• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Poll: Do you like Harman In-Ear target curve?

Do you like the Harman In-Ear target curve?

  • Yes, it is perfect.

    Votes: 50 45.9%
  • No, I like a little less mids.

    Votes: 30 27.5%
  • No, I like a lot less mids.

    Votes: 13 11.9%
  • No, I like more mids!

    Votes: 16 14.7%

  • Total voters
    109
OP
M

markanini

Major Contributor
Joined
Feb 15, 2019
Messages
1,719
Likes
1,768
Location
Scania
Based on your clarification, I would sort of agree, semantics aside. Rather than trying to use adjectives which are commonly used to describe the nature of sound colorations, I often find myself describing the sound imparted by describing a quality one can imitate with one's voice. In the case of a peak somewhere around 2 or 3kHz, I would suggest it sounds like the sound of the letter "a" as in the way we North Americans pronounce the word "ant." Human voices will take on a nasal "EHNNN" sound emphasis within the sound of similar vowels being spoken by the reproduced voice when there's a peak there. And that's what I hear too much of in IEMs which closely adhere to the Harman in-ear target curve or which may have an even stronger peak there. Everybody sounds nasal.
That sounds about right. I checked with ChatGPT and it considers the shouty range to be between 1-5Khz or 2-5Khz depending on how the question is phrased. Audio university considers the vowels A as in face, and EE an in bee :
Still I can't be sure it's my place to declare a specific range so I have to accept some ambiguity to avoid editorializing / inserting my own bias to the discussion. The same goes for discussing other ranges such as bass, this has previously been accepted as subject to individual preference to large degree, with an existing poll on ASR showing comparable numbers to Harmans own. I'm sticking to the the midrange here because this is what has been discussed the most as of late giving incentive for the poll. Threads with a more open discussion about Harman targets exist already.
 
Last edited:

Peterinvan

Active Member
Joined
Dec 10, 2021
Messages
294
Likes
234
Location
Canada
IMO there is an assumption in the Harman Curve that it is a target for an "average person" with "average hearing.

I would guess that the folks participating in this forum are in an older demographic - and have reached a stage in life where they can devote big chunks of disposable wealth to this hobby.

ASR gives us machine generated frequency charts that are useful. But for subjective reports I would suggest that we need to see a reviewer's Audiogram before we can relate to what they are hearing. I am in my 70's now. Here is my latest chart:

1671205933721.png
 

DavidMcRoy

Addicted to Fun and Learning
Joined
Jul 10, 2018
Messages
574
Likes
990
That sounds about right. I checked with ChatGPT and it considers the shouty range to be between 1-5Khz or 2-5Khz depending on how the question is phrased. Audio university considers the vowels A as in face, and EE an in bee :
Still I can't be sure it's my place to declare a specific range so I have to accept some ambiguity to avoid editorializing / inserting my own bias to the discussion. The same goes for discussing other ranges such as bass, this has previously been accepted as subject to individual preference to large degree, with an existing poll on ASR showing comparable numbers to Harmans own. I'm sticking to the the midrange here because this is what has been discussed the most as of late giving incentive for the poll. Threads with a more open discussion about Harman targets exist already.
That's a very interesting chart. It's fascinating how different people will ascribe different adjectives to describe those sounds.

For example, I would never have described the 4kHz and 8kHz regions that way. I'd say a 4kHz emphasis enhances "snap" from a percussive sound, and imparts "whininess" to voices.

I always associate an 8kHz peak with imparting "liquidity/wetness" to high frequency sounds.

"Air" and "a lot of distortion components" both live above about 10kHz. Working in TV broadcasting, I would usually treat distorted audio sources by peaking 8kHz and rolling off above that to subjectively "clean up" the distortion to make the offending sources sound subjectively cleaner.
 

