New Neumann KH 310 II and KH 350 Rumors: Leaked Pricing & Info

[Curvature]

Again, the theory based on measurements can be wrong, the measurements can be wrong, but ''science'' isn't.
I think it's important to qualify things properly.

You're equating science with truth while I treated it like knowledge (which is etymologically accurate) or a collection of methods.

 

[sysfc6]

I already know that the KH350 is a vertical 8-inch three-way speaker.

 

[dfuller]

1. Listening to opinions and looking at measurements can be very useful to weed out all the "not so great"/ subpar/ badly designed/ faulty etc... products, but above a certain "measurably/objectively good/great" level you just have to audition them.

2. A few years ago when I was planning an upgrade, I looked at the measurement differences between the D&D 8C, the Neumann KH420, and the Genelec 8361... their measurement data were some of the best at the time, objectively/scientifically they're all great. But than when I had a chance to compare these monitors side by side I quickly realized there is a huge character/sound signature difference between them. And that was true even when I've done everything I could to eliminate all variables (eg. distance/height/SPL level etc.).
So I'm 100% sure that
- the driver material and design,
- the power handling,
- the chosen crossover frequencies,
- ported vs sealed cabinet design,
- point source/coaxial vs traditional driver arrangement
-etc, etc.
so the sum of all of these design choices are way more important than what can be shown you on some measurement graphs. These design choices create clearly audible differences even when the actual measurements would suggest a much more "similar sound". Why? I guess because of it's the combined effect of many small differences. (Like when watching a great engineer working on really well recorded material. You see them doing all these teeny-tiny fader/setting moves on level, panning, EQ, Compressor, distortion/saturation etc., on their own all hardly audible BUT: when they're done, and you listen to the rough mix starting point and their finished work, bam, it's night and day.) I heard much greater character differences between these monitors than what I was expecting based on the measurement data.

3. All 3 of the monitors I just mentioned above were having different strength and weaknesses, all great but at the same time very different, they have clearly different "personalities".
E.g. the feeling what the bass punch and presence creates (while still maintaining mad amount of low end detail) on the 8C is nothing I've ever experienced on any other similarly sized monitors without subs. Or how the 8361 suggests some mix choices in the vocal region that are a bit less obvious on other speakers.
To put it simply: after testing these monitors I'm 100% convinced that a great sound engineer could deliver great results on any of these but the order of their mixing moves would be slightly different because each of these monitors pulls the attention in slightly different directions.
I'm oversimplifying it here, but while listening to a song the first time
- the KH420 sort of forced me to focus on the "big picture";
- the 8C pulled my attention towards the bass/ drums/ groove;
- while the 8361 directed my attention towards the vocals/guitars etc.
To put it simply: I could hear everything on all of these, but there is a clear difference where it took my ears/focus first + how much attention I needed on certain areas to pick up small details.

4. Small positioning differences can matter a lot! Auditioning/comparing in your own space is a must. And while doing your comparisons, be very precise when switching from one speaker to the other. Try to keep the variables like listening distance, listening height, SPL level etc. minimal. It's annoying, it needs a lot of positioning and measuring but well worth putting the effort in; this is the only way to actually hear the real differences between the speakers. Also worth doing REW measurements and analyzing/comparing the results of the different speakers. Just putting monitors side by side, completely different listening distances, often one at an equilateral triangle while the other/s thrown next to it far too wide out...
Look at this "review"/comparison video of the new KH310ii vs KH350 vs KH150:

I remember this post, this was on my thread about upgrading from 310s!

 

[Karmacoma]

You're equating science with truth

Where? Quote me.

I treated it like knowledge (which is etymologically accurate) or a collection of methods.

Fair, but both are very different definitions.
That's why I think it's best to talk about scientific observations, experiments or theories on one side and, I'll precise, the scientific method and more broadly the scientific enterprise on the other.
Saying ''this contradicts science'' or ''science is wrong" is incorrect, and that has nothing to do with truth or fallacy.
Saying this contradicts ''the'' science we have in that domain means something else.

 

[b73i57dc3f]

the actual measurements would suggest a much more "similar sound"

I don't quite follow? Just a quick glance at the directivity contour plots reveals that there are significant differences between all three speakers.

- the KH420 sort of forced me to focus on the "big picture";
- the 8C pulled my attention towards the bass/ drums/ groove;
- while the 8361 directed my attention towards the vocals/guitars etc.

