Originally Posted by xnor
Why look at something that tells something else indirectly
if you can look at it directly?
My understanding is that measuring phase is a real PITA, especially for headphones. Again, check the IF links. If this is mistaken, show me someone who is measuring, and publishing, phase response and other information that SWR gets at (even for speakers would be fine - I know such measurements are done for amplifiers, and DtoA converters, and other devices like that). CSD is not an indirect measure - it looks right at resonance and decay; and that's where the SA5000 and T70 argument makes sense; they're both very bright and have a ~10 dB bump around 9k. The difference is the SA5000 doesn't have a few treble ridges up there to compete with the T70. One therefore is just really bright, one is harsh and aggressive and really bright. And I think it would be remiss to declare one of them "better" than the other - there are people who enjoy both equally; is one group "unwashed?" Time domain does matter. Another great example is looking at the "Grado house sound" which relies on resonance. And just because you set this one up for me:
Why look at something that tells something else indirectly (FR) if you can look at it directly (CSD)?
And that's in response to:
"Resonances usually go hand in hand with peaks in the frequency response."
Really, do we have to keep going around and around? If I just tell you that you're the smartest person in the history of the world and that everything you say is the way, truth, and light, will that be enough? [/sarcasm]
Tyll stated in one of the articles I linked (and I won't keep re-linking it because I'm guessing nobody who's arguing with me is clicking anything I link, and is instead just blindly dismissing anything I say) that his measurements are attempting to look at three phenomenon:
Originally Posted by Tyll
Why bother with multiple measurements then? Well, the problem is that the data measured lives in three domains: time, frequency, and amplitude. Any single two axis (two domain) graph will only visually reveal information in the two displayed axis, and will hide information in the remaining domain. <...> For example, frequency response shows information in the frequency versus amplitude domains. Impulse response shows only time versus amplitude domains. I wish I could measure the acoustic phase response because then we could observe phase (time shift) versus frequency. (I currently measure the electrical phase on the impedance plot, which has little to do with acoustic phase. Acoustic phase is very difficult to measure as there's also a time delay in the loop between the driver in the ear that would have to be compensated for each headphone individually.)
Square wave response is a mix between time, and both amplitude and phase. Both frequency response and phase delays will have an effect on the square wave shape.
So, if the information contained in the frequency response measurement is close to the same thing as the square wave response, we should be able to predict what the square wave looks like from the frequency response plot, right? Abso-friggen-lutely!
And this brings me back to what I've been saying thus far (which I don't think anyone disagrees with, until it means that we can't keep posting about it) - you can modify some, but not all, of this because acting on the time domain is incredibly expensive and in some cases impossible without (potentially extensive/impractical) physical modification to the cans in question (for example, I'm sure it's *possible* to modify the Pro2900 to ring less, but I doubt it's *simple*). If you had a can that measured both absolute ruler-flat from 10-100k and produced an absolutely perfect loopback on CSD, with absolutely perfect phase and zero distortion, then yes with appropriate DSP processing you could probably make it sound like a lot of different cans or maybe even a lot of different speakers. But no such device exists, anywhere, for any amount of money. So it seems like a moot point to debate the merits of such an approach.
In terms of real-world cans, they will *always* have some flaws in those domains; the very best products try to minimize them (and often do a good job of it), but they are still not spot-on perfect. So the best you can do with equalization is attempt to address those flaws, not take a crack at transmutation. What's the problem with this reality?
To use an example brought up earlier, the HD 590 was mentioned - I found some measurements (in Japanese) for that can: http://www.geocities.jp/ryumatsuba/hd590.html
It actually looks pretty clean (I should qualify that Ryumatsuba measurements are suspect in terms of accuracy and what kinds of smoothing are applied - if someone can fluently read Japanese and could find out if he/she talks about their measurement methods, that'd be awesome; if you want to see an example of this discrepancy, there's an SA-5000 measurement on that site - compare it to Tyll's measurements of the SA5000. The assumption is that those measurements are relatively, but not absolutely, accurate.). And I'm guessing that with a bit of tweaking you can get it even cleaner sounding, wrt your individual hearing acuity, tastes, etc. But it still has flaws that prevent it from being transmuted into something else. From user reviews, it's supposed to be similar-ish to the HD 580.Edited by obobskivich - 6/10/12 at 9:00am