Resolve's B&K 5128 Headphone Target - you can try the EQ's.....

[amirm]

I just found it kinda amusing they are doing exactly what they are complaining about in the same comment. I would expect better from a moderator

And I with you.....

 

[Resolve]

"mischaracterize what transpires" and implication that it has happened before is what I said. I read a lot of insults about me but this takes the cake. Don't call my ethics into question and then turn it into an "olive branch." That is as tone deaf as it gets.

You are heavily reading a particular narrative into this. Mad_Economist has only ever wanted to have an open dialogue with you. And I maintain, if you are willing to approach this topic in good faith, with the genuine intention to clarify your position and engage with ours, we can figure out a way to have that dialogue. Feel free to reach out privately if you'd prefer and we can discuss it.

 

[Robbo99999]

FYI and general thinking points. Throughout all of this back and forth dialogue and discussion I want to point out two Major Factors to consider the context and motivations behind this confusing conversation and accusations being leveled. Certain members have brought up names of members that have been banned from this site for good reasons that are openly documented in several places on this site. Bringing them up and involving them here has absolutely nothing to do with this technical conversation except to try and attack Amir’s reputation and standing in the Audio Community. A cheap shot and low blow. Demonstrates they are loosing the technical argument and have resorted to Character attacks.

Secondly, throughout this entire conversation/argument one person is working on your behalf and seeks no business or sales from you. We don’t sell anything here, we have no Commercial Overlords that we must serve. We have absolutely No Commercials or Advertisements here. The other participant(s) is here solely for the purpose of getting you to buy the products that they market and sell. One participant is fighting on behalf of the Community and trying to help prevent us from falling for the sales and marketing ploy that this new measurement process antiquates the previous standard that has been used for years and years. Is proven to be effective and accurate with a solid and long history. A shell game is afoot and it’s up to you the reader to determine who has your best interests at heart and what is each participants motivations here.

I don't think it's that simple between "us" & "them", and I don't really see it as an "us" & "them". Resolve & Co are indeed bound to have some commercial motivations, but I think that can also be aligned with genuinely trying to get the most out of the 5128 now that they've got it.....and hopefully in terms of them doing some mightily good studies/research on creating a good target curve for the 5128 that can get close to what Harman achieved with GRAS (big ask as I've said though). Having some kind of "us" & "them" war is not necessary, all that is required is that ASR readers & members & if possible the headphone enthusiast community at large to become educated enough on the basics to realise the limitations & differences between GRAS & 5128, mostly in terms of the fact that GRAS has solid research behind the Harman GRAS target curve and currently the 5128 does not.

Resolve & Co I'm sure still want to do their best with the 5128 in terms of creating a target curve, we just have to see where they end up, they're not at that point yet. In essence, there's really not much to argue about until they've created their final target curve & we can see transparently the work that was done behind it. It's true that it would have made sense for them to have done the work first though before actually using it in their reviews, and if they're not being clear about the TBC & more shaky nature of their target curve they're currently using for it then that would mislead their user base to a degree. I think I can say that they've probably not approached the usage of the 5128 in the most ideal way, I think they've rushed their usage of it, but I think they're genuine about trying to get the most out their 5128 now that they have it in terms of trying to create a Target Curve for it, and I'd like to see where they end up with it. Either way it's not useful to have an "us" & "them" mentality about it.

 

[Robbo99999]

(There is so much tension between Amir and Resolve & Co that I think the only way it will stop in this thread is if they have face to face discussions about everything that surrounds this project & the discussions that have resulted, either in video chat or in person.....I don't really see it being sorted out here in this thread or in any thread......and it wouldn't be a face to face that was recorded either, it would just be a pow wow!)

(I think this is colouring a lot of what is happening in this thread, and it's influencing everyone that's reading it, whether it be ASR's or headphone.com readership.)

 

[amirm]

There is so much tension between Amir and Resolve & Co that I think the only way it will stop in this thread is if they have face to face discussions about everything that surrounds this project & the discussions that have resulted, either in video chat or in person.....