GaryH

Major Contributor
Joined
May 12, 2021
Messages
1,348
Likes
1,804
to avoid editorializing / inserting my own bias to the discussion
That ship has long sailed. The very first post of the thread:
Discussion about the Harman in-ear target has come about in amirm's review thread for of the TRUTHEAR x Crinacle Zero IEM. Some people find that IEM, and other Harman tuned IEMs sound fine. Others coined the term "shouty-gate".
shouldn't be there at all if you're claiming 'to avoid editorializing / inserting your own bias into the discussion', but if it is, it should be made 100% clear that the Truthear Zero is not in fact audiblly that highly matched to the Harman target, rather than conflating the two as you (and many others) do here, misleading those who've heard the Truthear into thinking they're listening to the Harman target, and increasing the likelihood these people will erroneously choose 'less mids' in this poll due to the Zero's broadband 2 dB excess above Harman in the upper midrange. Talking of poll options, there are 2 for less mids (wrongly suggestive that this is the most common opinion), but only one each for more mids and liking the target. And the latter is more restrictive than all the other options; it should have been 'yes, it's about right', or even just 'yes', rather than using the absolute superlative 'perfect' which is bound to decrease the number picking it. And yet, despite your best efforts to bias-prime the respondents and skew the results in favor of the narrative you're pushing that the Harman in-ear target is 'shouty', the outright majority still voted that they like it, in accord with the results from Harman's scientific research in the form of controlled blind listening studies, which are what actually matter, rather than pre-biased, ill-conceived forum polls.
 
Last edited:

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
That's a very interesting chart. It's fascinating how different people will ascribe different adjectives to describe those sounds.

For example, I would never have described the 4kHz and 8kHz regions that way. I'd say a 4kHz emphasis enhances "snap" from a percussive sound, and imparts "whininess" to voices.

I always associate an 8kHz peak with imparting "liquidity/wetness" to high frequency sounds.

"Air" and "a lot of distortion components" both live above about 10kHz. Working in TV broadcasting, I would usually treat distorted audio sources by peaking 8kHz and rolling off above that to subjectively "clean up" the distortion to make the offending sources sound subjectively cleaner.
Yeah to me it seems the 2-8 kHz adjectives need to be shifted down one, but these are likely geared towards musicians and mixing engineers given its source.
 
OP
M

markanini

Major Contributor
Joined
Feb 15, 2019
Messages
1,719
Likes
1,768
Location
Scania
That ship has long sailed. The very first post of the thread:

shouldn't be there at all if you're claiming 'to avoid editorializing / inserting your own bias into the discussion', but if it is, it should be made 100% clear that the Truthear Zero is not in fact audiblly that highly matched to the Harman target, rather than conflating the two as you (and many others) do here, misleading those who've heard the Truthear into thinking they're listening to the Harman target, and increasing the likelihood these people will erroneously choose 'less mids' in this poll due to the Zero's broadband 2 dB excess above Harman in the upper midrange. Talking of poll options, there are 2 for less mids (wrongly suggestive that this is the most common opinion), but only one each for more mids and liking the target. And the latter is more restrictive than all the other options; it should have been 'yes, it's about right', or even just 'yes', rather than using the absolute superlative 'perfect' which is bound to decrease the number picking it. And yet, despite your best efforts to bias-prime the respondents and skew the results in favor of the narrative you're pushing that the Harman in-ear target is 'shouty', the outright majority still voted that they like it, in accord with the results from Harman's scientific research in the form of controlled blind listening studies, which are what actually matter, rather than pre-biased, ill-conceived forum polls.
Wrong response for many reasons. Unlike you I did not focus on a measurement of a single sample. Seeking out more data shows Truthear Zero has high conformity to Harman-IE across multiple samples.
I chose the answers based on a sampling of user responses about Harman-IE tuned IEMs. The goal was to provide sufficient representation for each group I identified. 43 votes in it looks like I did OK with no answer getting too little or too much votes. Your claim that I sought out a certain outcome is unfounded and ridiculous.
BTW If you look at my post history you will find I've challenged unfounded critique about the Harman OE target many times and provided evidence for its universality, to those that refused to accept the data, by showing measurements of a multitude of well regarded headphones from different decades that happen to comply with the target with no input from Harman.
 