Isn't that in good agreement with what the measurements show? The 8C clearly shows the most elevated bass, while the 8361A clearly shows an elevation in the upper midrange in its estimated in-room response (due to wider directivity in that area).

 

[goat76]

Again, the theory based on measurements can be wrong, the measurements can be wrong, but ''science'' isn't.
I think it's important to qualify things properly.

So far the arguments made on this discussion against the theory are particularly weak, it's important for those who keep using this argument to understand that ("a lot of great engineers use ATC's"), and since we don't even know how the big ATC's measure, partly because the company -oddly- refuses to provide measurements, all of this debate, for or against ATC, for or against the theory is totally and utterly baseless.

When reading this thread, people have said they will be waiting for third-party measurements as they don't fully trust the measurements Neumann has provided for their loudspeakers, why is that?

ATC has said numerous times that they don't want to enter the numbers game, and the reason for that is that it's too easy to manipulate the numbers. The same reason many hear don't trust the measurements Neumann provides.

I have not seen anyone in this thread saying that ATC speakers are “better for mixing than other speakers”, as @Curvature put it. I find it a bit overly argumentative to suggest that. I find it better and more honest for the discussion if people keep things to what has actually been said, that ATC speakers are a great choice for studios for many more reasons than just sound quality, which may be true for Neumann speakers as well.

 

[kimmosto]

The latest paper I read on this subject was Audibility of Group-Delay Equalization, quite a while ago and those listening tests were done with headphones, not speakers.

GD of KH 310 hits between the limits in yellow. Ca. 0.8 ms at 500-1000 Hz is not really bad result though it is not clearly on the safe side.

Sure that study is limited due to headphones and click-type signal and limited frequency band and impractical method to produce phase distortion as GD peak. The reason was probably to study spectrum of audibility threshold. Spectrum due to IIR XO is a step with possible hump. That is audible down to bass range with close to same limits with 500 Hz. For example, all these are audible with repeated transient signal:

Commercial science most often bears some scent of the manufacturer's product portfolio, ensuring that possible weaknesses are not revealed. For example, hiding behind psychoacoustics filtered with test signals such as soft elevator music or limited frequency range.

When it comes to linear phase, I see no way of it reaching me in my normally reflective room and from what I understand my brain does not seem to care.
But I might very well be misunderstanding what you mean and I simply do not know what I do not know.

Distortion in source sound travels to listener in direct sound and reflections so transient content will be distorted due to temporal dispersion with speakers too. In addition, first ca. 3 ms of transients is direct sound also with loudspeakers.

Edit: Re-uploaded the screenshots at the usual 50dB scale

What a waste of empty space with such a good responses

 

[Ype]

GD of KH 310 hits between the limits in yellow. Ca. 0.8 ms at 500-1000 Hz is not really bad result though it is not clearly on the safe side.
View attachment 559025
Sure that study is limited due to headphones and click-type signal and limited frequency band and impractical method to produce phase distortion as GD peak. The reason was probably to study spectrum of audibility threshold. Spectrum due to IIR XO is a step with possible hump. That is audible down to bass range with close to same limits with 500 Hz. For example, all these are audible with repeated transient signal:
View attachment 559027
Commercial science most often bears some scent of the manufacturer's product portfolio, ensuring that possible weaknesses are not revealed. For example, hiding behind psychoacoustics filtered with test signals such as soft elevator music or limited frequency range.

Distortion in source sound travels to listener in direct sound and reflections so transient content will be distorted due to temporal dispersion with speakers too. In addition, first ca. 3 ms of transients is direct sound also with loudspeakers.

What a waste of empty space with such a good responses

Yes then we are understanding each other properly and I was thinking of the right paper.
Appreciate you taking the time!

And thank you, Neumann's engineering team are the stars of the show though, I just added some salt and pepper.
We'll know for certain when we see more measurements but I hope their MK2 to be even more excellent.

I think that is all I could add to this thread, will go back to reading.

 

[ozonepaul]

Yes, the 8C has more and deeper bass output than the 8361a and KH420, plus cardioid..
When you see a "bigger" speaker - it sounds "bigger".
No one would expect 8c, 8361a, KH420 to sound "the same"???

The measurements and science of listening is all there. There was no magical discovery when listening.
There will always be personal prefetence, but the spinorama provides way more informstion than people give credit for.

