I'd argue that the CSD measurements that Anax brought up are revealing of the differences between the two, but maybe not in the way that we might first interpret them. I don't think any of us would say that we hear the 009 as having a hashy or rough treble. On the contrary, I think most of us would say that it's on the smooth side. But I do agree with a point that Marv brought up in the past, w.r.t the Stax having a certain plastic sheen to the musical texture. Maybe that stat hash (stored energy) is part of what we're hearing there.

Milos: I would consider the idea that maybe your low listening levels prevent you from getting the full scope of all the microdetail in a given recording. Like you said, when you listen at such low levels (I'd guess around 50dB, maybe you could verify this for us), then a lot of that really low level information will be so small it can just get buried beneath ambient room noise and so on. I think a big reason that Marv, Anax, and myself often find ourselves in agreement w.r.t. gear is that we all tend to listen to closer to reference levels (~80dBA). That gives you the ability for realistic macro dynamics (~100dBA peaks) as well as all the rest of the plankton floating around below that.

Looks like there is a versus between HD800 and planar headphones. I'll try to provide my view on that and some ideas.

In my opinion, THD is a part of what hides plankton, it modifies signal, and harmonics of some signal may mask another simultaneous signal. This is the primary cause of lack of microdetail for most phones, but the HD800 has low enough THD not to have this issue. Stats and orthos are still better, but that's just one side of a coin.
CSD, on the other hand, shows for how long an "old" signal is present, and echo of the old signal could obscure the new one. HD800's CSD plot makes it evident as to why they are so detailed - they aren't feeding your ears old signal, they suppress the past well.

Initially I've wanted to say in this post that HD800's ring driver plays a big role in their fast decay, but after doing the math, I see that it isn't so. However, the ring driver still does play a role. Massive ortho and stat drives have a big difference between sound path run-lengths between the center of the driver to ear canal path and edge of the driver to ear canal path - high enough to cause some combing effect. Although, as ear isn't symmetric and sound from different regions across the edge of driver is affected in a different way, there's no obvious combing effect. Still, all that sound from the edges is somewhat old. Ring driver has much lesser difference while retaining high enough surface area.
Also, perhaps, the ring driver can also benefit from a phase plug? It could even really work with an ring driver, better than with an conventional one, as the driver is more or less, huh, uniphase relative to the "focus" of a plug. Would be better though if the diaphragm could move along a conical trajectory, although that's out of the question for dynamic driver. A planar with specifically knurled diaphragm could do that.

THB, I also never understood both the "fast" and "stopping speed" of diaphragms as in stats are perceivably superior to all else.

I tend to think that this mostly means proper air damping and relative lack of correlated resonances. Modern orthos aren't really that heavy in the diaphragm actually, but they have more stored energy in the driver as there are more "pockets" for air to be trapped within.

That said, I don't mean there aren't any resonances in diaphragmatic drivers. Beside fundamental resonance, there's also higher modes, which wouldn't be damped well, as both peaks and lows of the mode would occur on the same side of membrane and air won't be damping them really. The seemingly only way to get rid of that is indeed a stiff diaphragm with elastic edges - somewhat like high-frequency dynamic driver. MrSpeakers' Primes and Ether seem to use something like that. Although the descripion on site reads as if it's throughoutly knurled diaphragm, photos of Ether's diaphragm look more like it's only knurled at the edges, but the photo I've seen isn't good enough to really tell.

After that, we go to earside cup and earpad construction. Here we have pretty much all headphones on one side, and HD800 and Sony's MDR-F1 and MA900 on the other side. All planars and stats have "standard" construction of an flat baffle with driver in it and earpad attached. HD800 has that massive earside with semi-permeable mesh. One thing I can think of is that they've tried to decrease acoustic impedance of HD800 while providing at least some isolation and pressure chamber for bass. Sony went straightforward and essentially made a supersized on-ear headpho ne, fixing bass by using bass lens. Sony's approach should result in lower impedance and it seems to be so as they're apparently praised for the soundstage on par with HD800s while being much simpler in general. Our ears seem to detect acoustic impedance caused upon ears by headphones and limit the sound localisaton distance accordingly. The lower the impedance, the further  away can the perceived sound source be. The higher, the more in-head localisation becomes. IEMs may have perfect direction to sound source, but zero distance. But that's not our topic, albeit it is also a very interesting factor to investigate.

The earside construction of HD800 is non-flat, somewhat permeable and has velour pads - all together that provides an environment much less prone to reflections and standing waves than standard construction with an flat baffle. Most orthos and all stats further exaggerate reflectiveness by using leather pads, which have highly reflectant inner surface, to provide deeper bass. The baffles are often not even covered in anything, and when they are, it's usually thin felt which isn't really as good at absorbing sound as we want. Driver structures in ortho headphones are also non-covered, which is reflectant and bad again. Personally, when I first modded my T50RP I put acoustic foam on the whole baffle around the driver, but I couldn't listen to them for long - it was just painful to my ears, and I haven't heard anything particularly wrong in the sound. I noticed they're painful even with no sound and put foam all over the driver, leaving only the holes in the driver open. That fixed the issue and made sound noticeably better. There is, for example, the Fuzzor mod which does the same, so at least the modding community recognises that issue. Still, would be better if planar drivers were put in enclosures not so prone to reflections.