Attachments

  • Screenshot 2022-12-18 111152.png
    Screenshot 2022-12-18 111152.png
    23.2 KB · Views: 58
  • Screenshot 2022-12-18 111141.png
    Screenshot 2022-12-18 111141.png
    24 KB · Views: 61
  • Screenshot 2022-12-18 111128.png
    Screenshot 2022-12-18 111128.png
    24.2 KB · Views: 62
  • Screenshot 2022-12-18 111059.png
    Screenshot 2022-12-18 111059.png
    22.6 KB · Views: 64
Last edited:

GaryH

Major Contributor
Joined
May 12, 2021
Messages
1,348
Likes
1,804
Unlike you I did not focus on a measurement of a single sample. Seeking out more data shows Truthear Zero has high conformity to Harman-IE across multiple samples.
Unlike you I realize quality over quantity of data matters. (Also convenient how you left out the measurements that don't fit your narrative.) Those measurements use cheap, fake knock-offs of GRAS acoustic couplers from places like AliExpress. The measurements on ASR are using genuine GRAS couplers. As do Dr Sean Olive's measurements which show a similar excess in the upper midrange / lower treble, and that was using foam tips that are likely lowering response there by up to ~2 dB, so with the silicone tips (as most will use) the excess would probably be greater and match ASR's measurements even more closely. The fact is by the objective metric of our best headphone science (that for IEMs has a 91% correlation with actual preference ratings given in blind tests), when measured on genuine industry standard GRAS couplers, the Truthear Zero scores a predicted preference rating of 81% (which is a measure of its conformity to the Harman in-ear target) calculated from its frequency response. This is significantly below what Sean Olive himself defines as 'excellent' conformity: a rating of 90-100%.
I chose the answers based on a sampling of user responses about Harman-IE tuned IEMs.
Nah you just copied the options word for word (just exchanging 'mids' for 'bass') from someone else's equally biased and ill-conceived poll on here, probably hoping you'd get similar percentages to provide bogus 'evidence' for your 'shouty Harman IE!' narrative, but contrary to this it hasn't worked out for you with a large majority voting that they do in fact like the Harman in-ear target, in agreement with what the blind listening studies of their scientific research found.
 
Last edited:

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
The fact is by the objective metric of our best headphone science (that for IEMs has a 91% correlation with actual preference ratings given in blind tests), when measured on genuine industry standard GRAS couplers, the Truthear Zero scores a predicted preference rating of 81% (which is a measure of its conformity to the Harman in-ear target) calculated from its frequency response. This is significantly below what Sean Olive himself defines as 'excellent' conformity: a rating of 90-100%.
Using the preference rating alone without real regard to the nature of the deviations that cause it is a slippery slope to step out on. I have a set of these IEMs, and they have nearly the same tonality as my floor system and other headphones on the Harman target. The only real difference is they don't have the same bass extension, which could be the reason for the slight peaking seen before the roll-off in bass. Psychoacoustically this is supported in that adding a dB or two of boost before a response roll-off in the bass region can give the impression of additional bass extension. Generally, though, I would say that they comply with the target subjectively.

