Originally Posted by AndreYew
You can only determine this if you were present at the original recording session. For example, the Xds may have been more truthful to the actual recording, and the X2s may have been euphonic.
I think it is valid to hone in on the sound of something like a cymbal. I know what a
real cymbal sounds like - and there really isn't that much variation in the sound of
Additionally, I know well from the PHYSICS of how a piece of metal like a cymbal
is going to sound. I know I am going to hear the strike of the drumstick, at which time
the rest of the cymbal is silent because it doesn't know it has been hit yet. The rest
of the cymbal finds that out as the pressure wave propagates out from the hit site.
I know that the cymbal is going to ring before the waves reflect from the edges of the
cymbal. I know after the waves reach the edge and reflect and criss-cross the cymbal;
it has the distinctive "shimmer" sound.
Listen to a real cymbal; and you will hear the dynamics of the cymbal crash.
The Wilson X-2s presented all this information to me in a a life like way - just as one
would expect from a real cymbal. The Xd just flat out didn't do that. It didn't present
the detail - it just "mushed" it all together into a "shhh" sound.
I don't need to be at the recording session to know that the Wilson is not being euphonic.
John tried that argument the first time I posted this. He suggested that what was on the
recording was the mushed "ssshhh" sound - and that the Wilson created all these
transients and microdynamics because it was being inaccurate.
BALONEY. The Wilson speakers are not "smart enough" to take a mushed "sshhh"
and deconvolve from that the microdynamics that one expects from a cymbal.
No - the simple analog crossovers and drivers in the Wilson can't do that. Those
microdynamics HAD to be on the recording.
The DEQX in the XdA is really "smarter" than what is in the Wilson - it's a computer
after all. However, the computer is using its "smarts" in processing the sound in a
way that makes it worse.
The problem, as Keith Howard's article and the graphs of the various filters show; there
is a "conjugate" relationship between temporal and frequency response. If you attempt
to go for maximally flat frequency response - you mess up the temporal response.
Look at the three "all pass" filters - numbers 5, 6, and 7 in Howard's article at:http://www.stereophile.com/reference...ng/index1.html
These 3 filters have the best frequency response of any of the filters - ruler flat as
seen in the third column. However, they also have the rattiest
response of any of the filters.
That's what happens - when you optimize for the best frequency response; you get
inferior temporal response. Conversely, if you optimize for the best temporal response;
you don't get the best frequency response.
The best solution is to seek a compromise. The "best" solution won't be perfect in
either the frequency domain nor the temporal domain.
When someone shows you a curve that shows "perfect" response in one of these
two domains - it just means that they NEGLECTED
the other one.