The only way such face to face conversations make a difference is if the substance of the conversation is secondary to the manner with which we communicate. That is not so from my point of view. It is really simple: you either have research that a) shows the target curve for 5128 to significantly correlate with listener preference and b) that the 5128 is superior in this regard to modified 711 couplers in controlled listening test/measurement experiment, ala what Harman has done. Until then, it is premature to post content that assumes all this work has been done.

Tell me why I should have a conversation with them and back off from above position.

 

[Mad_Economist]

Oh yeh? What was this then:

"mischaracterize what transpires" and implication that it has happened before is what I said. I read a lot of insults about me but this takes the cake. Don't call my ethics into question and then turn it into an "olive branch." That is as tone deaf as it gets.

Quite literally - and as i said there - that is an assurance I was offering to you that even if you fundamentally mistrusted us, you could have a conversation with us on good faith without any fear that, even if we were maliciously motivated, we would be able to misuse it in any way. I offered this because you gave seemed to consistently read bad faith into what @Resolve and I have said in this thread, and that in past disagreements you've had of this sort (e.g. with Jude on Head-Fi), you've been misrepresented during or after the fact and I wanted you to be confident that it could not happen this time.

This was entirely an offer for your assurance that we would be held to good behavior, which is also why I was very explicit about being fine with chatting off the record if that were your preference.

Edit:
To be further clear, the "rough misunderstandings" bit references the various instances where you've read "an attack on everything we do here" into things I have written. This is a misunderstanding, given that I've been a reader of this site since its infancy, I generally like it, and if anything would prefer to see it flourish - which doesn't mean I have no complaints about it, including the pejorative and conspiratorial reading of my attempts to have a dialogue with you. But those occur in the open, with everything on display, so if communication in voice or in person didn't mend that problem, a recording sure wouldn't.

 

[amirm]

I like to take us back to this graph Harman has produced comparing the two fixtures:

Let's start from left, low frequencies (< 100 Hz). There we see that 5128 is showing substantially less energy there. That matches my experience that it is difficult to get a proper seal on the artificial head of 5128 as compared to the flat sides of the 45CA. This is documented by others as well. It has been said that 45CA overstates that energy. But this point is moot as the target is created with that already in mind. To wit, I have never, ever felt the need to boost the bass beyond what the target says for 45CA.

From 100 Hz up to about 6 kHz the two fixtures essentially agree. Between this and low frequency response, the story of the headphone tonality is basically told. What happens above is a minor thing but read below.

From 6 kHz and up, the two deviate. The 5128 shows a resonant peak whereas the 45CA mod shows an anti-resonance dip. With respect to over the ear headphones, I find that region to be a mess anyway. Any judgement in that region needs to be subjective and up to the listener as far as overall tonality.

Further, as the graph indicates, the Harman target or anything like that is going to mispredict the tonality with 5128 since it doesn't have that resonant peak. The simple solution is to modify the target to have that peak in it. Doing so just eliminates that issue with 5128 without accomplishing anything over the 45CA.

 

[amirm]

The raw measurement might be (I don't really know) but the compensation sure seems incorrect.

You noticed that too? Here it is again so others can follow:

I take "compensated" to mean difference between raw and target. This seems to just have some modifications of the same raw measurement.

 

[Blockader]

As a customer with a straightforward perspective, I have limited technical knowledge. I will express my opinions in a simple manner without providing extensive justifications. Here's how I perceive the situation:

1. BK 5128 is superior to GRAS KEMAR in accurately representing the acoustic impedance of the human ear. Despite any arguments made by Amir, it is evident to me that BK 5128 is a more precise device. Specifically, it excels in measuring headphones with high acoustic impedance, such as closed-back headphones or in-earphones. (Generally BK 5128 interacts closer to how real ears interact with the acoustic impedance of the headphone) Amir's side should address this point.
2. Now, what aspects should Mad/Resolve's side address? Well, there is a lack of comprehensive studies suggesting a reliable target for headphone preferences when measured with BK 5128. Amir is correct in highlighting this issue.
3. Mad_economist does not dispute this fact, while Amir denies the superiority of BK 5128 over GRAS KEMAR.