1. The KH420 (645 x 330 x 444 mm) is bigger than the 8C (485 x 270 x 380 mm) so I honestly don't get your point.
2. I never said I was expecting "the same" sound, I said
a.) ...clearly audible differences even when the actual measurements would suggest a much more "similar sound"...
b.) ... I heard much greater character differences between these monitors than what I was expecting based on the measurement data....
3. If you wish, for me there was some magical discovery: above a certain measurable/objective monitor quality, the link between small measurement data (including spinorama) differences and how it influences good mixing decisions is way more complicated than what I would have guessed.

Of course we are curious, i would love to know if there is something else we are missing here, but it has to be a rigorous scientific fact, you are asking us to believe you on base of anecdotal data only. The burden of proof is not in our side. We do not negate your experience but I am also an expert in my area, and biases have made me humble multiple times. we are not being unscientific, but the contrary.

I'm not asking anyone to believe anything based on anecdotal data: it's not an anecdotal data that loads of successful engineer's preferred monitor choice for work is ATC (often even after long and tedious monitor tests/comparisons). I suggested it's probably down to a bit more than marketing, blind brand loyalty or engineer voodoo. That's all.

"I really do not understand why loads of users on this forum are not curious why?"

And the circle of confusion goes round and round.

Do you really don't understand the needs for standards if translation is your goal?
Do you truly do not understand why having resonances added by your speaker, which color all sounds the same way, is undesirable?
Do you not see how the person you responded to before complaining about the Genelec mids is driving this exact point home, namely, a lack of standards?

I truly do not see how one could not, as you yourself auditioned speakers that would also be recommended here.
Part of you has to get it, surely.

"I find it laughable how can scientific inclined people be this unscientific."

*Again, my point is the view on speakers and standards in general, not a certain brand in particular.
I do not care if my speakers says Neumann, Genelec, ATC or PMC, I care about sound quality.

It is about the idea of standards and why they are important.
Neumann and Genelec follow the standards so I recommend them, it is as simple as that.

If either company puts a melon on the market I will not recommend it.
If ATC and PMC do not publish measurements or the measurements show audible resonances I will not recommend them.

How is any of this difficult to grasp without purposely misunderstanding?
Would you truly expect me to recommend studio's to buy speakers from Trust me bro Inc just because others mix on them?

Perhaps you should try first understand before trying to ridicule
Presuming you are a mixing engineer or mastering engineer, your job is literally to make good sounding recordings to be played back in people's homes.
I would hope you care about the preferences of the consumers and standards that will drive your industry forward.

So I am not quite sure why you are the one laughing.
See this is my final response to you, no hard feelings, but I am not paid to educate you on speaker design nor help you get better at what I presume is your job.
Or perhaps I am just mis-interpreting your post.

I do think you completely mis-interpreted my post. I don't even think our positions on the subject are that different. I agree with most of your concerns you started with "do you really don't", "do you truly not" etc. I spend most of my time in front of Neumanns (KH310s with KH750 and MA-1 and KH420s that I'm sure you would happily recommend); I never asked anyone to recommend ATC or any other brand; I don't find it hard to grasp what you say, I'm not laughing (but I still find "laughable that bashing ATCs is sort of a sport on ASR"), I'm not ridiculing anyone... and I would never expect you "to recommend studio's to buy speakers from Trust me bro Inc just because others mix on them".
But I would expect a scientific inclined community to be a bit more tolerant and curious about the preferences of engineers. That's all.

Two reasons:

1. There are almost no PMC or ATC measurements available.
2. The measurements which are available suggest that the speakers are not worth the money and don't live up to their reputation.

On the one hand, if PMC or ATC have undefinable qualities which make them better for mixing than other speakers, it means the science is not only wrong but deeply wrong, and that artists have no choice but to spend the kind of money they charge to make good music. On the other, if a lot of what is believed is simply false or misguided, then artists are much freer to make clear technical assessments of their equipment and not be bothered with so many irrelevant opinions about gear that contribute nothing to the craft.

I hope we can easily agree that the chance that science (as far as the method and usefulness of monitor measurements) is deeply wrong is near zero

+ In my previous post I went into detail why some ATCs are loved in the industry and I said that these "qualities/services" are all important in a pro studio.
You say: "The measurements which are available suggest that the speakers are not worth the money and don't live up to their reputation." My guess is that you make this assessment because for you "worth the money" does not include most of my points (from reliability to great customer care etc.).

I don't quite follow? Just a quick glance at the directivity contour plots reveals that there are significant differences between all three speakers.

Isn't that in good agreement with what the measurements show? The 8C clearly shows the most elevated bass, while the 8361A clearly shows an elevation in the upper midrange in its estimated in-room response (due to wider directivity in that area).