Also variations due to manufacturing is a lot different than those due to the inherent design. This doesn't mean that the Truthear Zero is non-conformant to the target, but that there will be a spread in responses. This means there will be a gaussian distribution that has most of them conforming reasonably well, with outliers that deviate in various ways. To say it doesn't conform means we need to see a clear trend in non-conformance in one or more areas that is seen in all samples. As far as single-sample deviations, there is not only variations in manufacturing to consider, but in the nature of the test itself in that there will be variations in how the IEM is inserted into the pinnae. The soft tips mean it wont stop at the same spot each time, and thus will rely on the person taking the measurement to use their own subjective judgement when inserting them. In contrast, a listener has feedback from the tip pressure and sound to determine when the IEM is at the correct position in the ear canal. Naturally I would not expect each sample to literally fall right on top of the target, but to track it within a reasonable tolerance band, something like ±2 dB. If the samples tested fall within that band (or whatever we deem acceptable), then its reasonable to say they are conforming to the target. If there are deviations, then one needs to actually make a meaningful assessment of their true impact since much of the response is relative and we have to choose where to put the measurement relative to the target.
 

someguyontheinternet

Active Member
Joined
Apr 16, 2021
Messages
194
Likes
334
Location
Germany
I'm not sure why this is still being misunderstood, but I've seen this train of thought in different threads popping up again and again (including here):
"The Harman target is used as a target therefore it must be perfect for all/most individuals. Any data suggesting othewrwise means the Harman target is worthless."

The Harman target is not the perfect frequency response for every individual. It is simply the best baseline frequency response with the lowest distance to individual preferences. EQing for individual preference by starting from the Harman target will be relatively easy because of this.

The votes also support this as we can see a big group in favor of the target, two groups of slightly smaller combined size in favor of minor deviation and one small group in favor of major deviation. Seems like a pretty "as expected" result to me which doesn't contradict the purpose and utility of the Harman target.
 

GaryH

Major Contributor
Joined
May 12, 2021
Messages
1,348
Likes
1,804
Using the preference rating alone
I'm not. Please read this post for a full excplication of the Truthear Zero's most significant deviation from the Harman in-ear target and why it's audible.
Also variations due to manufacturing is a lot different than those due to the inherent design. This doesn't mean that the Truthear Zero is non-conformant to the target, but that there will be a spread in responses.
This spread is small for the Truthear Zero. We know this from the low relative variance between the 9 units Crinacle measured.
As far as single-sample deviations
It's not a single sample. As I said the other measurement that uses a valid genuine industry standard GRAS coupler is Sean Olive's, which shows a similar deviation in the upper mids / lower treble as ASR's sample.
there is not only variations in manufacturing to consider, but in the nature of the test itself in that there will be variations in how the IEM is inserted into the pinnae.
You can tell the insertion depth from the spectral location of the coupler resonance. It's the same in both Sean and ASR's measurements.
If there are deviations, then one needs to actually make a meaningful assessment of their true impact since much of the response is relative and we have to choose where to put the measurement relative to the target.
The IEC industry standard (as Harman do) is to normalize at 500 Hz, for very good reasons.
 
Last edited:
OP
M

markanini

Major Contributor
Joined
Feb 15, 2019
Messages
1,719
Likes
1,768
Location
Scania
hose measurements use cheap, fake knock-offs of GRAS acoustic couplers from places like AliExpress
Speculative without proof of compromised accuracy. With calibration accuracy is comparable to GRAS coupler for IEM measurements. https://www.audiosciencereview.com/...ec-60318-4-clone-coupler-is-it-worth-it.23865 Like Cars-N-Cans touches on it needs to be understood that these systems are working within tolerances so obsessing about results to the dB will likely be a wasted effort. You may be conflating the requirements for OE measurements aren't comparable outside of GRAS units.

Nah you just copied the options word for word (just exchanging 'mids' for 'bass') from someone else's equally biased and ill-conceived poll on here, probably hoping you'd get similar percentages to provide bogus 'evidence' for your 'shouty Harman IE!' narrative, but contrary to this it hasn't worked out for you with a large majority voting that they do in fact like the Harman in-ear target, in agreement with what the blind listening studies of their scientific research found.
Are you aware of your line of questioning resembling a bad lawyer? You can't prove intentions yet claim you are certain...