Any measurement rig that minimizes these differences above 2khz and below 200hz is a welcome change in the community. We know that headphones tonally tuned similar to how speakers with controlled directivity in a room sound like. DFHRTF + room curve gives a response target very similar to that. (Harman RR1 was designed with a similar approach to this solution for example)

I believe that mad_economist and Resolve are showing an innovative approach to accurately capture how headphones measure as experienced on real individuals' heads. Why don't we simply acknowledge and appreciate the innovative approach of Mad_economist? And why don't we show respect for Amir's adherence to traditional methods, considering that the old ways offer a more reliable preference target compared to the DFHRTF + room curve approach? One shouldn't completely null the other.

 

[Robbo99999]

The only way such face to face conversations make a difference is if the substance of the conversation is secondary to the manner with which we communicate. That is not so from my point of view. It is really simple: you either have research that a) shows the target curve for 5128 to significantly correlate with listener preference and b) that the 5128 is superior in this regard to modified 711 couplers in controlled listening test/measurement experiment, ala what Harman has done. Until then, it is premature to post content that assumes all this work has been done.

Tell me why I should have a conversation with them and back off from above position.

They are trying to get the 5128 to a point where there is a valid target curve for it - they could discuss their rough (or detailed) plans with you face to face & in a non-recorded fashion so neither they nor you (in terms of your reaction to those plans) have to commit to airing initial plans or non-complete plans. You've also been bringing up aspects related to "motivation" and "motive" surrounding their work, which is more than just factual scientific debate but is by definition related to people's character & personal motivations - so it wouldn't remain a sterile scientific discussion, hence the benefit of the non-recorded face to face. It could also help to alleviate any associated personal tensions or misunderstandings.

 

[Mad_Economist]

I like to take us back to this graph Harman has produced comparing the two fixtures:

[...]
Further, as the graph indicates, the Harman target or anything like that is going to mispredict the tonality with 5128 since it doesn't have that resonant peak. The simple solution is to modify the target to have that peak in it. Doing so just eliminates that issue with 5128 without accomplishing anything over the 45CA.

I have deja vu all over again!

I decided to respond to this one by making a publicly viewable google sheet - please feel free to "make a copy" if you feel inclined to play with it. If I recall Sean's presentation correctly, shortly after that slide should be one about the adapted Harman target for the 5128, which was produced using the average of the difference between the headphones measured on the 45CA with the Welti pinna and the 5128.

From Sean's twitter, we have this very interesting plot

which showed the average difference in response for a set of over-ear headphones for different fixtures, and a population of humans, measured with blocked canal mics. Since it happens to include both the "official" Welti pinna clone, and the commercially available KB501x that @amirm and, amusingly, Sean on his twitter use, we can plot the average difference between these pinnae, and compare it with the 5128's difference to the 60318-4+Welti pinna - this is what we get:

That is, the KB501x also has a quite noticeable peak in the treble relative to the Welti pinna, and if we "decompensate" the Welti pinna measurement (adding the delta rather than subtracting it, as with the 5128), we get a scary looking result as well

Will measurements using the KB501x (or for that matter, the KB500x, the KEMAR/43AG equivalent) be compensated to adjust for this average difference? Should we cast them aside, due to the substantial impact (20% preference prediction difference!) of the different pinna? Well, that wouldn't be my recommendation, but based on your commentary I take it that you'll be adding this adjustment to future reviews?

From 100 Hz up to about 6 kHz the two fixtures essentially agree. Between this and low frequency response, the story of the headphone tonality is basically told. What happens above is a minor thing but read below.