You're right, there are differences and you are free to say that my experience was in agreement with what the measurement show. For me though it was surprising. If you wish, my interpretation of those measurement graphs did not prepare me for such obvious audible differences. And even more importantly, it did not prepare me for e.g. how much influence it will have on the order of my mixing choices.
As I said in my earlier post: "A TEST SIGNAL IS NOT MUSIC, it's not the same thing. Hence measurements (based on these test signals) can help make informed predictions about roughly how an actual monitor will play back real music, but still those measurements are "only" good for informed predictions; it's not a 100% full representation of the monitors real life behaviour when it receives the complex (ever changing) signals of real music." If you agree with what I just said, than you can say that my informed predictions were off. You can even say that yours would have been better, I'm absolutely fine with that.
I said earlier that "the way a measurement signal "tortures" a tweeter, a mid or bass driver is NOT the same as how an actual musical event in an extremely busy or extremely sparse (or everything in between) mix "tortures" a tweeter, a mid or bass driver." In my head it's a bit like the controlled environment data collections of crash tests (which clearly improves driver/passenger safety) vs the complexity of real life accidents.


Finally, I completely agree with goat76: it would be great if the discussion would be based on what has actually been said.

 

[dfuller]

I'm not asking anyone to believe anything based on anecdotal data: it's not an anecdotal data that loads of successful engineer's preferred monitor choice for work is ATC (often even after long and tedious monitor tests/comparisons). I suggested it's probably down to a bit more than marketing, blind brand loyalty or engineer voodoo. That's all.

Probably it's just that - and this I'd say a hard thing for people who aren't engineers to understand, at some level - that even a speaker that isn't "perfectly flat" can still be a useful tool, if it reduces revisions and speeds up work. ATC aren't really even out of the realm of "neutral", either - are they as flat as a Neumann or Genelec, no, but they're certainly in the realm of "good enough" that I don't think they meaningfully impede users from making good mix decisions on them.

 

[Purité Audio]

But why settle for ‘good enough’, particularly when one considers the performance/cost ratio when more transparent designs are available, presumably ‘engineers’ enjoy the challenge?

 

[Ype]

1. The KH420 (645 x 330 x 444 mm) is bigger than the 8C (485 x 270 x 380 mm) so I honestly don't get your point.
2. I never said I was expecting "the same" sound, I said
a.) ...clearly audible differences even when the actual measurements would suggest a much more "similar sound"...
b.) ... I heard much greater character differences between these monitors than what I was expecting based on the measurement data....
3. If you wish, for me there was some magical discovery: above a certain measurable/objective monitor quality, the link between small measurement data (including spinorama) differences and how it influences good mixing decisions is way more complicated than what I would have guessed.




I'm not asking anyone to believe anything based on anecdotal data: it's not an anecdotal data that loads of successful engineer's preferred monitor choice for work is ATC (often even after long and tedious monitor tests/comparisons). I suggested it's probably down to a bit more than marketing, blind brand loyalty or engineer voodoo. That's all.



I do think you completely mis-interpreted my post. I don't even think our positions on the subject are that different. I agree with most of your concerns you started with "do you really don't", "do you truly not" etc. I spend most of my time in front of Neumanns (KH310s with KH750 and MA-1 and KH420s that I'm sure you would happily recommend); I never asked anyone to recommend ATC or any other brand; I don't find it hard to grasp what you say, I'm not laughing (but I still find "laughable that bashing ATCs is sort of a sport on ASR"), I'm not ridiculing anyone... and I would never expect you "to recommend studio's to buy speakers from Trust me bro Inc just because others mix on them".
But I would expect a scientific inclined community to be a bit more tolerant and curious about the preferences of engineers. That's all.


I hope we can easily agree that the chance that science (as far as monitor measurements) is deeply wrong is near zero

+ In my previous post I went into detail why some ATCs are loved in the industry and I said that these "qualities/services" are all important in a pro studio.
You say: "The measurements which are available suggest that the speakers are not worth the money and don't live up to their reputation." My guess is that you make this assessment because for you "worth the money" does not include most of my points (from reliability to great customer care etc.).



