On Wed, 5 Jan 2005 22:38:03 +0200, "Iain M Churches"
wrote:


"Jim Lesurf" wrote in message
...
In article , Iain M Churches
wrote:

"Don Pearce" wrote in message
...

First the capacitors. Go through many samples with a meter to make
sure they are as far as possible equal in value. Install them in a
switcher box so they can be changed without delay.

OK. The service technician at the studio has promised to select with a
bridge carefully matched samples from those we give him.

I would wish to know the details of the measurement. Frequencies used (if
sinewave), distortion levels, signal levels, measurement system
calibration, etc. Ideally, I'd wish to know values measured across the
audio band, and get a match to a tight level.

Just a moment. Not long ago, the general opinion seemed to be that we
would not be able to tell the difference when changing between two
caps of the same value and voltage working from different makers.
Now we are being asked to match them to a tight level :-))

Yes, to ensure that they are *exactly* the same value. If this proves
to be impossible, then you should run the tests twice, once with the
Jensen being the slightly larger value, and once with the Jensen being
the slightly smaller value.

If you are using a single-frequency cap bridge, you may need to consider
the effects of series ESR, etc. Given the points made elsewhere about
the dc you may also find it useful to measure with dc applied, and perhaps
even before/after use to see if this has had some effect!

The original claim was that a listening panel would be able to
differentiate between an industrial/commercial grade coupling
capacitor andJensen capacitor.

Afraid I am quasi-paranoid about such things as I've spent 20+ years
working on precision measurement systems for people like the NPL 8-

It shows:-) But I am sure it is a good thing, and that we can all sleep
more
safely in our beds:-)

I believe that Jim is just covering the bases, I myself would be happy
with a simple check of value at 100Hz and 10kHz.

I think I'd like to know things like the frequency and phase response with
each cap in circuit, again with a fair degree of precision.

Those are surely things that will have to be determined afterwards, to find
out why there is a difference in sound, if one can be established.

Agreed, there's little point in intensive investigation *before*
you've observed an audible difference.

We could use a random numbers generator to pick from three numbers, and
feed these to the desk automation.

Erm... how are you planning to do the switching?

As mentioned earlier, The source material and the console both run to time code
and so we can program the changes to TC. This way we will be able to repeat the
experiment with great accuracy. It has been suggested that we should use odd
bar counts (not many people will expect this) in multiples of fives, sevens and
elevens, and switch within the bar:-)

I'm not sure about this TC business, it seems to introduce an
additional stress, and has nothing to do with identifying the Jensen
per se.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

"Stewart Pinkerton" wrote in message
...
On Wed, 5 Jan 2005 22:38:03 +0200, "Iain M Churches"
wrote:


"Jim Lesurf" wrote in message
...
In article , Iain M Churches
wrote:

"Don Pearce" wrote in message
...

First the capacitors. Go through many samples with a meter to make
sure they are as far as possible equal in value. Install them in a
switcher box so they can be changed without delay.

OK. The service technician at the studio has promised to select with a
bridge carefully matched samples from those we give him.

I would wish to know the details of the measurement. Frequencies used
(if
sinewave), distortion levels, signal levels, measurement system
calibration, etc. Ideally, I'd wish to know values measured across the
audio band, and get a match to a tight level.

Just a moment. Not long ago, the general opinion seemed to be that we
would not be able to tell the difference when changing between two
caps of the same value and voltage working from different makers.
Now we are being asked to match them to a tight level :-))

Yes, to ensure that they are *exactly* the same value. If this proves
to be impossible, then you should run the tests twice, once with the
Jensen being the slightly larger value, and once with the Jensen being
the slightly smaller value.

Yes OK. They will be matched to be exactly the same value, with the Vwkg
as stated on the component.

I believe that Jim is just covering the bases, I myself would be happy
with a simple check of value at 100Hz and 10kHz.

OK. That's how it will be done:-)

I'm not sure about this TC business, it seems to introduce an
additional stress, and has nothing to do with identifying the Jensen
per se.