Were this true, their scores would be more similar - it's reasonable to say that the difference below 1khz can be attributed to leakage response, which isn't a consistent factor, however

This is documented by others as well. It has been said that 45CA overstates that energy. But this point is moot as the target is created with that already in mind. To wit, I have never, ever felt the need to boost the bass beyond what the target says for 45CA.

This is, it must be noted, not a static effect - it's variable between headphones, placements, and operators. Indeed, the more problematic version of this is (atypically very high leakage vs. human experience) is directly what prompted Todd to make his modified pinna, because excessive leakage on the fixture was producing meaningful errors in predicting preference. It's an area where, as Sean has noted, in situ measurements on humans would be highly useful to determine the actual perceived timbre of headphones.

You noticed that too? Here it is again so others can follow:

I take "compensated" to mean difference between raw and target. This seems to just have some modifications of the same raw measurement.

Ironically, this is a perfect example of the sine illusion, which is why I'm pushing for compensated measurements. @Resolve is welcome to post these actual measurements, but I just did a quick scrape with vituix, and performed the compensation both by dividing in REW and arithmetically in Excel

The only deviations I see are tiny dips and peaks due to the scraping, but the "compensated" plot in your screenshot is, indeed, the result of subtracting the target from the raw response.

 

[isostasy]

@Mad_Economist I think I follow what you're saying in theory but @amirm's measurements don't show this treble peak and look more similar to the Welti pinna measurements in your spreadsheet than the inferred KB5000:

I've followed his reviews for at least a year or so now, and looked through many past ones, and a dip rather than a peak between 8-10kHz seems to be the consistent feature of the measurements on his equipment.

There's a high chance I'm completely misunderstanding here though.

 

[markanini]

@Mad_Economist I think I follow what you're saying in theory but @amirm's measurements don't show this treble peak and look more similar to the Welti pinna measurements in your spreadsheet than the inferred KB5000:

I've followed his reviews for at least a year or so now, and looked through many past ones, and a dip rather than a peak between 8-10kHz seems to be the consistent feature of the measurements on his equipment.

There's a high chance I'm completely misunderstanding here though.

Maybe because Amirs rig uses a hi-res coupler, similar to the GRAS RA0402.

 

[Blockader]

@Mad_Economist I think I follow what you're saying in theory but @amirm's measurements don't show this treble peak and look more similar to the Welti pinna measurements in your spreadsheet than the inferred KB5000:

I've followed his reviews for at least a year or so now, and looked through many past ones, and a dip rather than a peak between 8-10kHz seems to be the consistent feature of the measurements on his equipment.

There's a high chance I'm completely misunderstanding here though.

you are completely misunderstanding it. Basically during measurements, acoustic impedance of headphones interact with the acoustic impedance of the coupler. The measurement you see is a result of that. The measurements taken with different measurement rigs show different responses because measurement rigs might have difference couplers with different acoustic impedance. You can't see these differences mad mentioned unless the GRAS measurement is compared to the BK 5128 measurement.

 

[Mad_Economist]

@Mad_Economist I think I follow what you're saying in theory but @amirm's measurements don't show this treble peak and look more similar to the Welti pinna measurements in your spreadsheet than the inferred KB5000:

I've followed his reviews for at least a year or so now, and looked through many past ones, and a dip rather than a peak between 8-10kHz seems to be the consistent feature of the measurements on his equipment.

There's a high chance I'm completely misunderstanding here though.

Yes, I should probably mark my sarcasm a bit more clearly - my rhetorical point here is not that Amir should do this, but rather that "compensating" a 5128 measurement cannot transform it into a 45CA measurement. Different pinnae and different headphones interact in novel ways, and attempting to compensate one HATS into another is foolhardy. This is, indeed, part of why measuring with multiple ears is so interesting - and, thus, a large chunk of why I was keen on the 5128 (and why I keep obliquely referencing in-ear mic measurements, which we really need to do more of).

My point here is that ears are not substitutes for each other. They are "non-fungible" so to speak. It is not that Amir has an inferior alternative to Todd's pinna, it's that he has a different pinna, and the data it produces is itself quite interesting.

