OK, so I'm a bit late. Doing a little fine tuning, but life has been in the way. The bottom line is at least for my set, the frequency response is a bit smoother above 2K for the LCD-2, but you can see that the transient response of the RP2 is much cleaner, with essentially no output after 0.4ms, while the LCD-2 has an excellent frequency response it has significantly more energy over time across the spectrum.
I may differ from some people on this, but to me I have always found that a system with real DARK space between notes correlates well to fast/clean transient response, and the waterfall clearly shows the RP2 is performing better here (important caveat: my test room is not totally isolated, but most noise occurs below 2K. Dark space is super-audible to me, and the faster the transient response (without being "zingy" and under damped) the better.
I'm trying a few fun experiments on one of my sets to see if I can smooth that roll in the response between 2 and 5KHz, and will wait to publish until I am satisfied that the time and frequency domain are as good as I can get them.
Test comments:
1) I do not have a dummy head or measurement ear, and these curves are not compensated. An uncompensated chart will have a definite tilt down from 1K, with several large peaks, so the real value here is not showing "ruler flat frequency response" but rather to do a rough frequeny and time domain comparison of two really nice phones on the same test bench. As many people know the sound of the LCD-2, this gives a relative context, not a "absolute" measurement like I'd get off a full headphone test rig. I'm looking at how to find a file to apply the HRTF to the raw data. I also have not calibrated my mic, so who knows what the absolute curve would look like. However the TIME DOMAIN response is not affected much by the HRTF or the mic calibration, and serves to show how fast the decay is, and hence how good the "quiet between the notes" is).
2) As the LCD-2 is open-air I can't rule out that at least some of the noise is ambient, however the falloff rather makes me think this is not in fact the case). So I've been working on getting the transient response improved, even if it makes the frequency response a little bumpier...
Teaser:
RP2 O2 v1.2: virtually no output after ~0.4ms
LCD-2 on same rig: significant energy in the system after 0.4 ms, in fact out to about 0.7ms. This is excellent, too. Look at some of Purrin's work, as noted above, the frequency curves may not align, but the time domain should be similar. Most phones are WAY worse than the LCD-2. Again, the LCD-2 is a bit less peaky in response, but not as fast on the attack and decay.