The EQs were done with convolution because it would’ve been way harder to do it manually.
This is the difference between the two headphones on each of our heads.
This part was actually far less interesting than some of the other testing we did, since this was before we got our HRTFs. So really what we did here was just match the in-situ HpTF (headphone transfer function), and we should have that video coming out in the next few weeks.
But if you are EQing properly, that is by ear, then how the bass measures doesn’t matter does it? And my experience is still that there are some headphones that I can’t get the lower bass levels I want. The HD800S being one of them.
Having said that, playing around with the bass on my Hifiman Mini Shangri-La recently does give some evidence that I am wrong on this. What happened is that I realized that the sub bass I am hearing is actually about 10hz higher in frequency than I thought. So in round numbers I was increasing bass around 30hz when I should have been increasing it around 40hz.
I’m saying the in-situ FR in the bass wouldn’t be the same as what the graph shows. Like… we actually discovered this to be a factor even between left and right ears. So yeah you’d have to do it by ear. The only issue there of course is that it’s harder to be precise.
While we can’t sell these in-ear mics, I would consider making more of them. Even without the HRTF to calibrate it, it’s useful to get a sense of how the bass performs. Unfortunately it’s not our design to sell, so we’d probably need to modify that or come up with a new design.
Unironically doable - though one of them being an on-ear might pose a challenge. At the very least they wouldn’t be perceptually identical though because of that. Pinna crush do be an issue.
Perhaps we need to do the equivalent of the Grado tape mod on ourselves. Something that can be stuck in and on the ear to smooth out the bumps and create a better sound path. I wonder if anything like this has been tested. Obviously not sane to try with the middle or inner ear.
But this topic was supposed to be about perception science. Habituation to a particular FQ? Sound pressure level? Perceptual differences with loudness? Time of day? Amount of sleep? Background noise? Or my favorite, training and experience.
I think @generic posted a lot about this, I’ll have to try and search the forum.
My favorite too
My experience with time of day/amount of sleep is that the best listening is done between 2am and an hour or so after getting up for the day - assuming I went to sleep around 10pm the previous night. In other words being well rested is paramount for me. Unfortunately the time of day I WANT to be the best listening is right after work. Spoiler alert: it isn’t a very good time for me to do focused listening.
When I talk about training and experience, it’s that once you start “listening critically” you can’t really stop. I pointed this out years ago with people learning about typography for the first time.
Some don’t know about serifs and sans fonts, about X height and kerning, Don’t know what glyphs or ligatures are. Once you point it out, they can’t stop seeing it.
Agree this is a thing. When I talk about training and experience it’s in relation to neuroplasticity and training your brain to get better at listening critically.
Interesting results. My apologies for asking what is probably another OT question, but I wondered if you had similar difference curves you could post for the HD600 and HD800S measurements on your HBK and GRAS rigs, for comparison?
Not for those units unfortunately. But also we haven’t done this convolution for those two heads. I’m not sure we will since that would require sending those units to me.
But I wouldn’t exactly expect the results to be that different from the bands you see similarity / variation in the above. We haven’t explicitly drawn this parallel yet since this is just four heads and there’s not really a reason to… but notice how the variation in adjustments widens and narrows similarly to the preference bounds? We expect one of the reasons for the preference variation in those regions is down to this HpTF difference across heads, where minimal adjustment is needed in the more consistent bands, and more drastic adjustment would be desirable in the more variable regions.
Fwiw, the difference between these two headphones looks pretty similar when comparing the HD800S and HD600 measurements of different graphers using different rigs. And the graphers’ measurements also look fairly similar to some of your in-situ plots.
There are some differences in the fine grain detail in the treble, which is to be expected. The bass on Caleb’s HD800S in-situ plot appears to be a liitle outside the norm though. I assume that’s due to the HD800S not sealing as well on his head. Some of the variations in the ear-gain region are also interesting.
I think you’ve also observed that the treble gets darker on the HD600 as the pads wear in. And that also cropped up in some of the measurements I looked at by different graphers. The HD800S looks substantially brighter in the treble when compared with HD600’s with worn pads versus with fresh pads. Maybe differences in clamp and pad compression also influenced your in-situ measurements in the treble in a similar way?
