That analogy isn't actually as poor as you seem to think...
Yes, it's true that most of us rarely go above 55 mpH (although, to be fair, some of the interstates around here are posted at 70 mpH, and people often drive 80 on them).
However, most people also agree that a sports car with a listed top speed of 120 mpH will usually be a little more responsive going from zero to 60 on a freeway on-ramp than an econo-box that tops out at 65.
This may not be very important to all of us, but it would not be fair to say that "you can't possibly notice a difference" either.
It's also worth noting that most folks have a somewhat skewed idea about what different frequency ranges actually sound like.
What most of us think of as "very high treble frequencies" actually starts around 12 kHz or so....
And anything much above that is more experienced than heard as a sound.
(So most speakers and other components that sound "noticeably dull" probably start rolling off around 12-13 kHz - and a roll-off at 18 kHz is pretty darned subtle.)
However, it often seems to be the case that things that deliver response past 20 kHz often ALSO do a better job of delivering the upper audible frequencies.
Transient response is not just a matter of being able to deliver single sharp transients that have exactly the correct wave shape.
The reason that many tweeters have poor transient response is that they store an excessive amount of energy and release it over time.
This makes them unable to accurately reproduce a short high frequency transient (what they deliver instead is a somewhat dulled transient - followed by a long tail of gradually decreasing ringing).
This is what causes the sort of time domain blur that can make a wire brush cymbal sound like a leaky steam valve.
(In other words, the poor transient response is symptomatic of other issues, some of which are clearly audible in the audible range of frequencies.)
So, to say that in simple terms, while response to 50 kHz isn't necessary for good sound quality...
That ability may be at least suggestive that other problems, which might be audible, also aren't occurring...
(If something can deliver 50 kHz accurately then you probably don't need to worry about it's ability to deliver 15 khz cleanly.)
In the case of something like a DAC.....
One way of looking at things is that "ultrasonic ringing shouldn't be audible because we cannot hear the frequencies at which it occurs"...
But another way of looking at it is that, if you have a lot of ringing, then energy that should have been delivered in a short burst is now being smeared out over a relatively long period of time...
This could be the reason that you can have two DACs - both with excellent frequency response (which is a time-averaged measurement)...
Yet, on one, a wire brush hit on a cymbal sounds like many tiny impacts, but on another it sounds more like a burst of hiss from a leaky steam valve.
Both have the same (correct) overall amount of energy in the proper range of frequencies... yet one delivers it as a series of separate sharp peaks while the other blurs it all together.
This is the exact thing we're talking about when we talk about things like "how much ringing" and "whether there is more ringing before or after the transient".
For example, many people insist that, since natural sounds tend to produce ringing after but not before the event , a filter that has no pre-ringing, even at the cost of increased post-ringing, sounds more natural.
The "post-ringing" added by the DAC is said to be well masked by the post-ringing already present in most naturally occurring transients.. for example plucked guitar strings.. while pre-ringing virtually never occurs in nature.
Yet, for others, whether the ringing occurs before or after seems to make little difference, but the lowest overall amount of ringing produces the more natural sound.
(Perhaps different folks have become trained to recognize different subtle flaws... and to ignore others that are equally subtle.)
I'll reiterate my current summary view on the subject...
I have definitely heard high-resolution recordings that sounded noticeably better than their corresponding "CD quality" versions.
I personally doubt that the difference is there because they are being delivered in a high-resolution format.
However, as a consumer, I have no say in how the music I purchase was produced...
Therefore, to a degree, the reason why one release sounds better than another is somewhat moot to me...
If the re-issue sound audibly better then purchasing it is justified.
And, in most cases, when multiple resolutions are offered for sale, the highest resolution version is "the original", and the others have been created by down-sampling that original.
And, since down-sampling introduces yet another processing step, and another opportunity for coloration to be introduced, I am willing to pay a little extra to get "the file that is one generation closer to the original".
(The fact that the 16/44k version could be audibly as good as the 24/96k version isn't terribly significant to me unless I am assured that it actually is as good.)
(And, once I've purchased the best version, I see little benefit to down-sampling it myself, then confirming that my new version is audibly identical (assuming it is), just to save a few cents worth of disc space.)
Your car is built to be able to break 90MPH....doesn’t mean you’re gonna drive around like that.
Bill
LOL - I speeeet on joor analogy! Eef you're on cruise @ 55, does it matter whether the top end ees 80 or 180?
Ok - My analogy was just as funny as yours...
But if you can't hear 20KHz, I really don't think that having 50KHz information in the recording is going to help you hear the transients...
Yeah - I've read the article too about the super tweeter that everyone thought really opened up the sound, but I'm definitely skeptical.