Nothing extra for the listener. And locking the sound source and the
console together, and switching precisely to within a subframe of
the TC will allow us to reproduce the same test over and over
without the slightest discrepancy.
This will in fact reduce the stress for the operator.

Thanks for your useful comments.
Iain

On Wed, 5 Jan 2005 22:44:31 +0200, "Iain M Churches"
wrote:


"Jim Lesurf" wrote in message
...

FWIW I am not really interested in "how long people take", but in their
ability to just detect (or not!).

Agreed. Though one would expect the change to register within
a bar or two.

Hence introducing time as a factor is one
I would personally avoid as I fear it complicates the real issue.

There is no time factor as such, except that it will be interesting to see
which member of the panel has the sharpest perception:-)

However, provided there is no time pattern which the listener can predict
or deduce, your protocol seems OK.

We like the idea of the panel just having to listen, and press a cue button
if and when they hear a change. Nothing else.

The problem you will have with this approach is that of interpretation
of results - how close constitutes a hit? Correlation is going to be a
real bitch, and I think you are going to wind up not knowing whether
the effect is audible or not. Far better to present discrete chunks of
music with defined start and stop points, and simply ask the ABX
question at the end of each.

d

Pearce Consulting
http://www.pearce.uk.com

In article , Iain M Churches
wrote:

"Jim Lesurf" wrote in message
...
In article , Iain M Churches
wrote:


I would wish to know the details of the measurement. Frequencies used
(if sinewave), distortion levels, signal levels, measurement system
calibration, etc. Ideally, I'd wish to know values measured across the
audio band, and get a match to a tight level.

Just a moment. Not long ago, the general opinion seemed to be that we
would not be able to tell the difference when changing between two caps
of the same value and voltage working from different makers. Now we are
being asked to match them to a tight level :-))

My concern is not with "the general opinion". It is with being able to
carry out a test whose results can then be regarded as having some validity
and reliability. If you don't ensure the values are closely matched then
you have no evidence that a slight change in value is either causing an
audible change or hiding one. So if you concern is to test audibility of
"type" of cap (i.e. chosen dielectric, method of construction, etc) then
you have to control/match those values which might confuse or confound the
test.

Bear in mind that some people claim that effects like the tan delta are a
'cause' here. (IIRC your Jensen rep has said this.) If so, then very
careful matching is implied as if this *is* the cause it implies human
sensitivity to very small differences.



If you are using a single-frequency cap bridge, you may need to
consider the effects of series ESR, etc. Given the points made
elsewhere about the dc you may also find it useful to measure with dc
applied, and perhaps even before/after use to see if this has had some
effect!

The original claim was that a listening panel would be able to
differentiate between an industrial/commercial grade coupling capacitor
andJensen capacitor.

Yes. The snag is that such a claim may be trivial to support if the caps
differ markedly in value. Discovering that different values can produce
audible changes would hardly be a startling result. :-)


I think I'd like to know things like the frequency and phase response
with each cap in circuit, again with a fair degree of precision.

Those are surely things that will have to be determined afterwards, to
find out why there is a difference in sound, if one can be established.


The problem with doing that "after the event" is that people can then
decide whether they think the match was "good enough" already knowing your
test results. I think you need to establish in advance that those
interested all agree that the level of match is close enough to ensure that
a difference in gain, response, etc, is not the 'cause' (or masking some
other effect).



We could use a random numbers generator to pick from three numbers,
and feed these to the desk automation.

Erm... how are you planning to do the switching?

As mentioned earlier, The source material and the console both run to
time code and so we can program the changes to TC. This way we will be
able to repeat the experiment with great accuracy. It has been
suggested that we should use odd bar counts (not many people will
expect this) in multiples of fives, sevens and elevens, and switch
within the bar:-)

Sorry, but you have not said the physical switch method or how it will be
operated. I was not asking about the protocol but the physical circuitry.

Well you be using mechanical switching, electronic, what type, etc?

