Engineering White Papers suggest that only A/B comparison.. [Archive] - Audio & Video Forums

PDA

View Full Version : Engineering White Papers suggest that only A/B comparison..



Tony_Montana
07-21-2004, 01:32 PM
.....done in real time (which mean no delays) is needed to recognize difference caused by cables.

The article suggest that:"Taken [cable's effects] altogether (ie., inductance, resistance, interference, skin effect, capacitance), these ugly realities introduce various dynamic and time-related forms of signal distortion which are very difficult to quantify with simple sine-wave measurements. When complex waveforms have their harmonic structures altered [by a cable], the sense of immediacy and realism is reduced. The ear/brain combination is incredibly sensitive to the effects of this type of phase distortion, but generally needs direct, A/B comparisons in real time to recognize them."

I agree with above statement 100%, especially doing it in real time. Any cable testing protocol that is not done in real time and have "delayed" comparison (such as DBT) might indeed introduce human error into testing result, making the results colored.

Also note that article didn't say anything about comparison should be sighted or not :D

http://www.cardinalproaudio.com/main/speaker.htm

Pat D
07-21-2004, 07:03 PM
.....done in real time (which mean no delays) is needed to recognize difference caused by cables.

The article suggest that:"Taken [cable's effects] altogether (ie., inductance, resistance, interference, skin effect, capacitance), these ugly realities introduce various dynamic and time-related forms of signal distortion which are very difficult to quantify with simple sine-wave measurements. When complex waveforms have their harmonic structures altered [by a cable], the sense of immediacy and realism is reduced. The ear/brain combination is incredibly sensitive to the effects of this type of phase distortion, but generally needs direct, A/B comparisons in real time to recognize them."

I agree with above statement 100%, especially doing it in real time. Any cable testing protocol that is not done in real time and have "delayed" comparison (such as DBT) might indeed introduce human error into testing result, making the results colored.

Also note that article didn't say anything about comparison should be sighted or not :D

http://www.cardinalproaudio.com/main/speaker.htm
If you can avoid significant delays in switching under sighted conditions (and this can be done relatively easily with the right equipment, such as an ABX comparator), then it can also be done under blind conditions. N'est-ce pas? So your premise seems to be incorrect. Indeed, the DB audition would lose considerably in sensitivity without fast switching times so that any delay between comparisons is only a tiny fraction of a second.

mtrycraft
07-21-2004, 09:24 PM
.....done in real time (which mean no delays) is needed to recognize difference caused by cables.

How do you do a no delay comparison? A switching has delay, even if it is very fast. Don't understand what they mean.

The article suggest that:"Taken [cable's effects] altogether (ie., inductance, resistance, interference, skin effect, capacitance), these ugly realities introduce various dynamic and time-related forms of signal distortion which are very difficult to quantify with simple sine-wave measurements. When complex waveforms have their harmonic structures altered [by a cable], the sense of immediacy and realism is reduced. The ear/brain combination is incredibly sensitive to the effects of this type of phase distortion, but generally needs direct, A/B comparisons in real time to recognize them."

Yes, that is what they suggest. So what? I didn't see them offer evidence, did you?


I agree with above statement 100%, especially doing it in real time. Any cable testing protocol that is not done in real time and have "delayed" comparison (such as DBT) might indeed introduce human error into testing result, making the results colored.

DBT is a protocol. Has nothing to do with delays, real time, nothing.

Also note that article didn't say anything about comparison should be sighted or not :D

I wonder why? Maybe an email to the author for a clarification?

Tony_Montana
07-22-2004, 01:49 PM
Indeed, the DB audition would lose considerably in sensitivity without fast switching times so that any delay between comparisons is only a tiny fraction of a second.

That was exactly the point. If there are delays (probably more than a second), then it doesn't matter if it is sighted or DB, it will have tendency to be colored since memory will be involve.
Some claim that DBT is the only method that is valid and flawless. But if there are delays between testing subjects, then it might not be :)


How do you do a no delay comparison? A switching has delay, even if it is very fast.

