put aside the fact the the IE research only used data from an internal panel of 10 employees from Harman. Let's even put aside the fact that what I and many others labelled as "Harman IE 2019" is simply smoothed over version of Harman IE 2017 made for Listen Inc. and does not have the same statistical backing as the OE research or even the 2017 study.

All I need to do is to cite is the title of research that went into developing IE 2017 2016:

Olive, S., Welti, T., & Khonsaripour, O. (2016, August). The preferred low frequency response of in-ear headphones. In Audio Engineering Society Conference: 2016 AES International Conference on Headphone Technology. Audio Engineering Society.

Respondents in this study (if you can even call it that) did NOT have any high shelf or slope controls. Harman IE research only accounted for preferences in bass response (level and frequency of a 2nd order low shelf filter), which while admittedly better than nothing, is still a far far cry from the rigour that was established in OE research.

Anyone who claims to follow the scientific methodology should be able to acknowledge the limitations and shortcomings of the Harman IE research, and not put the same level of faith in it relative to the OE research.

But fine, I will play along.

I am criticising the Harman IE research and how it is not as rigorous as the OE research. If one holds so much reverence and faith in the OE research, then it only makes sense to question the IE research which was conducted with different parameters and validation methodology, all of which I had pointed out in my posts.

The "old data" is generated under the IEC 60318-4 standard, now being challenged by the "new data" under the ITU-T P.57 Type 4.3 standard. With more data and research done on the new standard, time will tell whether Type 4.3 offers more accurate data, but based on current testing I'm inclined to agree especially for low frequencies. However if my usage of Type 4.3 OFFENDS anyone, I still continue to generate measurements with my IEC 60318-4 system in parallel.

RE: my curve, I had offered it as an alternative to Harman IE but the concept behind it is still based on the parameters that created the initial OE curve. Users there were allowed to adjust for both low and high shelf filters, and thus my own curve (using the Zero:RED as an example) would represent the subset of listeners that prefer less of a bass shelf, and more of a slope.

Does this mean the subset of listeners who would prefer "my curve" would be more substantial than the subset of listeners that prefer the current IE curve? Currently we cannot say for sure as there have not been any validation tests, but it would be fairly unproblematic to set up a blind test with a regular Harman-2019-tuned Zero and a Zero:RED, alongside additional EQ to have each of them fit their respective targets more precisely. That is in the works now (all I can say at this juncture).

 

[lazarian]

Funny how you'd quote me and ask "Established how?", and then completely ignore (and delete!) the rest of my post where I proceed to establish how.

But fine, I will play along.

I am criticising the Harman IE research and how it is not as rigorous as the OE research. If one holds so much reverence and faith in the OE research, then it only makes sense to question the IE research which was conducted with different parameters and validation methodology, all of which I had pointed out in my posts.

The "old data" is generated under the IEC 60318-4 standard, now being challenged by the "new data" under the ITU-T P.57 Type 4.3 standard. With more data and research done on the new standard, time will tell whether Type 4.3 offers more accurate data, but based on current testing I'm inclined to agree especially for low frequencies. However if my usage of Type 4.3 OFFENDS anyone, I still continue to generate measurements with my IEC 60318-4 system in parallel.

RE: my curve, I had offered it as an alternative to Harman IE but the concept behind it is still based on the parameters that created the initial OE curve. Users there were allowed to adjust for both low and high shelf filters, and thus my own curve (using the Zero:RED as an example) would represent the subset of listeners that prefer less of a bass shelf, and more of a slope.

Does this mean the subset of listeners who would prefer "my curve" would be more substantial than the subset of listeners that prefer the current IE curve? Currently we cannot say for sure as there have not been any validation tests, but it would be fairly unproblematic to set up a blind test with a regular Harman-2019-tuned Zero and a Zero:RED, alongside additional EQ to have each of them fit their respective targets more precisely. That is in the works now (all I can say at this juncture).

Exciting, thanks for your work. Definitely plenty of hardliners around that will do anything in their power to try and bring down any and all discussion that questions existing research or aims to further or enhance it. Very counter intuitive to the actual scientific process.
End of the day we all get awesome products at very affordable prices, it's totally win win.