TBH Iain if you have not already done so it may be useful if you have a
look at previous 'ABX' and 'double blind' studies that have been carried
out for things like this. The point is that many of these end up being
criticised for various (alleged) defects which then allow people who do not
agree with the results to dismiss them. If you do not check this you may
simply be wasting your time as people will reject your results.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

In article , Iain M Churches
wrote:

"Jim Lesurf" wrote in message
...

FWIW I am not really interested in "how long people take", but in
their ability to just detect (or not!).

Agreed. Though one would expect the change to register within a bar or
two.

The problem here is that your expectation may be unjustified so you are
implicitly adding an extra 'expectation'. I would have said that the time
would vary wildly from one person to another, from one passage of music to
another, etc. Unless the change was stunning obvious I would personally
spend some time listening before coming to a decision. I would not
personally feel I had to react in a bar or two. Others may not react as I
do.

Hence introducing time as a factor is one I would personally avoid as
I fear it complicates the real issue.

There is no time factor as such, except that it will be interesting to
see which member of the panel has the sharpest perception:-)

The problem here is that you are using "sharp" to mean "quick". I am more
interested in the possibility that some people may detect much smaller
changes than others, but may well take longer to do so.



However, provided there is no time pattern which the listener can
predict or deduce, your protocol seems OK.

We like the idea of the panel just having to listen, and press a cue
button if and when they hear a change. Nothing else.

I'm afraid that I am not included in your "we" at this point. Not sure who
is. Do you mean that those who will take part in the test wish to perform
it in that manner?

My concern is less that it should be performed as the listeners wish than
that its results should be robust against later criticism. Otherwise you
may have an ejoyable time, but not provide useful results. :-/


I would also wish to have a lot of measured data on the performance of
the system to establish the level of any 'uncontrolled' effects which
may influence the results. Ideally, this would be in advance so that
any 'contentious' points could be sorted out before actual listening
tests.

Does that mean you would like to see the console schematic?

Well, if you wish me to be able to make detailed comments on my views of
the reliability and applicability of your results, then yes, I think I
would wish to have a lot of detailed info of the kind I have been
discussing about the physical arrangements, circuitry, protocol, etc. I am
concerned with all sorts of possible 'stray' or unanticipated interactions.

FWIW with *all* measurement systems my main concern is "what is wrong with
this measurement system that will limit the accuracy or reliability of the
results?" Such limits always exist, but when human perception is involved
the problems are much more difficult than normal lab-work . My experience
is that time after time people do ABX or DB tests, but don't establish
these limitations clearly by nailing down the details. Thus risking not
getting satisfactory results.

Bear in mind that *whatever* results you get, someone will find it does not
fit their views. :-)

As an academic I was used to the idea that you make research proposals or
publish results, and then everyone else sets out to criticise your work and
point out your failings. Part of the process. The real challenge is to do
work that stands up well to such challege. You won't convince everyone, but
the better-founded your method, etc, the more people you will satisfy when
they find you have 'done your homework'. :-)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

In article , Stewart
Pinkerton
wrote:
On Wed, 5 Jan 2005 22:38:03 +0200, "Iain M Churches"
wrote:


Just a moment. Not long ago, the general opinion seemed to be that we
would not be able to tell the difference when changing between two caps
of the same value and voltage working from different makers. Now we are
being asked to match them to a tight level :-))

Yes, to ensure that they are *exactly* the same value. If this proves to
be impossible, then you should run the tests twice, once with the Jensen
being the slightly larger value, and once with the Jensen being the
slightly smaller value.

That is an interesting idea. However it may fall foul of the test being for
"different" rather than "better". i.e. a difference either way may be
noticable.

I would hope, though, that we can choose values which are similar enough so
that people can agree in advance that the caps are - in this respect - ones
we can regard as 'identical'.

If you are using a single-frequency cap bridge, you may need to
consider the effects of series ESR, etc. Given the points made
elsewhere about the dc you may also find it useful to measure with dc
applied, and perhaps even before/after use to see if this has had
some effect!

The original claim was that a listening panel would be able to
differentiate between an industrial/commercial grade coupling capacitor
andJensen capacitor.

