Blind? You tested two headphones without knowing which is which? How did you get over how they feel different on your head? And did you repeat the trials and perform statistical analysis to rule out chance? Do you have the protocol write up you can present?
Tests were not about specific headphones, but about preferred frequency response. They were conducted blind in three runs, two tests done on two different acoustically-open headphones, and the last on a closed back headphone. I wasn't swapping between the headphones in the same test; each test was one headphone, I just did three different entire runs of all of the following stages, one for each headphone.
Each test begins with the headphone being equalized to the subject's Diffuse Field HRTF as measured at their Ear Entrance point. Following that, the subject is able to adjust the sound for 3 different pieces of music in 4 stages:
- Stage 1: a flat tilt from 20-20000 Hz, adjustable in ±0.5dB steps, can call this "Tilt A"
- Stage 2: a high shelf at 500 Hz aiming to mimic the behavior of Toole's in-room speaker target/the rising DI behavior of typical good monopole speakers, adjustable in ±0.5dB steps, can call this "Tilt B"
- Stage 3: A "contrast" adjustment of the preferred result in Stage 1 with proportional low/high shelves (adding or cutting equal amounts of bass and treble simultaneously), adjustable in ±0.5dB steps, can call this "Tilt + Contrast A"
- Stage 4: The same "contrast" adjustment but affecting the preferred result in Stage 2, can call this "Tilt + Contrast B"
After identifying the preferred adjustment in the above 4 tests for all 3 pieces of music, a MUSHRA test where the subject compares and ranks the results arrived at in the above tests, along with un-adjusted Diffuse Field (low anchor) and Diffuse Field with the approximate Harman's 2018 filter adjustments (hidden reference), ranking them in order of preference for each piece of music.
Following this MUSHRA step, there's another round of MoA where the highest rated choice in the MUSHRA test for each piece of music is allowed to be altered once more with Harman-like adjustments (105 Hz Q: 0.707 low shelf, 2500 Hz Q: 0.4 high shelf, 3000 Hz Q: 2.0 ear-gain peaking filter) to see if this preferred result can be improved further re: the subject's preference.
Finally, there is one more MUSHRA test where the results of the previous MoA stage is tested against the same low anchor, hidden reference, and Stage 1/2/3/4 results from the prior MUSHRA test (to see if the newly adjusted response is actually better than the others).
The final result that is ranked highest for each piece of music in this final MUSHRA test is, of course, the most preferred response able to be achieved in this test for that subject and that piece of music, but we also get a ranking of how all of the included curves are received by the subject.
This whole process is then repeated for two other headphones to see if the differences in acoustic openness/noise floor/good ol' run variation results in drastically different results.
For one, my results were generally very consistent across runs; roughly the same bass-to-treble delta for each piece of music across the three tests was preferred, within a small margin of bass to treble delta... but importantly, the "Harman filtered Diffuse Field" hidden reference always, 100% of the time, scored second to last, only ahead of the low anchor (un-adjusted Diffuse Field). I would call this curve scoring second to last in all 18 blind MUSHRA tests (3 songs, 2 MUSHRA tests for each headphone) statistically significant.
And as completely subjective commentary on why this was the case (that you can feel free to discard):
it was always the bass shelf & how it integrated with the midrange that made it sound obviously wrong/not preferable to me. How it colored transients of bass drums (too much fundamental/early overtone energy) and snare drums (too little fundamental energy to match the tone color of the bass) created a meaningful disconnect in how each instrument was presented, such that they felt like they are given unequal attention/emphasis by the headphone's tuning.
It's totally fair to ask these questions, skepticism is welcome. I hope that explains it sufficiently enough, but feel free to ask further questions if you'd like.
What is uncontroversial according to science, is to dismiss your writings as random statements that cannot be proven. This is especially so in the way you described and justified them.
To be sure, you can think whatever you want about the sound of any headphone. But the moment you ask us to believe it, then in this forum, we ask that you provide reliable proof according to how controlled listening tests are performed.
Considering there are scientific—in this case meaning "evidence-based"—explanations for my writings in this thread, I wouldn't call them random nor would I say they should be so easily dismissed.
