Phil wrote:
Patrick Turner wrote:
[snip]
About the same kind of things get said about Auricaps, that after fitting
them they "burn in" and
begin to increase the quality of the sound in subjective ways.
I replaced regular Wima polyprops in a client's preamp recently and did one
channel before the other.
We had an AB test with various music recordings in mono to find if he could
choose which channel had the
Auricaps and which one had the old Wimas ( which presumably were "burned
in" ).
He could not identify better than 50% of the time or say which channel was
better, different, or had the Auricaps or Wimas.
I switched between one channel and the other, asking him to identify where
the "better" caps were or not,
but he just couldn't pick it out accurately. I smiled when he said "that's
definately the better one with the auricaps"
after switching from Auricaps to Wimas.
I heard no changes, it was all fine wine to me.
There would have been no point in measuring the thd of the amps with
different caps because the
thd caused by good quality plastic film coupling caps in high impedance
circuits between
tubes is almost unmeasurable.
Remember, Patrick, the A-B double blind test doesn't work in audio,
because it fails to achieve its primary purpose, namely the "isolation
of a variable." When you switch back and forth, you test two variables,
not one, the capacitor, and the ability of the mind to remember audio
passages. The Boulder amplifier company noted in a letter to the Audio
Critic that when they changed a cap WITHOUT TELLING ANYONE, they would
get tons of letters saying "What did you do, this is great?!", and yet
they knew from actual trials that in AB comparisons, no one could
distinguish between the old and new caps. The letters proved that (1)
there was indeed an audible difference, and (2) that their AB trials --
some of them even double-blind -- were incapable of detecting this
difference. The reason is obvious: By switching back and forth, you test
not only the component, but also the mind's ability to "fill in the
gaps," to remember, when hearing the old cap, the information that was
learned when listening to the new cap.
Perhaps by playing a section of music, say, 6 times in a row, and
switching to the new cap at some point unknown to the listener (after
the 2nd, OR 3rd, OR 4th, etc. time), you will find that the listener can
identify when the switch was made, BUT, you can only play a given piece
of music ONCE. Who knows? Perhaps by performing a test that, unlike the
standard implementation of the double-blind test, you will be able to
identify those componenets that truly are superior -- at least, in
combination (use, as some have suggested, a Teflon first, a Mundorf
last) -- and then make an amp that is far superior to anything you have
made so far. However, note that when a change is indeed detectable, and
initially preferable -- perhaps because it provides information unheard
before -- that over the long term, a change from a good component to a
bad component (or circuit) will result in a less satisfying system,
which is not listened to or enjoyed as much, again, even though the
change to that component or circuit INITIALLY seemed to be an
improvement. Perhaps testing listener's reactions when going both from
cap A to cap B, AND from cap B to cap A (using different selections for
each test, of course, so that you don't wind up accidently, but also
stupidly, SHIFTING from a test of the cap, to a test of the mind's
ability to "fill in the gaps"), will help to identify the "better" changes.
Phil
Thanks for your contribution about AB tests.
AB tests are all I have and AB tests do indicate many listeners cannot discern
which capacitors are being used in an amplifier.
Its as simple as that, and further more I have changed from polypropylene to
polyester
or vice versa nobody has heard any difference to the sound.
Next time you build a speaker crossover, change from a 5 uF bipolar electro
to a plastic film cap and notice the complete lack of sound quality change.
Patrick, did you actually THINK about what I said? If I plug your ears
with wax, and then test to see if you hear a difference between amps or
speakers, would you consider that to be a valid test?
Of course not.
I am not objecting
to your using AB-tests, I am pointing out to you that using a TYPE of
AB-test that causes, actually forces, the test to shift from being an
accurate test of the actual differences between two components (this
assumes that you have two components that really do have significant
differences), to the ability of a listener's mind to "fill in the gaps"
when a less revealing component is substituted, is stupid.
I tested Auricaps A-B to see if any detectable difference could be detected at all.
None was detected better than chance.
The two of us at the test wanted to know if the Auricaps were detectably " better
sounding"
than Wimas.
I heard no changes whatsoever, and I don't care if the makers of Wima and Auricap
and their groupies think I am cloth eared.
The other guy was asked simply which he liked better about 10 times with a range of
repeated music tracks
and I kept a tally of what he thought sounded better. I was the only one making the
switch from
the Wimas to Auricaps, and recording his reaction to the different capacitors.
It was 50-50 score for Wima and Auricaps.
There was no attempt to force anyone to do anything or to
proove that A-B tests become a test of humans rather than the gear being compared.
Of course it IS a test of humans. Humans go around saying A is audibly better than B.
I ask then to sit in the hot seat and have a listen to A and then B, without
knowing whether it is A or B.
Sometimes they will say there's a big difference when you have not changed from A to B.
AB tests make fools of so many people and some are in the high places of
audiophool-dom,
and some are miffed that anyone may challenge their private theories about what sounds
better,
and proove they have no clue really.
I've seen many a dude create arguments to denigrate simple A-B tests to wheedle out of
having
to put up or shut up.