 

[peniku8]

FYI and general thinking points. Throughout all of this back and forth dialogue and discussion I want to point out two Major Factors to consider the context and motivations behind this confusing conversation and accusations being leveled. Certain members have brought up names of members that have been banned from this site for good reasons that are openly documented in several places on this site. Bringing them up and involving them here has absolutely nothing to do with this technical conversation except to try and attack Amir’s reputation and standing in the Audio Community. A cheap shot and low blow. Demonstrates they are loosing the technical argument and have resorted to Character attacks.

Secondly, throughout this entire conversation/argument one person is working on your behalf and seeks no business or sales from you. We don’t sell anything here, we have no Commercial Overlords that we must serve. We have absolutely No Commercials or Advertisements here. The other participant(s) is here solely for the purpose of getting you to buy the products that they market and sell. One participant is fighting on behalf of the Community and trying to help prevent us from falling for the sales and marketing ploy that this new measurement process antiquates the previous standard that has been used for years and years. Is proven to be effective and accurate with a solid and long history. A shell game is afoot and it’s up to you the reader to determine who has your best interests at heart and what is each participants motivations here.

Financial incentive to sell us what exactly? Are headphones.com readers rushing out to buy 5128's?

 

[markanini]

Financial incentive to sell us what exactly? Are headphones.com readers rushing out to buy 5128's?

No, OTOH a larger revenue can be had by upselling headphones, based on hype. That's where providing graphs from a novel rig, that can't be humanly interpreted, rendering them useless, comes in handy.

 

[Guess it]

Your behavior here has indicated precisely that having it be on the record is a good idea!

Amir... I'm going to make this as clear as I possibly can. I want to give you every opportunity to engage with the actual points being put forward in good faith, absent the accusations of ulterior motives or your continued attempts to discredit what we're doing, which is so far predominantly what you've done here in this thread and quite frankly in any other interaction I've had with you. With that said, I believe your position on this topic is an important one to address, and in so doing we can get to a better outcome and we can all understand each other, clearly and professionally. If this is something you are interested in, we can figure that out.

If not, that's fine too - but I'm not going to continue trading barbs with you about any of this. Have a good weekend.

So what is the motivation behind being a reviewer for a store front? Just for the access to a nice hats?
Does receiving an expensive hats not give you a sense of authority as a reviewer which you rely on for customers and that your viewers defer to you partially for?
Are you not in competition with amir? All this subtle discrediting is really beneficial to headphones.com not just for traffic or image but because authority over the narrative means maintaining the ability to manipulate perception of and alternative to their consumer facing justifications (such as the hard to validate 5128) that give leverage against their suppliers etc. The potential to aim for hidden variables in the equations you want strangers to involve with is massive given your bold plan so I'd probably look to disclose rather than deflect in the face of honest speculation about your intentions instead of writing off accusations as spurious slander, Hon.

Don't forget you're filling the convenient vacuum tyll left who was known for his genuine care with no qualms abstaining from potential paths for the sake of integrity despite not always having full independence.

Also yes, science is historically racist, wanna talk about how it can be neoliberal presenting false choices to trick the well meaning into supporting elitist structures such as luxury goods markets? Metrology isn't a high ground so it's really funny to pretend provides purity of opinion and incentive or can't be used for economic hemming. Hope you like your garments. That Audio Science Server is tight! Shame they're not public like a forum.

 

[amirm]

They are trying to get the 5128 to a point where there is a valid target curve for it - they could discuss their rough (or detailed) plans with you face to face & in a non-recorded fashion so neither they nor you (in terms of your reaction to those plans) have to commit to airing initial plans or non-complete plans.

You don't call a meeting first and then figure out the agenda later. I have stated what needs to happen: a proper research project as done by Harman. I have seen no receptivity to that. If they do want to do it, then the blueprint is there in Harman papers and if they have questions, they can ask Sean.

