Steven Sullivan wrote:
S888Wheel wrote:
From: "Rich.Andrews"
Date: 6/15/2004 3:52 PM Pacific Standard Time
Message-id:
With all of the discussion regarding "tweaks" and "mods" that has been
prevalent, I was wondering not if any of them had any merit, or hold even
then slightest chance of making a difference, but whether or not one could
devise a quantifiable test to prove the claims made. I think it is up to
the person making the claims to prove them.
In the medical field there is anecdote and there is proof. Without proof,
an anecdote is just that, a nice story. An anecdote could also be an
indicator that some effect is happening, but the anecdote by itself
substantiates or proves nothing.
For example, how can one devise a test to prove that XYZ product not just
sounds but also measures "significantly" different than the $0.49 variety
available at Walmart? If we are able to view and manipulate single atoms,
there must be a way to measure and quantify and therefore qualify an
effect claimed.
It is as if we are in the early days of Hi-Fi placing speakers in cabinets
of various sizes until we find something that sounds good. We are trying
all manner of substances without a clue as to what is going on.
As near as I can tell, those making claims of speaker cables,
interconnects, etc are just guessing at what is going on. They don't know
and even if they did, they can't prove it with measurements and tests
using laboratory equipment. There are some theories floating around, but
no one has proposed any experiments to prove these theories. I believe
that if we fully understand a mechanism, then we are able to produce a
better product than all of the guesswork done previously.
This begs the question of how would one go about proving these
unsubstaniated claims.
r
--
Nothing beats the bandwidth of a station wagon filled with DLT tapes.
I think one could start by comparing actual signals. If a given tweak makes
absolutely no measurable difference in the signal then it can't possibly make a
difference in the sound.
What's to prevent someone from claiming, 'you haven't measured the *right thing*'?
Along with the ever-popular 'not everything can be measured'?
And, too, a measurable difference is not necessarily audible.
And I repeat, we cannot be sure that everything can be measured. No
researcher in sound or signal processing could be taken seriously if he
said otherwise, given the advances in measuring properties of the signal
that are made each week and reported in the journals. Now, on the other
hand, if two outputs produce exactly the same signal down to the 96 kHz
sampled bit, then they are indeed "the same". Comparing two digitized
signals can be done simply, just look at their matrices and see whether
the cells all have the same numbers.
But I don't think this is what people mean when they say there is "no
measurable difference," they are usually talking about staring at some
graph or chart or something that has been computed as a property of the
signals. And there is good reason to try this, since the bit-identity of
two signal waveforms is really not at all correlated with two signals
seeming to "sound the same." Drastically different signals can be
easily made which sound the same, because of the variety of effects to
which the ear is not sensitive.
But, alas, once you break away from simply comparing two signals (i.e.
their matrices) to see whether they are in fact the same (not unlike
using Unix 'grep' to compare two text files), you can no longer be
certain of your assertions to the effect that your failure to measure
any difference represents everyone's inability to hear any difference.
-Sean