You're right, there are differences and you are free to say that my experience was in agreement with what the measurement show. For me though it was surprising. If you wish, my interpretation of those measurement graphs did not prepare me for such obvious audible differences. And even more importantly, it did not prepare me for e.g. how much influence it will have on the order of my mixing choices.
As I said in my earlier post: "A TEST SIGNAL IS NOT MUSIC, it's not the same thing. Hence measurements (based on these test signals) can help make informed predictions about roughly how an actual monitor will play back real music, but still those measurements are "only" good for informed predictions; it's not a 100% full representation of the monitors real life behaviour when it receives the complex (ever changing) signals of real music." If you agree with what I just said, than you can say that my informed predictions were off. You can even say that yours would have been better, I'm absolutely fine with that.
I said earlier that "the way a measurement signal "tortures" a tweeter, a mid or bass driver is NOT the same as how an actual musical event in an extremely busy or extremely sparse (or everything in between) mix "tortures" a tweeter, a mid or bass driver." In my head it's a bit like the controlled environment data collections of crash tests (which clearly improves driver/passenger safety) vs the complexity of real life accidents.


Finally, I completely agree with goat76: it would be great if the discussion would be based on what has actually been said.

Appreciate the response, thank you for also driving towards clarity!
Will not add more noise but I too think we are mostly in agreement, at that point is simply a question of prioritization and execution.
I too would not want to compromise on some of the points you made either, but they do not negate the concerns.

My time and energy is rather limited these days but wanted to end our interaction on a hopefully positive note.

 

[ozonepaul]

But why settle for ‘good enough’, particularly when one considers the performance/cost ratio when more transparent designs are available, presumably ‘engineers’ enjoy the challenge?

Even if I accept the presumption that ATCs are only "good enough": the answer is perhaps partly because in a pro studio the question is not only performance vs cost ratio but also reliability/customer care/serviceability vs cost ratio.

+ If I'm a bit cheeky, than I guess following your logic buying an Alfa Romeo (compared to the competition it's expensive, it's even prone to mechanical failure - which the ATC isn't -) doesn't make any sense. Weirdly enough it's one of the most popular brands amongst car enthusiasts, racers, car journalists etc....presumably they also enjoy a challenge.

 

[Karmacoma]

I'm not asking anyone to believe anything based on anecdotal data: it's not an anecdotal data that loads of successful engineer's preferred monitor choice for work is ATC

It is an appeal to anecdote/popularity/authority.
It's important to understand that it is a very weak argument to choose ATC.

Once again not saying ATC's are bad. I just don't know.

 

[Ladipoe@98]

The 310 II has my attention. I’d be more interested in the measurements than the leaked specs, though.

 

[ozonepaul]

It is an appeal to anecdote/popularity/authority.
It is a clear fallacy. It's important to understand that it is a very weak argument to chose ATC's.

Once again not saying ATC's are bad. I just don't know.

Where is the appeal to anecdote/popularity/authority in me saying that "loads of successful engineer's preferred monitor choice for work is ATC"? What if the popularity is based on actual performance and not marketing and woodoo? What if authority is demonstrably paired with knowledge, experience and wisdom? And what if it's not an anecdote?

- If you just run through the gear lists of professional commercial studios in London, Paris, Prague, Berlin, Budapest, New York, Los Angeles, Nashville, Buenos Aires, Santiago, Sao Paolo, Istanbul, Tokyo...and you often bump into ATCs than it's not an anecdote.
- If you just randomly checked studio visit videos from channels like Rick Beato, Produce Like a Pro, Studio Time, Agartha Studio Tours etc. and you often bump into ATCs than it's not an anecdote.
- If you happen to work in loads of different studios and you recognize that, "hm so many studios are kitted with ATCs" than it's not an anecdote.
- For God's sake, if northwardacoustics (one of the most well known high-end studio design company) who built "more than 200 projects spanning over 20 countries, 4 continents" pretty much only works with ATC than where is the appeal to anecdote in what I said?
Instead of arguing with me, look at their list of works from 2009-2016: https://northwardacoustics.com/on-the-job/on-the-job-2026/
Do you see anything else than ATCs? If not than why are you so argumentative when I just stated the fact that "loads of successful engineer's preferred monitor choice for work is ATC". That's not an anecdote, that's a fact.
I suggest you to read what Thomas Jouanjean from Northward says about why he builds his high-end studios around ATC systems:
"The reason I stick with ATC is that they are simply, in my experience, the best speakers you can get: they are the best compromise..."
His full post is here (worth reading some of his other posts as well):
https://www.audiosciencereview.com/...c-speakers-monitors.17448/page-13#post-938413
And if even after reading his posts, checking his references etc. you still think that what I said is just "appeal to anecdote/popularity/authority" than I honestly can't help you.