Back on earpads, IMO, the way out of bass-reflection dilemma is simple -   use a leather pad without any leather on inner side - just open foam/fiber filler. That way they would be even less reflectant than standard velour pads while maintaining seal. Again, I've lined inner side of my T50RP's earpads (beyer gel pads) with acoustic foam, and that increased clarity a bit and made the sound more comfortable. That's subjective though, as I've made the mod so I should be inclined to hear improvements every time I do something. The aforementioned pain thing, though, surely wasn't so subjective.

So it seems quite obvious that HD800s enclosures are designed with proper goals in mind, even if they aren't perfect perhaps, while ortho drivers and stats are put in enclosures that doesn't really do them justice.

Oh, and additionally, I've once read an note about filling an horn loudspeaker with reticulated foam, to reduce reflections within the horn. It could be applied to headphones as well, perhaps cutting out clear paths in foam over driver/ openings in the driver, perhaps in a form of a phase plug (for HD 800). Could further improve CSDs.

Sorry for my English, haven't written anything in it for a long time.

graean,
I don't have much background either.

I'd say that the process of making frequency response linear (doesn't matter objective or subjective, everyone has own goals) is tuning. Damping is one of the methods, particularly it's adding resistance in the air flow path to even response out. Can be uneven indeed

It's hard to define what plankton is, as microdetail as a term is just a thought up word with no fixed meaning behind it. Everyone understands his own. In my opinion, microdetail is just that - low-level sounds that are often obscured by distortions of some kind. I have to say though, resonances don't really take away. They actually add old signal, retain its energy and release  it when the old signal is off, potentially masking low-level sounds. Could be really nullifying them though if the signals are close enough in frequency and erroneous signal is out of phase. Could even exaggerate them. By the way, one of thoughts of what causes brightness on not high-freq-heavy phones is that - phase shifts of different signals might cause them to add up to produce extremely high singular amplitude peaks.

True take-away of sound isn't really a thing. Though, again, I've read once that voice coil heating in dynamic headphone can cause transient sensitivity loss due to magnet heating. There were even some calcs  showing it isn't something unprobable. Such kind of distortion could really take away low-level sounds that occur after loud passages.

Mind you that frequency response isn't the beginning, it's the end result of all the resonances, dampings, other stored energies and dissipators. There is also phase responce, which together with FR can tell you everything about what really happens down there, but its derivatives are easier to understand. And they again can't account for signal-dependent distortions like thermal distortion mentioned above.

HD800s are technically great, many things done for a first time. Ring driver is a great thing. Housing is also a masterpiece compared to other phones, even if not 100% for the result, but at least for the goal. It adressed acoustic impedance and helped tame earside reflections by the way. Of course, they still can be made better - even just taking reflections in deeper consideration, like with Anax mod. Mass manufacturer also might not be really able to adress these issues due to requirements of mass production, quality control and perhaps not good enough repeatability of materials required.

Lastly, I wanted to say not about damping, it's a well known and huge topic, I am only adressing the importance of fighting reflections on the earside, which has just began getting traction. You don't want to affect damping with it, it's another thing. The mods you've mentioned are about that too, as I understand. You've mentioned damping some spaces and not damping others, true, damping might leave some parts of the driver undamped - a bass lens would be an example. But damping should provide at least some restriction to direct sound flow, that's what it's about. If damping isn't in the way of sound directly radiated from driver, it's not damping.

Oh, sorry for the harsh-looking post, I didn't mean that. I just want to be settled on terminology. Talking with different understanding of termins would be like talking in mutually unknown languages.

Though, again, I've read once that voice coil heating in dynamic headphone can cause transient sensitivity loss due to magnet heating. There were even some calcs  showing it isn't something unprobable. Such kind of distortion could really take away low-level sounds that occur after loud passages.

Lets say we play an HD800 really loud, say 105dB AVERAGE SPL.
That takes 5mW of power. 
Lets assume a thermal resistance of free air. (100^oC/W)
The voicecoil would heat up about 0.5^oC (0.9^oF).

Assuming 120dB AVERAGE SPL. = 160mW would increase the voicecoil temperature by 16^oC (29^oF)

Personally I would not be afraid of heating up the relatively very large magnet to levels where the magnetic fieldstrength would be compromised without destroying my eardrums or the voicecoils (if I had an HD800) in the process ... 

Assuming 120dB AVERAGE SPL. = 160mW would increase the voicecoil temperature by 16^oC (29^oF)

Personally I would not be afraid of heating up the relatively very large magnet to levels where the magnetic fieldstrength would be compromised without destroying my eardrums or the voicecoils (if I had an HD800) in the process ... 

Curie point of ~300 deg C  is needed for permanent neodymium magnet to lose domain arrangements. I have seen a speaker transducer catch a fire, but I think it's magnet was ok afterwards.
Perhaps at some temperatures closing 200 deg C the lacquer around coil wire would melt away, causing shorts?
It got me thinking when probing an amp build that had 8V dc.