[Moonwalker]

Hi!

I read the graphs on Headroom site, and while the FR graph seems pretty accurate, I was surprised to see that PX-200 (closed!) have such a low distortion like HD-650, where PX-100 (open, highly regarded model) has quite high harmonic distortions.
For me, it was more surprising, because I heard almost no distortion when auditing PX-100 but I heard quite a lot of "clamshell" closed-type resonances from PX-200. Could it be the graphs are mistakenly placed or something?

Thanx
Moonwalker

 

[Moonwalker]

So it seems no one can confirm or explain this strange discrepancy between proven sound signature and seemingly wrong distortion charts on Headroom...

Moonwalker

 

[sgrossklass]

Closed cans do tend to measure lower in terms of harmonic distortion indeed (look at the old HD250 II or the HD280). Now "closed can" distortion isn't harmonic (nonlinear) distortion, it's more of a comb filter effect, i.e. it only affects frequency and phase response (linear distortion) - in other words, simple multipath propagation. (This accounts for the characteristic dips. It does seem that the less visible peaks are more audible to us though. Example) We may even be able to hear distinct echoing if the time delay is high enough, or at least a "smearing" of sounds. This may provide a greater degree of spaciousness (e.g. W5000) but can also kill it (closed cheapies) depending on the specifics. (The better closed cans tend to have a damping foam ring around the back of the driver to eliminate the more critical longer paths.)

BTW, in theory it should be possible to eliminate this "closed can" distortion. One would need to extract the reflection related part of the impulse response (not trivial methinks, but maybe someone comes up with a bright idea - MLS based analysis perhaps?) and feed a deconvolver with it through which the actual signal is also routed. OK, deconvolution seems a little computing intensive at this point and is patented at that, but it *should* work. A smarter way would probably be applying the inverse transfer function in frequency domain after FFT, similar to how EQs work.

 

[Moonwalker]

Hi sgrossklass!

Many thanks for the enlightement! Your information proved to be very in-depth and sorry for questioning accuracy of Headroom charts

Moonwalker