Where some differences are found, its often a case of its all good wine,
and a cabernet sauvignon is just different to a shiraz, not necessarily better or
worse.
Ignoring Phil Allison's useless, very old debating technique of throwing
out a bunch of general criticisms without providing any supporting
examples or facts (apparently he can say "that's stupid" and it
mystically comes true! I wish he would mystically declare that I'm rich
...), look again at what the makers of Boulder amps said. They found the
issue of the STANDARD AB-test confusing, because on the one hand they
got your results -- caps A and B could not be distinguished in a normal
AB test -- and yet another type of double-blind test, which
automatically occured when they changed from using cap A to cap B in
their amps without telling anyone, frequently resulted in a ton of
letters saying how much the amp's sound had changed.
I have changed amps surrepticiously from an 85-85 class A tube amp to 2 x 300W SS
with a group of 4 audio guys one evening here and NONE noticed any change.
Both amps sounded fine, maybe because they were made by the same guy with the same eye
for
detail about what was important to get right.
Scientifically speaking, using the standard rules of logic, it
NECESSARILY FOLLOWS from the experience of the people at Boulder that
the standard AB test, which switches back and forth between two
components when listening to a single piece of music, has a low level of
resolution, period.
Well yes, unless large differences in the gear are present.
Now, I THINK, given that the normal version of the
AB test does allow the listener to memorize the music -- and this next
conclusion is NOT something that absolutely MUST be true from the rules
of logic -- that the reason for this drop in resolution is because the
standard AB test actually winds up testing two variables, not the one it
was designed for, namely the actual differences between components, AND
the abilty of the mind to "hear" what it has remembered from previous
runs, thereby overwhelming the actual data provided to the mind from the
ear. In other words, we cannot ASSUME that the mind hears only the data
stream that the ear is sending to it WHEN the mind has repeatedly heard
both the current data stream (cap A) and a very similar data stream
(cap B) before. Does anyone have an INTELLIGENT, scientifically HONEST
and VALID objection to that statement?
AB tests are FLAMBOOZLING TO MANY PEOPLE.
OK, so darn what?
My proposed solution is simply an attempt, and I'm not certain that it
will work, to greatly improve the resolution of the AB test by removing
this second variable, by removing the ability of the mind to overide the
data actually presented to the mind by the ear. This isn't a mindless
"components do too sound different!" rant, it's an honest, serious
attempt to provide an intellectually valid assessment of the needs and
requirements of accurate testing. If AB tests that really do fully test
the real differences between components show that no differences between
film caps exist, then I will say that those of us who claim to hear such
differences are indeed stupid. I won't be happy about it, but I WILL say it.
Why don't you (and that uselss idiot, Allison) ask yourself a couple of
simple questions? One, assume that there really are no differences
between caps A and B (hell, just use the same cap in all six runs!);
would my proposed test consistently show that differences do exist? I
say no. Do you agree? Two, given two caps that really do have
significant differences, will my proposed test have at least some chance
of reliably detecting that difference? I say yes, although at this
point, we don't know for certain whether it will have more, less, or the
same sensitivity as the standard AB test. However, IF it turns out that
my test does indeed have much greater resolving power than the standard
AB test -- due to heretofore unrecognized defects in the current
implementation of the standard AB test -- then wouldn't ANYONE designing
audio products want to have access to those superior test results?
Wouldn't you? Unless someone can prove, using the rules of logic, that
my test is inferior to the standard AB test, we really won't know for
sure until someone tries it.
Gee Zuss!
Such logic I ain't ever had to wade through before.
I just ask Joe, which do you like better, A or B?, and switch a few times
on the same bit of music.
Joe anwers to the question about 10 times where I try deliberately to trick him,
and if the results proove he did no better than guess, then not much difference exists.
Sometimes its a chalk and cheese process, say you change between different 6CG7 brands
in a line stage without NFB. Inexplicable sonic changes are heard, thd is below 0.02%,
so distortion isn't the reason, all tubes measure about the same......
I've witnessed 4 guys all agreeing that tube A was far inferior to tube B,
where B was NOS Siemans 6CG7, and A was a new Sovtek 6CG7.
Maybe the Sovtek would mature like a fine wine with some age.
Who knows? repeating the test months later after ageing the poor tube in a test circuit
might improve the sound.
Some AB tests raise more questions than can be answered, but the real audiophile is
nervous and even neurotic about component choice, and is always worrying about what
choice sounds better.
He does constant AB tests, and he worries how time is affecting his gear, and the real
wear on
his system is due to so many changes. They don't just set it up and listen for 20
years.
They are always thinking it could be better.
Well sometimes their tweaking leads to a downgrading unless thay have two systems which
they
compare so to be able to leap frog the quality of one over the other.
That was the method I chose to build fine sounding speakers.
Patrick Turner.
Phil
The sound is dertermined by factors of recording standards, room quality,
speakers, source, amplifiers etc, and all the many variables within those
categories
People like to think they hear a difference between capacitors and cables but
when challenged to proove they can detect any difference at all they routinely
fail
to hear anything better than 50% chance would predict.
