I tend to avoid that place. However I just posted roughly the same reply in the Ultrasone Club thread, here:
I also know that there was, at some point, an Impedance Tutorial in Sound Science that I contributed not insubstantially to at one point. It wasn't "dead to rights" accurate, but it was a lot better than a lot of the nonsense that crops up from time to time.
Anyways, it seems that German models in general seem to fit into that trope (Sennheiser, Beyerdynamic, and to a much lesser extent, Ultrasone). The HD 598s have actually been measured, here:
http://www.innerfidelity.com/images/SennheiserHD598.pdf (posted so there's no need to guess)
Compare this to a stable headphone:
Regarding the Sony cans - Sony headphones (like many other Japanese models) tend to be pretty stable*, but the MDR-F1 and MDR-MA lineup are explicitly so, because they have a Zobel network included for each driver; the goal being to flatten out the impedance. The downside (and this isn't explicitly a downside to Zobel networks) is apparently a very low nominal Z - the F1 and the MA900 are both 12ohms nominal, which presents an issue for a lot of equipment.
Basically with the HD 598 (or something like the HD 580, which I used to own) if you hook them up to a low Zsource, they can sound "thin" or "bass light" and with a high Zsource they will gain some body (at least imho - others might view it as too much bass). Personally I enjoyed the HD 580 on the 470R taps of my Yamaha more than anything else. Yes you could get the same result with a graphic or parametic EQ. But in the case of something like the M50, it really will not matter because they aren't reactive relative to Zsource. My suspicion is that a lot of people who subscribe to "all amps are different" are basing this on experience with a Sennheiser or Beyerdynamic headphone and trying out amplifiers with wildly varying Zsource figures.
* Some various Sony cans' measurements:
http://www.innerfidelity.com/images/SonyMDRV6.pdf (here's an outlier)
The MDR-F1 have actually been measured before (http://www.geocities.jp/ryumatsuba/mdr-f1.html), but the reliability of that suite is somewhat suspect (there is A LOT of post-processing/smoothing going on with those measurements), and there is no impedance data. Also it's in Japanese.
To the topic of the thread: right now I am listening to Ultrasone PRO2900s. And enjoying it.
Just noticed in your signature, you list the HD 595, not the 598 - apologies. Here's the 595:
http://www.innerfidelity.com/images/SennheiserHD595.pdf (a lot of the same but not identical)
I've pretty much come to the same conclusions. Japanese headphones's in general are pretty nonreactive. I wonder how that happened... Also, the Zobel network thing is pretty cool. Perhaps an electronics build is in order for me...