I've already provided Oratory1990's target and the user opinions, the later of which I was open about being not conclusive, merely warranting discussion. You must have some emotional investment to a certain outcome if you can't accept this. The poll you linked to is different because Harman acknowledged the segments that prefer different bass levels. We don't currently have a definitive conclusion on why some people perceive Harman IE mids elevated, forget segmentation models for now.
 
Last edited:

GaryH

Major Contributor
Joined
May 12, 2021
Messages
1,348
Likes
1,804
Speculative without proof of compromised accuracy.
Listen to the experience of professionals in the field:
https://www.reddit.com/r/oratory1990/comments/ukw7sp/_/i8k703r
With calibration accuracy is comparable to GRAS coupler for IEM measurements. https://www.audiosciencereview.com/...ec-60318-4-clone-coupler-is-it-worth-it.23865
Still not actually reading your own links I see:
The calibration files were poorly done, and therefore useless. So we need a way to calibrate the clone couplers with better coupler like the GRAS or B&K couplers. This is quite a challenge if we don't have access to exactly the same IEM measured by GRAS or B&K couplers and compare the measurement with their measurement result.
If you're presenting measurements that use cheap knock-off equipment as valid, you need to provide clear evidence that every one of them has been correctly compensated and so is commensurable with proper measurements on genuine equipment. And crucially, you'd have to do this for each IEM (in this case the Truthear), because differences in acoustic impedance of the fake couplers from the genuine ones will result in different compensations for each IEM (as the IEMs themselves have differing acoustic impedance which will interact with the differing acoustic impedance of the fake couplers in different ways). Read this if you don't understand what I'm talking about. This effect can actually be seen practically in the post you linked but failed to examine properly, with one IEM's measurements with the clone differing by 1 dB from the genuine coupler over a pretty broadband range in the lower to mid-treble. Who knows what the error would be for the Truthear. And that was with one clone that randomly happened to produce the 'best' results, another one may be far worse (potentially up to 5 dB off as Oratory says above). It's a complete lottery, and this is why these knock-off coupler measurements cannot be trusted.

I've already provided Oratory1990's target and the user opinions, the later of which I was open about being not conclusive, merely warranting discussion.
Nope, your absolutist position without valid evidence is clear from previous comments, please don't suddenly now pretend otherwise:
First the In-ear target needs to be improved. Find the cause for the mid-elevated result in it's current state
a wide range of listeners like it better than Harman

The poll you linked to is different
That was to expose the fact that you've just copied the options verbatim from that poll to yours, despite claiming:
I chose the answers based on a sampling of user responses about Harman-IE tuned IEMs.
It's this kind of intellectual dishonesty that calls into question many of your unfounded claims.
 
Last edited:

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
I'm not. Please read this post for a full excplication of the Truthear Zero's most significant deviation from the Harman in-ear target and why it's audible.

This spread is small for the Truthear Zero. We know this from the low relative variance between the 9 units Crinacle measured.

It's not a single sample. As I said the other measurement that uses a valid genuine industry standard GRAS coupler is Sean Olive's, which shows a similar deviation in the upper mids / lower treble as ASR's sample.

You can tell the insertion depth from the spectral location of the coupler resonance. It's the same in both Sean and ASR's measurements.

The IEC industry standard (as Harman do) is to normalize at 500 Hz, for very good reasons.
By your metrics, then, something like the Stealth is also non-compliant since it has excesses and shortfalls relative to the target that are easily visible. Amir’s measurements of the Truthear show a slight excess in the treble region with the usual bump in the bass like several other headphones and IEMs exhibit. Even our most “compliant” headphones here at ASR (Stealth and Expanse) don’t conform since he took artistic license with the bass response and deviated from the target. Edit: I would also argue that the slight treble lift in the Truthear could be to help give good subjective balance with respect to the bass response. We dont have any real info on what the overall design process was, or what decisions were made for what reasons.