But where directed to instal whatever capacitors make people happy, then I
always just instal them.
The customer is king.
Customers are not dumb, and generally can discern which system they
may prefer.
Unfortunately, I cannot give them the opportunity to auditon 20 different
output transformers and that number of different amp topologies.
I can only build amps with low noise, low thd, low Rout and wide bandwidth.
I mostly generate sales because what I make sounds better in a total way to what
they were used to.
Many amp makers simply rely on getting the numbers right first.
Tweaking around at the edges with capacitor brands and cables appears to actually
do
an extremely small amount of improvement to the sound.
Many amp makers, perhaps myself included make such fine sounding gear that
usually the changing of caps or cables is simply unecessary.
I have no doubt that using the most expensive caps or cables constitutes
very good engineering practice, and I go right along with anyone who says
that brand X is darn good stuff.
But should they allege brand X is better than brand Y, then sometimes I must
humbly request
proof of their allegations. Where they refuse such proof, and cite all the perils
of AB testing
and its spurious nature, then so be it, I leave them with their refusals and
obfuscations
and proceed to build amps without the benefit of their wisdoms, but with whatever
parts they request,
but otherwise to my own standards which work well and with complexities that are
quite incomprehensible to the vast majority of the most seriously minded
audiophiles.
Its easy to make a triode amp that sounds blameless, regardless of what caps or
cables are used.
Patrick Turner.
"Burn in" when used about capacitors is another phrase which means
that an owner has become used to the pain and expense of changeing coupling
caps, and has
mentally accepted them, like getting used to a new maid, even though she
doesn't work any
better than the old maid.
The same things are said and done with cables and line conditioners and a
few other absurd things
that ppl like to suggest make a difference.
Audiophiles like to credit themselves with super-abilities that us ordinary
mortals do not possess.
They place their hearing abilities above everyone else. In some cases, a
few DO have
better hearing, they understand words of opera and songs better than myself
or anyone else and can hum along
in perfect pitch, and their brain is obviously better connected to the
music than most of
the rest of the population. They even understand the purposes of the
emotional intentions expressed
in the music by the composer. But alas in simple AB tests where they are
asked to identify equipment changes
they don't fare much better than anyone else with supposedly normal
hearing.
I have no problem challenging their sacred cow beliefs so to find truth and
thus be better
equipped to build fine systems for such people.
I have always enjoyed my above clients system with or without his Auricaps.
He has an excellent turntable with hi-end MC which renders vinyl better
than most SACD, CD DVD etc,
and nevertheless has excellent silver disk replay systems.
He is very happy and that's what counts. His system is one of very few I
like listening to
all night, and I would rate it better than my own, but mainly because the
source
devices, ie, silver disc and black disc devices are better than anyone
else's.
Maybe the caps make a difference, but I really doubt it, it could be blind
insurance,
"put in these acclaimed caps just in case.."
Acclaimed by who? where does the BS start? with a bribe of someone in an
audio magazine
to give a good review?
Usually when the guys I know here start raving about capacitor sound they
do so
WITHOUT A REFERENCE SYSTEM FOR COMPARISONS.
So they just change something such as caps without comparing to a second
reference system
and say they hear better sound.
But are they? What is the reality?
I give them full permission to fool themselves, but none to fool me.
I have two complete systems in my lounge, and many comparisons have been
made.
The change from having crossover points in a 3 way speaker system from 1kHz
and 5 kHz to
250Hz and 3 kHz was astounding once I learnt about LCR theory, applied it,
and built
a decent measurement system.
Later when i changed from cheap asian made speaker drivers to SEAS drivers
there was another revelation, and I realised I'd wasted a lot of time on
crap made in China.
The tighter magnetic gap tolerances and cone materials in Norwegian
speakers
resulted in clearer and more precise accurtate music.
Caps and cables have made imperceptible differences.
I tried Allen Wright foil types of interconnects, twisted pairs, no change.
Maybe the system was already as good as it was gonna get.
I hear little difference between good class A tube amps and my class A
mosfet
amps.
I thought I could here something different until I did the careful AB test,
and asked
a couple of other folks around.....
Some SS amps are disgraceful though.....no doubt about that.
Patrick Turner.
No one showed me real measurements 
I wouldn't care if they are
cleverly disguised inner chewing gum wrappers rolled into capacitor
looking packages. I like how they sound.
And, I know someone who has an Audio Precision system in his home. He
gives me lots of good advice. He never mentions his toy gives him the
insight of God Almighty. Perhaps he is too wise for such silly
posturing
Changing anything in my Audio Dungeon may affect my perception of
Music. I may notice 8^0
Changing nothing in my Audio Dungeon might be good for earnest
evaluations of components, but I Ching is a rascally Coyote ...
I enjoy Music through cheap SS amps and cheap speakers, too. I do not
think they sound the same as my fancy stuff, but, Music hath charms to
soothe the savage technoid.
I enjoy the Quest. I enjoy the comradery. I suffer the neo-wizards
lightly. Forgive them, Father, they just like to talk, too ... just
like me.
Happy Ears!
Al