+ Why do you say "it's important to understand that it is a very weak argument to chose ATC's"? Who argued for choosing ATC? I never did, I even mentioned that I'm on Neumanns. But even if someone would have argued, why is it a weak argument? What if someone not even an engineer, just a vocal editor said something as simple as:
"even though I tried my friends KH310s, and I also tried two Genelecs, (the 8351 and the 8361) in my editing room, I tend to pick up sibilant sounds faster on my SCM25 and I find it much easier to judge vocal editing moves on the SCM25 than on those other monitors. So I feel like I make decisions quicker, hence I work faster. On top I don't like monitors with too much bottom end in this room, it can easily muddy the sound, plus I find that on my SCM25 there is a clarity in the mid range that I couldn't find on those other two monitors in this room. So I prefer working on these SCM25s here, this is why I chose them". How is something like this a "very weak argument"?
It's great that someone like you would probably tell this imaginary SCM25 user that: "no, no, you should have chosen the 8361 because it reaches 10Hz further down in the bass, it's more powerful and it's spinorama graph is better because x-y and z..." but what's the point? If on his preference list picking up sibilant sounds quickly is much higher than let's say bass extension than what is the point of this argument?

I know my example above was oversimplified and silly but I hope it helps to make sense of my argument.
Anyway, I'm out for now, I really don't feel like I can add any more useful info to this thread.

 

[goat76]

What people often miss when blindly criticizing a certain speaker brand, saying there are better choices out there, is that mixing and mastering engineer has a result to deliver and does so on a weekly basis. That is what counts in that business, and if the clients wouldn't be happy with result and that could be blamed on the choice of studio speakers, the engineer would not hesitate for a second to swap them for something else.

It's quite silly and short-sighted to sit on an audio forum like this and criticizing the professional audio engineers on their choice of work tools, when these engoneers obviously deliver satisfying results week after week. For some engineers that loudspeaker can be a Neumann, and for another it can be ATC’s.

 

[Curvature]

Where? Quote me.

Fair, but both are very different definitions.
That's why I think it's best to talk about scientific observations, experiments or theories on one side and, I'll precise, the scientific method and more broadly the scientific enterprise on the other.
Saying ''this contradicts science'' or ''science is wrong" is incorrect, and that has nothing to do with truth or fallacy.
Saying this contradicts ''the'' science we have in that domain means something else.

I think we should move to PM or another thread because this is off-topic and I'm fairly sure the exchange will end up a little long.

 

[Curvature]

You say: "The measurements which are available suggest that the speakers are not worth the money and don't live up to their reputation." My guess is that you make this assessment because for you "worth the money" does not include most of my points (from reliability to great customer care etc.).

I think most of the points are unassessable. There are no stats or public attempts for a brand-level study about speaker use, reliability, etc. Likewise, I don't know of any serious way to assess translation. Which for me means that we are pitting reputation against reputation if we aren't interested in measurements, and reputation tends to drift wildly depending on who you talk to, which part of the world you are in, and so on.

Lots of great music has been made using many different speakers. The question is if using speakers that show better measured results would make the artist's (or engineer's) job easier and the listener's time better. I'm no scientist, I'm a civilian. But my earnest effort to read the science has pushed me toward a few simple conclusions about how to understand speakers and away from the reams of hearsay that defined the pro scene for me since I was a kid.

This sort of information counts for a lot: https://www.audiosciencereview.com/...leaked-pricing-info.72465/page-6#post-2669027

 

[Curvature]

GD of KH 310 hits between the limits in yellow. Ca. 0.8 ms at 500-1000 Hz is not really bad result though it is not clearly on the safe side.

Sure that study is limited due to headphones and click-type signal and limited frequency band and impractical method to produce phase distortion as GD peak. The reason was probably to study spectrum of audibility threshold. Spectrum due to IIR XO is a step with possible hump. That is audible down to bass range with close to same limits with 500 Hz. For example, all these are audible with repeated transient signal:

Commercial science most often bears some scent of the manufacturer's product portfolio, ensuring that possible weaknesses are not revealed. For example, hiding behind psychoacoustics filtered with test signals such as soft elevator music or limited frequency range.

What's the problem, exactly? We have a small number of studies about a measurement whose impact is extremely difficult to assess. Sure I would like GD, like THD, as low as possible, but I have no idea what the "sound" of GD is.

Y axis is milliseconds. Neumann data. Graphs are noisy due to extraction.