You make it sound like it’s grossly non compliant with the target, but it’s actually within reason. I would not call a slight 1-2 dB treble lift a substantial deviation, esp. given all the uncertainties in the measurements. Is it audible? Slightly, but not so much that I would call it “shouty.” On top of this, the Harman target is an aggregate of many users. There will be physical and perceptual variations from individual to individual that easily exceed that. It’s a crusade to split hairs. If we insist on this level of granuarity, then we need to define an actual tolerance range for the Harman curve, and use that as the metric. This is doubly-so for things like IEMs that cost about as much as a meal at a fast-food restaurant. For the cost, this IEM is actually quite reasonable in its conformance to the target. It is, after all, a consumer product and not a calibrated transducer.

And speaking more generally I certainly wouldn’t use an IEM alone as a reference for one’s preference of the Harman Target as has been pointed out since even if it had full compliance in the regions in question the IEM will not have the perceptual bass extension of a large headphone that complies and goes down below 20 Hz. It will just blow it out of the water. Anyone voting on just one sample will likely be in error.
 
Last edited:

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
And for reference, here is the Stealth, which as hailed as SOTA at the time. The sample that was measured and reviews has some minor warts. There are slight deviations up to about 1 kHz and then a minor shortfall at 4 kHz. Complies? If that shortfall were across all samples would we say they are recessed? Now the Truthear does land above the curve over the treble region by a dB or two depending on the measurements, maybe more for some samples, but if that really matters, then we need to state how much deviation is acceptable for compliance.


1671489965860.png
 

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
And to go even farther, we have standardized on one fixed frequency as the reference. Using your Moondrop example, can we be sure that will be a peak? What if it’s perceived as being a shortfall thereafter by some people since there is a broad lift in the upper bass/lower midrange region? I would argue no since the 1-3 kHz region is where the ear quite sensitive and will likely be the perceptual region that helps to define the overall tonality there. And it’s not so narrow that it won’t be audible. The ear is integrating, after all. It will be responding to the average SPL over that region. Very narrow high Q disturbances are inaudible because they span such a short range of frequencies. There is little overall energy there unless it’s extreme, and the broad response means it gets lost and is inaudible. Here I would say that the peak and dip will be audible as tonality error of some degree. It’s also been smoothed, so smaller variations are not visible.
1671490811790.png
 
OP
M

markanini

Major Contributor
Joined
Feb 15, 2019
Messages
1,719
Likes
1,768
Location
Scania
Listen to the experience of professionals in the field:
Are you electing oratory1990 to the niche of top critic of budget measurement equipment? Individuals wishing to achieve comparable results on a low budget couplers go for individual calibration. Technically more ambitious than using provided calibrations oratory1990 talks about in your quote. Precisely the part of my link where you stopped reading, the part after OP acknowledged the limitations before moving on to describe and validate his solution. Not disingenuous at all on your part...

Your response is the kind you would make when you have nothing of substance to offer because you can't provide any proof of compromised accuracy.

BTW anyone can collect out of context statements to make someone look "bad", it's not that impressive.
 

GaryH

Major Contributor
Joined
May 12, 2021
Messages
1,348
Likes
1,804
Are you electing oratory1990 to the niche of top critic of budget measurement equipment?
He's a professional acoustic engineer with a master's in metrology who literally measures and helps design headphones/IEMs for a living. He certainly knows what he's talking about and how to judge the accuracy of acoustic couplers.
Individuals wishing to achieve comparable results on a low budget couplers go for individual calibration. Technically more ambitious than using provided calibrations oratory1990 talks about in your quote.
You obviously haven't bothered reading the second link I provided in which he explains why you cannot properly even individually compensate for a coupler that does not have the same acoustic impedance as a genuine one, as a single compensation would not be accurate for all IEMs.
Precisely the part of my link where you stopped reading, the part after OP acknowledged the limitations before moving on to describe and validate his solution.
Except he doesn't validate it for all IEMs, which you'd have to do due to the above acoustic impedance issue you ignored (at the very least in this case for the Truthear Zero if you're presenting data on it as valid and comparable to that from genuine couplers).
Your response is the kind you would make when you have nothing of substance to offer because you can't provide any proof of compromised accuracy.
Hah, sure. I've presented detailed explanations and multiple links to try to get you to learn about the actual science behind all this explained by professionals in the field, but you've just ignored them. I'm obviously just wasting my time. You're the one who's presented random graphs you've found on the internet of measurements using fake knock-off equipment as 'valid' data. The onus is on you to provide evidence specifically of this data's accuracy. You've failed to do so.
BTW anyone can collect out of context statements to make someone look "bad", it's not that impressive.
Nah you're doing just fine with that all by yourself, context and all.
 