To be clear, I'm not the one so often calling them "undynamic" as... well this word is pretty useless for actually explaining things & has no consistent meaning. I'm just using all of the information available to offer a guess (and potentially even an explanation) for what people using this language are
actually referring to with the terms they're using. Considering my main issue is how it colors the core transient elements of the music (drums), I think it's at least somewhat likely these are (some of) the frequency response elements people are actually referring to when speaking erroneously about DCA's designs lacking "punch and slam." (Not "erroneously" in that their subjective experience is wrong/invalid, but rather that using language like this is vastly inferior to using frequency response to explain one's issues with the sound.)
And I am telling you that terms like "slam" are made up, non scientific terms. The moment you resort to it, it shows that you have no use for proper evaluation of sound and just repeating stuff you read in subjectivists reviews.
No disagreement here, but IMO we can only lead people to evidence-based explanations of the terms they're wrongfully using by actually connecting it to evidence in the frequency response... which entails using the word they're using in the explanation so they can understand how incomplete/imprecise the word alone is at actually explaining anything.
Nobody is going to have their mind changed re: using meaningless subjective language without being offered a better alternative, and it would be pretty difficult to explain a term well enough to change minds without saying the term itself, yk?
Dan's observations are just that. His subjective observations. On every new headphone they design, I get pre-production samples and provide feedback. Dan shares what he thinks about their sound but they don't enter my equation in that regard.
This is an interesting piece of information I wasn't aware of. Do you feel that what you describe—what seems like a pretty close working relationship like this—biases you at all when doing your final evaluation? I would assume not, but I know if I were given input on a product before release and I chose to review it, I would always want to disclose that in the body of the review so people could make up their own mind about how that may or may not bias the review. Not to say you're hiding it of course, you were very forthcoming with that information just now, I just don't remember ever seeing you mention it prior in the body of one of your DCA reviews.
My assessment starts with measurements and then, I create filters to change the deviations from target. This method provides much more reliable data than random listening comments you are throwing out.
This doesn't actually provide much data that readers couldn't get from just looking at the measurement vs. the target, and if anything I would actually argue that looking at data from that fixture alone is not sufficient evidentiary basis to bolster subjective impressions.
We know the fixture isn't relevant to the target used, we know Dan's headphones are fairly load-sensitive, and we know that none of us are the measurement fixture ourselves. One could argue trying to connect a subjective impression to a measurement/target methodology with issues such as these would produce high likelihood for error/disconnect.
And as mentioned, my comments aren't random, they're based on blind listening tests.
If I copied what you are doing, everyone is entitled to criticize me just the same. Indeed, I think without methodology, it is impossible to make reliable subjective remarks about a headphone. It is for this reason that you see me silent on comparisons of Core against the others. I simply do not have reliable information to provide. Listening to one headphone, and then later to another, makes for random data, not anything usable.
I'm not asking you to copy what I'm doing, I'm just asking that you hesitate to assume you're the only one with defensible methodology. I get that you and the other reviewers at my firm have had disagreements, but I'm not them and I'd appreciate you not carrying their baggage into making assumptions about me or my methodological rigor. I care deeply about laying threadbare the uselessness of subjective language, and I am actively working towards this goal in a way I feel is pragmatic and sensible.
As said, IMO the best way to unhammer the nail of nonsense in subjective language is not by refusing to engage with it, or by saying "I'm doing it right and you're doing it wrong" because that doesn't actually help or convince anyone... especially when
any of the existing methodologies used by reviewers I'm aware of can be shown to have clear issues. We choose to use a variety of data in hopes it maximizes potential relevance to the consumers of the content, in order to hopefully show what about the sonic event—as described through frequency response—is likely contributing to the effect people are trying to describe with the subjective language.
On a completely different, much more relevant to the thread note: I finally got to hear the Aeon CORE today. Had the same issues with the bass/low-midrange that I usually have, but with a resonance issue I don't usually get on Dan's designs around 5 kHz. Upper midrange was excellent though, probably Dan's best upper midrange tuning for me.