Afraid I am quasi-paranoid about such things as I've spent 20+ years
working on precision measurement systems for people like the NPL 8-

It shows:-) But I am sure it is a good thing, and that we can all
sleep more safely in our beds:-)

I believe that Jim is just covering the bases, I myself would be happy
with a simple check of value at 100Hz and 10kHz.

I'd like accurate values for a range of frequencies as this would help
check for some of the claimed 'effects' I've read about in magazines, but
where no actual evidence ever seems to appear. Also to be able to model
some of the relevant cap imperfections.

FWIW If measurable, I'd also like to know the cap leakage resistance under
the conditions of use.

I think I'd like to know things like the frequency and phase response
with each cap in circuit, again with a fair degree of precision.

Those are surely things that will have to be determined afterwards, to
find out why there is a difference in sound, if one can be established.

Agreed, there's little point in intensive investigation *before* you've
observed an audible difference.

I disagree. :-) The worry I have is that after the event people well then
say, "Ah well, now you've published your responses we can see that is shows
why you [heard]/[didn't hear] a difference - the audible change was [caused
by]/[masked by] the response differences. Thus my preference is to get such
figures in advance and invite people to say if they seem close enough not
to matter *before* they find out if the listening results agree with their
own views or suspicions.

I would alway want to characterise a measurement system *before* making
serious measurements. Otherwise I may be wasting time making observations
which simply aren't reliable.

I'm not sure about this TC business, it seems to introduce an additional
stress, and has nothing to do with identifying the Jensen per se.

I am similary concerned about that.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

In article , Don Pearce
wrote:
On Wed, 5 Jan 2005 22:44:31 +0200, "Iain M Churches"
wrote:


We like the idea of the panel just having to listen, and press a cue
button if and when they hear a change. Nothing else.

The problem you will have with this approach is that of interpretation
of results - how close constitutes a hit?

Another concern I have with the above method can be explained as follows.

If the cap is 'switched' during each session, how long a delay before a
'detection' will be cease to be meaningful?

If the cap *is* switched during each session, perhaps sometimes more than
once, you can argue that a 'hit' could be produced just by waiting for some
time then saying you hear a change. Provided the listener waits for long
enough the chances are a switch *will* have occurred, but they could just
be guessing.

The only way to avoid this I have thought of is to arrange that in over 50%
runs no switching at all occurs, and the delays between switch events is
very long. However this takes a lot longer.

Similarly, if more than one switch event occurs during a run, does only
'detecting' the second event not count? Deciding this could be a nightmare.
It will then be detecting a possible response to two (or more) successive
and potentially cancelling events. Hence it becomes ambiguous what may have
been 'detected' - if anything.

The method is I think also really a form of 'AB' test, not 'ABX'. The
advantage of ABX is its ability to help deal with some of the statistical
and interpretations effects like those I have mentioned and which occur
with AB methods.

As I have said, I would prefer ABX where the *listener* can switch as and
when they choose, and can take as long as they wish before deciding what
they think 'X' may be. This avoids 'wait long enough and you may be right,
or you may be guessing' type problems. It also avoids things like a faint
click on switching being picked up as a subconscious 'cue' since the user
would already know when switching occurred, but would not know from that
the identity of 'X'. I would have said that this maximises the ability of
the listener to hear genuinly audible changes whilst minimising the effects
of some spurious chances clouding the results.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

"Jim Lesurf" wrote in message
...
In article , Iain M Churches
wrote:

Just a moment. Not long ago, the general opinion seemed to be that we
would not be able to tell the difference when changing between two caps
of the same value and voltage working from different makers. Now we are
being asked to match them to a tight level :-))


My concern is not with "the general opinion". It is with being able to
carry out a test whose results can then be regarded as having some
validity
and reliability. If you don't ensure the values are closely matched then
you have no evidence that a slight change in value is either causing an
audible change or hiding one. So if you concern is to test audibility of
"type" of cap (i.e. chosen dielectric, method of construction, etc) then
you have to control/match those values which might confuse or confound the
test.

This started out as the kind of test one should be able to do at home, in
one's own listening room. It has grown into something of gigantic
proportions,
and I feel that the ensuing unwieldiness will make it difficult to draw
conclusions. In this respect, I think that Mike's method has a great deal
to be said for it.