Last edited:

GaryH

Major Contributor
Joined
May 12, 2021
Messages
1,348
Likes
1,804
By your metrics, then, something like the Stealth is also non-compliant since it has excesses and shortfalls relative to the target that are easily visible.
If you're referring to the predicted preference rating, that's not my metric, it's Dr Sean Olive's. And according to his professional measurements, the Stealth achieves 'excellent' compliance with the Harman target with a rating of 90%. If you're referring to the HzdB 'metric' for a deviation, as you can see from those measurements, it doesn't have any large (high in amplitude and/or low Q) deviations, with the greatest being around 900 HzdB between ~600 Hz and 1.5 kHz, around 10 times less than the Truthear Zero's 10 kHzdB deviation. I'm not pretending this is accurate, but it's better than just looking at amplitude of deviations, which it seems you're still falling onto the trap of.
You make it sound like it’s grossly non compliant with the target
No I haven't. I've repeatedly said its compliance is very good, but not excellent, and so its frequency response should not be conflated with the actual Harman target and be used to judge the latter.
I would not call a slight 1-2 dB treble lift a substantial deviation
If you still think a low amplitude but broadband (low-Q) deviation is 'slight' after I explained in the post I linked why this isn't slight in terms of audibility, referring to Dr. Floyd Toole's scientific research on the matter, then you need to read his book to understand this, I've done all I can.
Is it audible? Slightly, but not so much that I would call it “shouty.”
You're one person. Younger listeners with better hearing sensitivity around the frequencies the Truthear Zero is elevated above the Harman target are likely find it more shouty.
For the cost, this IEM is actually quite reasonable in its conformance to the target.
None of this has anything to do with value for money, it's about an objective measure of audio performance and how close the Truthear Zero is audibly to the Harman target. (But incidentally, the Zero is not particularly good value for money compared to other cheaper IEMs, especially considering its other flaws.)
And speaking more generally I certainly wouldn’t use an IEM alone as a reference for one’s preference of the Harman Target as has been pointed out since even if it had full compliance in the regions in question the IEM will not have the perceptual bass extension of a large headphone that complies and goes down below 20 Hz.
This poll is about the in-ear Harman target. This is distinct and differs from the over-ear target likely precisely due to the lack of any tactile bass afforded by IEMs, compared to at least a small amount of this from headphones (obviously still far less than speakers). But in terms of actual sonic bass extension, most IEMs are actually better than the majority of headphones due to the former operating primarily in pressure-chamber conditions, and with a perfectly sealed front volume this enables linear bass down to ~0 Hz. The Truthear fails to do this due to front volume venting, resulting in its relatively poor sub-bass extension (~5 dB short of Harman at 20 Hz) for an IEM, despite the 'subwoofer in an IEM' marketing.
 
Last edited:

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
If you're referring to the predicted preference rating, that's not my metric, it's Dr Sean Olive's. And according to his professional measurements, the Stealth achieves 'excellent' compliance with the Harman target with a rating of 90%. If you're referring to the HzdB 'metric' for a deviation, as you can see from those measurements, it doesn't have any large (high in amplitude and/or low Q) deviations, with the greatest being around 900 HzdB between ~600 Hz and 1.5 kHz, around 10 times less than the Truthear Zero's 10 kHzdB deviation. I'm not pretending this is accurate, but it's better than just looking at amplitude of deviations, which it seems you're still doing.