 

[julian_hughes]

Funny how you'd quote me and ask "Established how?", and then completely ignore (and delete!) the rest of my post where I proceed to establish how.



But fine, I will play along.

I am criticising the Harman IE research and how it is not as rigorous as the OE research. If one holds so much reverence and faith in the OE research, then it only makes sense to question the IE research which was conducted with different parameters and validation methodology, all of which I had pointed out in my posts.

The "old data" is generated under the IEC 60318-4 standard, now being challenged by the "new data" under the ITU-T P.57 Type 4.3 standard. With more data and research done on the new standard, time will tell whether Type 4.3 offers more accurate data, but based on current testing I'm inclined to agree especially for low frequencies. However if my usage of Type 4.3 OFFENDS anyone, I still continue to generate measurements with my IEC 60318-4 system in parallel.

RE: my curve, I had offered it as an alternative to Harman IE but the concept behind it is still based on the parameters that created the initial OE curve. Users there were allowed to adjust for both low and high shelf filters, and thus my own curve (using the Zero:RED as an example) would represent the subset of listeners that prefer less of a bass shelf, and more of a slope.

Does this mean the subset of listeners who would prefer "my curve" would be more substantial than the subset of listeners that prefer the current IE curve? Currently we cannot say for sure as there have not been any validation tests, but it would be fairly unproblematic to set up a blind test with a regular Harman-2019-tuned Zero and a Zero:RED, alongside additional EQ to have each of them fit their respective targets more precisely. That is in the works now (all I can say at this juncture).

I quoted a specific assertion in order to ask a specific question about that specific assertion. If I'd thought the rest of your post had explained it then I wouldn't have asked the question. Sorry for asking a question My Lord! Are you now A God Who May Not Be Questioned?

I asked a very, very simple question and you respond with a snotty remark and then a series of evasions. Harman did things the right way to make their curves, and put it out there how they did it and why they stood by it. They never claimed to beyond criticism. You treat a simple question like it was some kind of heresy and then rely on conjecture and supposition to avoid offering a real answer.

You have a *very* obvious commercial interest in making your curve and your dataset some kind of de facto standard. Enough of this "we cannot say for sure" and "it would be..." and "is in the works" and "all I can say at this juncture". What is next? "The Check is in the post"?

Also what the f*c* are "technicalities"? Just a long word for "because I say so"?

An edit: it's fine to criticise established standards and conventions and methods, that's part of progress and improvement. But when you set yourself up as an expert in opposition to those things you seek to undermine but have nothing more credible to offer than "in the works" and also adopt the big "I AM" attitude if anyone asks even a not very hard question......that sucks.

 

[lazarian]

I quoted a specific assertion in order to ask a specific question about that specific assertion. If I'd thought the rest of your post had explained it then I wouldn't have asked the question. Sorry for asking a question My Lord! Are you now A God Who May Not Be Questioned?

I asked a very, very simple question and you respond with a snotty remark and then a series of evasions. Harman did things the right way to make their curves, and put it out there how they did it and why they stood by it. They never claimed to beyond criticism. You treat a simple question like it was some kind of heresy and then rely on conjecture and supposition to avoid offering a real answer.

You have a *very* obvious commercial interest in making your curve and your dataset some kind of de facto standard. Enough of this "we cannot say for sure" and "it would be..." and "is in the works" and "all I can say at this juncture". What is next? "The Check is in the post"?

Also what the f*c* are "technicalities"? Just a long word for "because I say so"?

An edit: it's fine to criticise established standards and conventions and methods, that's part of progress and improvement. But when you set yourself up as an expert in opposition to those things you seek to undermine but have nothing more credible to offer than "in the works" and also adopt the big "I AM" attitude if anyone asks even a not very hard question......that sucks.

Can you point to where Crinacle is being overly pushy saying his curve is the one and only, and not just creating a product to see market reaction, and to use in further testing he's organising?

Sorry wording probably a bit easy to misunderstand, I am curious if there's been previous statements made that got you to the point you made here. Without hunting through every Crinacle response, recent responses point to him not liking the IE research, finding it limited compared to OE, and wanting to see if his own preference curve is more liked. How he goes about that is obviously cause for future debate as to whether it would partially invalidate or add to the existing research.