Bear in mind that some people claim that effects like the tan delta are a
'cause' here. (IIRC your Jensen rep has said this.) If so, then very
careful matching is implied as if this *is* the cause it implies human
sensitivity to very small differences.

The Jensen rep, just like Morgan Jones, suggests that this *might* be
an explanation. He also mentioned to me that better manufacturing methods,
materials and tighter control parameters (all of which are reflected in the
price of the capacitor) probably have a marked significance.

If you are using a single-frequency cap bridge, you may need to
consider the effects of series ESR, etc. Given the points made
elsewhere about the dc you may also find it useful to measure with dc
applied, and perhaps even before/after use to see if this has had some
effect!

The original claim was that a listening panel would be able to
differentiate between an industrial/commercial grade coupling capacitor
andJensen capacitor.

Yes. The snag is that such a claim may be trivial to support if the caps
differ markedly in value. Discovering that different values can produce
audible changes would hardly be a startling result. :-)

Looking at a few schematics, CJ, Radford, etc they simply state the value
and working voltage of the cap, so one would presume that if the
component to be tested closely met these two parameters then
it would be as satisfactory component for test as it is for manufacture.

The problem with doing that "after the event" is that people can then
decide whether they think the match was "good enough" already knowing your
test results. I think you need to establish in advance that those
interested all agree that the level of match is close enough to ensure
that
a difference in gain, response, etc, is not the 'cause' (or masking some
other effect).

If there is no audible difference, then the test will end there, and we
shall
all retire to the lounge bar of the Frog and Nighgown. If there are
differences,
then it will be for the following discussions and perhaps furthert tests to
find
out the reasons.


Well you be using mechanical switching, electronic, what type, etc?'

FET switching at the input of the console. One pair of inputs active, and
the other muted.




TBH Iain if you have not already done so it may be useful if you have a
look at previous 'ABX' and 'double blind' studies that have been carried
out for things like this.

I have been sent a lot of most welcome pointers. I was hoping that Arny,
who it sems talks about ABX on his website would have joined this
discussion. Are you there, Arny?

The point is that many of these end up being
criticised for various (alleged) defects which then allow people who do
not
agree with the results to dismiss them. If you do not check this you may
simply be wasting your time as people will reject your results.

At the outset I was simply interested to satisfy myself as to whether
I could hear an audible difference. Then I added nine others, all involved
in
some way or another in music or audio, to confirm what (I hope) I can hear
or
not hear. I certainly don't regard this as a waste of time, and am not
really to
worried about whether or not I can convince the whole world:-)

There seem to be a multitude of greatly differing opinions as to how the
test should be conducted. One of the panel is a technical reviewer, whose
magazine is interested to cover the test and its findings.

Iain

On Thu, 6 Jan 2005 14:15:42 +0200, "Iain M Churches"
wrote:


This started out as the kind of test one should be able to do at home, in
one's own listening room. It has grown into something of gigantic
proportions,
and I feel that the ensuing unwieldiness will make it difficult to draw
conclusions. In this respect, I think that Mike's method has a great deal
to be said for it.

I think you have complicated everything to a huge degree by doing all
the studio and time code stuff. All you really need is a toggle switch
on the amplifier to decide which cap is in circuit. After that it is
just protocol. That reduces the test to its essence, and really does
guarantee you meaningful results - provided you run the trial well.
All that other stuff will just provide points of contention to a
degree that will invalidate the results.

d

Pearce Consulting
http://www.pearce.uk.com

On Thu, 6 Jan 2005 14:15:42 +0200, "Iain M Churches"
wrote:

Looking at a few schematics, CJ, Radford, etc they simply state the value
and working voltage of the cap, so one would presume that if the
component to be tested closely met these two parameters then
it would be as satisfactory component for test as it is for manufacture.

It's probably worth mentioning that the caps don't have to be exactly
on the nominal value, they just have to be the *same* value.

--

Stewart Pinkerton | Music is Art - Audio is Engineering