No I haven't. I've repeatedly said its compliance is very good, but not excellent, and so its frequency response should not be conflated with the actual Harman target and be used to judge the latter.

If you still think a low amplitude but broadband (low-Q) deviation is 'slight' after I explained in the post I linked why this isn't slight in terms of audibility, referring to Dr. Floyd Toole's scientific research on the matter, then you need to read his book to understand this, I've done all I can.

You're one person. Younger listeners with better hearing sensitivity around the frequencies the Truthear Zero is elevated above the Harman target are likely find it more shouty.

None of this has anything to do with value for money, it's about an objective measure of audio performance and how close the Truthear Zero is audibly to the Harman target. (But incidentally, the Zero is not particularly good value for money compared to other cheaper IEMs, especially considering its other flaws.)

This poll is about the in-ear Harman target. This is distinct and differs from the over-ear target likely precisely due to the lack of any tactile bass afforded by IEMs, compared to at least a small amount of this from headphones (obviously still far less than speakers). But in terms of actual sonic bass extension, most IEMs are actually better than the majority of headphones due to the former operating primarily in pressure-chamber conditions, and with a perfectly sealed front volume this enables linear bass down to ~0 Hz. The Truthear fails to do this due to front volume venting, resulting in its relatively poor sub-bass extension (~5 dB short of Harman at 20 Hz) for an IEM, despite the 'subwoofer in an IEM' marketing.
My hearing still goes to 16 kHz and just beyond, which should be acceptable to judge the tonality. Further the deviation I pointed to in the Stealth at around 2-3 kHz is around 2 dB and unlike the Truthear is an actual tonality defect and not some part of the response lifted. Is it an issue? I seriously doubt it. Additionally the Truthear has some boost to compensate for its lift in the bass region to offset for the lack of extension. How can we know that it will still sound balanced and reasonably neutral if the slight broad lift is taken out? Once could do tests, but I can say it will likely be detrimental. It has the same overall tonality as my EQd 2020 Sundaras and my floor system, but sans sub bass. I’m pretty sure it would make it start to sound a touch boomy. Also “better than the majority of headphones” doesn’t say much given how often they exhibit shortfalls in bass, and are likely capable of more bass that IEMs.

The problem here is that there are few IEMs that do conform more closely, and applying PEQ will never be perfect since the parameters will not fully track the deviations, and they are often based on one sample. You are treating these measurements and the Target like a precision micrometer and confidently ruling they do not conform closely enough without ever considering all the other factors involved. People are governed by evolutionary biology, not measurement fixtures. The furor over the Harman target’s bass is proof of that. Further, Sean commented that the lack of tactile bass means nearly 3 dB MORE bass is needs to be simulated over headphones to get a vis-a-vis perceptual response with respect to speakers (the study was with respect to car audio, but should still apply). This goes a long way to a potential explanation of the objections to its bass content. The slight treble lift is but one of a multitude of issues. here. Even handing out SOTA headphones you would still have divergences in opinions for a multitude of reasons.
 

cbracer

Member
Forum Donor
Joined
Jun 13, 2021
Messages
60
Likes
44
Location
California
We all seem to agree the correct sound for a speaker is a flat response. Not preference, but correct. So go listen to one, and then try your harman curve IEM and tell me it sounds similar. I'm not saying the exact same, that's not possible, but similar. I can't understand how anyone would not thing the harman curve has way too much upper base, low mids. It distracts from some music that use more of the full range of sound. It will sound great with some music like certain classical, but try just pop music and tell me it doesn't have too much bass compared to a flat response speaker? Sure add in a sub tuned to get the full range. Of course we all have preferences in sound, but is the harman curve correct to represent sound from a flat response speaker? No it's not. Do people like it, sure, but it's clearly not for audiophiles.
 
Top Bottom