Anyway it’s an interesting topic, but maybe one that deserves its own thread.
If you have some idea of the amount of wear on your HD600’s pads though, and when the headphone was made (ie before or after 2020), then I suspect you could probably put together a fairly reasonable EQ curve to emulate the HD800S’s FR using the measurements done on the rigs. And there are a number of graphers who seem to have some fairly decent plots of these two headphones to work with.
Sure but that’s part of the HpTF behavior across all heads. With respect to the other graphers out there, as you say a lot of the differences there include pad wear, but since for us this was all the same unit, pad wear wasn’t a factor - and it was a fresh set of pads.
You could but not for the fine-grained features that are person-specific. The thing to keep in mind is that the FR result is emergent from its relationship to the ear and the ear load, and so placing that on a different head is going to yield various features in the FR that are unique to that head.
Still… I personally believe that doing so might be enough to create the soundstage effects people experience with the HD 800 S. So it’s worth trying.
So, here’s a recent issue I just ran across. Whilst using the ATH-ADX500 with a GR Research OTL headphone amp, I heard an effect that I’ve not noticed before. Listening to a recording of the best of Keely Smith, was able to hear the ambience associated with the recording studio. I’ve listened to that recording countless times, and never have heard this from any headphone before. The micro detail from this headphone is remarkable.
I don’t think adjusting FR by itself would cause this. Thoughts?
Pretty much guaranteed that’s FR. Like even if the headphone weren’t min phase (or broken in some other way), those characteristics come from FR relationships.
Here’s the issue: Don’t think this effect can be duplicated by adjusting the EQ alone. It would seem to me that it has more to do with the driver itself, in the ability to respond to the incoming signal in such a manner as to allow the user to hear both the voice and the slight echo from the studio or the hall (decay and reverb).
Found this explanation about the construction of the ADX5K:
“ The ADX5K packs a huge 58mm dynamic driver, a tungsten-coated diaphragm utilizing what Audio-Technica are calling “Core Mount Technology”. The driver, baffle, and voice coil are all mounted in one integrated unit without any moving parts. This was done to reduce unwanted resonance and create high rigidity for a better transient response. Translation: the driver is designed to resolve forward and rearward movement more quickly and accurately, meaning it can deliver fast detail and decay notes in as realistic a fashion as possible. And I will take Audio-Technica at their word for it, as it is a wholly accurate-sounding implementation. “
Perception of a closed space absent any gear comes from how your ears hear echoing and your brain’s ability to perceive small differences in both time and volume between your ears. Walk into a room and you can feel if it’s a closet, bedroom, or auditorium.
So is this FR If so, perhaps we have different definitions of FR.
I don’t dispute that room treatments can profoundly affect the sound by altering the amount of different frequencies reflected.
So record sounds binaurally in a room and play them back on different headphones. If I can perceive the room size on my estats but not with my HD6xx how does that now boil down to FR and not compliance (I think that’s the right term) ?
Is it the headphone or the interaction of the headphone with an OTL tube amp?
I have a pet theory that one thing tube amps can do through the added harmonic distortion is trick my brain into perceiving room effects that would naturally come from listening to speakers.
Also I have heard from a knowledgeable audiophile that the tube amp harmonic distortion can also cause us to perceive more detail, rather than the expected less. I haven’t read a study on this so take it with a grain of salt but it also wouldn’t surprise me.
How’s that for getting this back on topic? 
Fair point to make. I’ve also noticed this effect with the Topping D900/A900, so it’s not just the OTL tube amp.
Subjectively, it does seem a bit more pronounced with the tube amp, especially as it’s connected to the Canever DAC, which uses Lundhal transformers in the signal chain.
The problem is that what a thing SOUNDS like isn’t synonymous with its cause. If something in the recording is getting enhanced by the playback equipment - as in it’s IN the recording and not an emergent property of the playback equipment (distortion or something being broken), then FR relationships are causing the enhancement. That much we can be confident in.
But again, it might not sound like it’s an FR thing, because you’re not interpreting that in terms of specific frequency bands (yet). To you it sounds like what you described. But once you can interpret it in terms of FR, it becomes clear which bands need to be adjusted to impact that quality of the playback.
