snip
Thank you for agreeing with me!
I'm not sure what measurement graph you're referring to when you're speaking of the «overshoot». Looking at the Inner Fidelity measurements, there's certainly a massive overshoot at the start of the square wave cycles at 30 Hz and 300 Hz, but that is a common feature of all headphones there. As I see it, Tyll has used «uncompensated» signals for the impulse measurements, which by nature have an upper-midrange hump. Additionally most of the test subjects also show some high-frequency accentuations. Both features will reflect themselves in the square-wave responses as an accentuated initial impulse – within the relatively low-frequency square waves. An overshoot – or any square-wave/impulse shape – always has a connection to the amplitude response, and the 6-kHz peak followed by a still somewhat elevated upper treble will definitely find its corresponding expression in impulse diagrams of any sort.
I don't believe the 6-kHz peak is really intended by Sennheiser. Headphone developers don't have total control over the sound, otherwise someone would have designed the perfect sounding headphone. Most likely they found it to be bearable or even attractive for a large public. If it's really caused by inner reflections between membrane and magnet system (my suspicion), it's hard to avoid in such a ring-membrane design.
snip
When I looked into the earcup and saw those (relatively) large, flat, reflective surfaces, I knew they needed to be addressed. And the driver mounting ring being a flat metal resonator/reflector pointing directly at my ears was also in need of attention.
And I wasn't alone in recognizing these as issues to be addressed.
Many attempts of material choices and configurations later, I had come up with a 'passable' solution dealing with these reflective aspects to the design, but it hadn't 'solved' the primary issue that I heard, namely Listener Fatigue.
I later began calling it
tLFF (the Listener Fatigue Factor) and I had drilled down into it enough to be able to track changes and determine improvements, or not.
And I was able to quantify my results well enough to see correlation between my experiments and the results such that I became 'clued into' what the benefits and net effects were.
This helped me in dialing in what changes to try next and to more quickly determine what truly was
'better', or not.
And this was all well before I began directly addressing and fussing with the overshoot issue.
And the prevailing understanding all along was, and remains, that the 6 to 7KHz 'spike' was the culprit.
And yes I used the leading edge of a 20Hz square wave/step response as the single measurement to address this overshoot issue.
And yes again, I do agree that overshoot is a designed in aspect and has been widely adopted as a desirable performance characteristic by many manufacturers, in that it is the exception to NOT see considerable overshoot on the leading edge of a square wave/step response.
So my experimentation shifted gears and I started finding ways to reduce the amount of overshoot.
It took a couple of years and many thousands of hours of listening, to find the right combination of materials and configurations to significantly reduce the amount of overshoot, AND not mess with the rest of the 800's desirable characteristics.
My latest (and final) Gen-6 results, which have been nicknamed the 800-J-mod, has greatly ameliorated the
tLFF issue, for the better.
And what I have also determined is that the 6 to 7KHz treble peak is NOT a problem that needs to be addressed, as the PRIMARY problem in order to 'solve' the
tLFF issue.
It's the extra added 'unnatural' energy generated at the leading edge of any steep rise time wave form that is the primary source of
tLFF.
It is irritating, especially at elevated listening levels.
Indeed, I have applied 'corrective' EQ (both before and after Gen-6 results were obtained) and it makes very little difference in terms of not just
tLFF but also to the overall SQ that has been achieved.
And to the point I no longer use ANY EQ, except for the extreme bottom end (12Hz to ≈100Hz) which is a mirror reflection of the 800's natural roll off.
And one of the unexpected benefits of lowering the amount of overshoot is an increase in the
LEDR (Leading Edge Dynamic Response), meaning the acoustic sound pressure that is heard and FELT when a percussive impact is played, has taken a decided step up, which is oh so appreciated.
So instead of acoustic sound pressure being generated where it shouldn't be, along with it's dissipation thru time, a more precise waveform is created with a more 'natural' dissipation thru time.
And as I stated in my write up, the results I have obtained thus far could be taken to whole new levels of performance, given the necessary tools and resources, in order to 'nail' the amount of both overshoot and undershoot in order to better 'square off' the response to a square wave signal.
But after I had achieved Gen-6 performance, I don't feel the need to spend any more $$$$ in order to explore this one set of experiments any further, as there are other avenues and experiments that are in need of exploring.
IOW
tLFF has been reduced to the point that it no longer is an issue for me, it no longer 'sticks out' enough that it needs to be fussed with, and along with the unexpected improvements that were also noticed in,
LEDR,
Moar is Less (The
DRC (Dynamic Range Control) goes up but it doesn't get louder, there is just MOAR there),
REALNESS The acoustic presentation becomes all the more REAL,
HMMAIAA (Hearing My Music As If Anew Again, hearing additional nuances and subtleties etc.),
HE (Hear Everything, as in the entire acoustic space, all of the '
voices' (any source of sound), massed instruments with greater individuation etc.),
and as such I have declared 'victory'.
Which in turn, has allowed other projects and experiments to move to the front burner.
JJ