I have also asked them to refrain from publishing results on 5128 until these issues are sorted out and we don't create confusion in the market. I already see people posting mix of 5128 and 711 here. It is a mess and getting worse. Again, no receptivity on this either.

You've also been bringing up aspects related to "motivation" and "motive" surrounding their work, which is more than just factual scientific debate but is by definition related to people's character & personal motivations - so it wouldn't remain a sterile scientific discussion, hence the benefit of the non-recorded face to face. It could also help to alleviate any associated personal tensions or misunderstandings.

None of this is an issue with respect to what I am asking above. You could be my best friend doing what they are doing and my response would be the same.

 

[solderdude]

I have deja vu all over again!

I decided to respond to this one by making a publicly viewable google sheet - please feel free to "make a copy" if you feel inclined to play with it. If I recall Sean's presentation correctly, shortly after that slide should be one about the adapted Harman target for the 5128, which was produced using the average of the difference between the headphones measured on the 45CA with the Welti pinna and the 5128.

From Sean's twitter, we have this very interesting plot

which showed the average difference in response for a set of over-ear headphones for different fixtures, and a population of humans, measured with blocked canal mics. Since it happens to include both the "official" Welti pinna clone, and the commercially available KB501x that @amirm and, amusingly, Sean on his twitter use, we can plot the average difference between these pinnae, and compare it with the 5128's difference to the 60318-4+Welti pinna - this is what we get:
View attachment 297853
That is, the KB501x also has a quite noticeable peak in the treble relative to the Welti pinna, and if we "decompensate" the Welti pinna measurement (adding the delta rather than subtracting it, as with the 5128), we get a scary looking result as well
View attachment 297854
Will measurements using the KB501x (or for that matter, the KB500x, the KEMAR/43AG equivalent) be compensated to adjust for this average difference? Should we cast them aside, due to the substantial impact (20% preference prediction difference!) of the different pinna? Well, that wouldn't be my recommendation, but based on your commentary I take it that you'll be adding this adjustment to future reviews?


Were this true, their scores would be more similar - it's reasonable to say that the difference below 1khz can be attributed to leakage response, which isn't a consistent factor, however
View attachment 297856

This is, it must be noted, not a static effect - it's variable between headphones, placements, and operators. Indeed, the more problematic version of this is (atypically very high leakage vs. human experience) is directly what prompted Todd to make his modified pinna, because excessive leakage on the fixture was producing meaningful errors in predicting preference. It's an area where, as Sean has noted, in situ measurements on humans would be highly useful to determine the actual perceived timbre of headphones.


Ironically, this is a perfect example of the sine illusion, which is why I'm pushing for compensated measurements. @Resolve is welcome to post these actual measurements, but I just did a quick scrape with vituix, and performed the compensation both by dividing in REW and arithmetically in Excel
View attachment 297857
The only deviations I see are tiny dips and peaks due to the scraping, but the "compensated" plot in your screenshot is, indeed, the result of subtracting the target from the raw response.

It would seem that the 8kHz peak is in all measurements and therefor should be in the 'error correction' curve but for some reason the 'target' only has a too small 'bump' there.
Why not completely correct it ?
That would make a lot more sense than 'get used to seeing it'.
That also explains the Stealth measurement from Jude.

Seeing as there is a substantial and more 'sensible' under reporting of the bass in the 5128 ? you can remove the 'Harman' bump in the target and make the target in the lows almost flat.

Is that leakage only ? Why on earth would one leak so much that there is a -10dB drop on open headphones. That leak would be substantial. Something else going on ?

Ha.. that explains exactly why I came up with my own bass correction (done by ear and eyeballing other HD650 measurements) for my pinnaless FP (so no 9kHz dip).

Seems to match the research above perfectly. It is done in hardware (including a 16kHz mic related correction, not shown) so 'raw', for me, is already corrected and do not need to apply a target afterwards.

I guess the red trace is the 5128 ?