QUOTE(Klyith @ Aug 21 2005, 10:31 PM)
I meant flatness in terms of the frequency response graph, decibels vs frequency. There's probably a better word than "flat" for that, I dunno. Linear maybe. Anyways, reproducing the sound in the most accurate way possible.
For example, here are headphone.com's measurement graphs of the Sennheiser HD 555 (open):

versus the AKG K271S and Ultrasone HFI-550, two of the better sealed cans:

Sorry, but this is not any proof. You are comparing drastically different products. Also, these graphs are not actual frequency response graphs of the headphones, but they are measurements of headphones on a dummy-head, which has a serious effect on the high freuqncy measurements. You do not see the actual behaviour of the headphone itself in these graphs. Also, you do not target a measurable as flat respones as possible with headphones, as you would not be compensating for the required response deviations to sound 'flat'. Just one example: a tailing off of high frequencies at a specific rate is required to compensate for the HRTF functions on a head listening to a non-direct coupled to ears sound source in natural circumstances.
QUOTE
I think that's not true. There are problems with a sealed headphone that will never go away, like the resonance frequency of the enclosed space. The designers can do various tricks to reduce that, but it never will completely disappear.
There is nothing inherantly wrong with a sealed headphone, and resonance of the 'sealed' space is related to the air volume and compliance/driver, just like a sealed speaker enclosure. This is used usually used to advantage to control the low frequency response of a headphone, if it's a true sealed unit. The anomoly, if there was one, would be located in the low frequency spectrum, in the form of a frequency response anomoly. Not a problem for any system if the enclosed volume was specifically chosen to match the characteristics of the drivers; just like when calculating the cabinet volume for a subwoofer. As for resonances equaling the half-wavelength of the spaces in the volume in the rear chamber, this is also no problem whatsover if the headphone was designed properly. The distances are so small, that the resonances related to 1/2 wavelength modes will not become any issue until the treble band. One example of how this can be eliminated by using a solid paper/plastic backer on the back of the driver to prevent any resonances or interaction at high frequencies with the rear volume[basicly preventing high frequencies from entering the rear enclosure], and then drilling a small hole in the back to allow for air pressure to freely pass between rear volume[to allow for use of the volume for low frequency shaping/enhancement]. This is a method used on several headphones. I have measured the response of one such example[MDR-7506] and the method was worked so well, that using any density of acoustic damping material in teh back enclosure made no difference to the mid/hf response. Another example is to use sufficient acoustic dampening materials. It is trivial to absorb/damp the frequencies in question using thin layers common materials such as wool, fiberglass, open cell urethane foam, etc. If you think that having a sealed rear section is a real problem, then please sit back and consider that almost every one of the world's best tweeters have enclosed backs.
QUOTE
I didn't say that good sealed headphones don't exist, just that they are never (or rarely) as good as open designs. They are certainly less common and more expensive than comparable open headphones. Some people prefer closed headphones, which is fine, but I personally wouldn't recommend them to someone without various warnings, particularly on the comfort issue.
Sealed headphones do not appear to the interest of the market, except for professional use.
QUOTE
I also have to laugh at your list of examples. Three can't be bought for love or money, one is just very very expensive ($400), and one that's reasonable ($200) but is not as good as something $50 cheaper, as demonstrated above..
You can not prove 'better' sounding by the graphs above. Headphone.com graphs are not terribly useful[look at the low resolution/high smoothing, the huge dB scale] and I already noted that these measurements are made on dummy heads. These measurements can not be directly interpretted to represent what you might hear[once you enter the upper midrange and treble bands], if you are trying to somehow correlate these with speaker frequency response graphs.
-Chris