Not if it is done at a fraction of second (time it takes to toggle the switch). It probably take that much time for ear to process the data anyway :)


Yes, that is what they suggest. So what? I didn't see them offer evidence, did you?

They don't need evidence. You can try for your self and find out. I did

http://forums.audioreview.com/showthread.php?t=2537&highlight=equalizer


DBT is a protocol. Has nothing to do with delays, real time, nothing.

May be or may be not. If there are too much delays in DBT, then human bias might color the results since memory will have a role. And you know how volatile human memory and its perceive(ness) is :)

skeptic
07-22-2004, 07:41 PM
"The article suggest that:"Taken [cable's effects] altogether (ie., inductance, resistance, interference, skin effect, capacitance), these ugly realities introduce various dynamic and time-related forms of signal distortion which are very difficult to quantify with simple sine-wave measurements. "

The mathematics of waveform analysis have been known since Laplace and Fourier well over a hundred years ago. They will never become obsolete because they are correct. The people who promote certain types of audio equipment including special cables have negated the value of these analyses but offer none of their own as a substitute.

I think it's interesting that we can analyze the waveforms of millions of chemicals using infrared spectroscopy and identify virtually every single one using this method, we can set standards for the waveforms for broadcast and reception of radio signals up into the microwave region, we can analyze the waveforms of light and x-rays from outer space so faint they couldn't even be detected a few decades ago and deduce the structure of galaxies billions of lightyears away, we can analyze other radio waves and deduce the structure of subatomic particles but we cannot measure or understand the waveforms of audible distortion in speaker wires. Doesn't sound plausible to me.

As for DBTs, what other method do those who dismiss well run experiments to test whether a difference can be detected or is the result of random guessing offer as a substitute?

It seems to me all I hear are anecdotes about how great some product or other sounds and half baked theories from squirrel brained wanabee scientists about obscure theories of fermi velocity of electrons, grain boundries in metals, strand jumping, but no hard evidence of any kind whatsoever. They may not have anything to convince a scientist or engineer but then most of the market for their products are not scientists or engineers.

mtrycraft
07-22-2004, 08:44 PM
That was exactly the point. If there are delays (probably more than a second), then it doesn't matter if it is sighted or DB,

Of course it still matters. The longer the time delay the less effective the memory. However, in a sighted test, memory is not the issue but the perceptions that are grossly unreliable :)

Some claim that DBT is the only method that is valid and flawless. But if there are delays between testing subjects, then it might not be :)


Two issues. One is of the time delay, other is in protocol. In a sighted listening, no matter how rapidly you switch, it is still has no meaning as it is still uncontrolled for bias.


Not if it is done at a fraction of second (time it takes to toggle the switch). It probably take that much time for ear to process the data anyway :)


You can do such rapid switching with the ABX box.
By the way, what are you comparing in such a protocol? a soundtrack between what is played now and what was played before, unequal events. And, you only address the switching. How long of the track you listen before switching? Or, is it a constant switching so you can confuse yourself?



They don't need evidence.

That is not how science works.

You can try for your self and find out. I did

What am I comparing? Compare what was played in the past with what is playing now? No, you cannot compare a running tape and have any meaning.





May be or may be not. If there are too much delays in DBT, then human bias might color the results since memory will have a role.

Not bias, but less accurate results. No contest here. Never claimed it.
But, you still cannot run a continuous tape and compare two different events. Has no meaning.

And you know how volatile human memory and its perceive(ness) is :)
Yes, :)

skeptic
07-23-2004, 02:58 AM
"You can do such rapid switching with the ABX box.
By the way, what are you comparing in such a protocol?"

But what ARE you comparing? If you are comparing two wires, you are comparing two elements both of which are less than perfect. Because wires are passive elements, you can AB one wire against a shunt, a virtually ideal real world element that the wire can be compared to on an absolute rather than a relative basis. To the degree that the wire fails to match the performance of the shunt, it is flawed. Both it's degree and nature can be determined, if not by hearing then by measurement. If an audible difference exists, it is important to know based on an analysis of the waveform differences if it is due to linear or non linear distortion. Linear distortion can be corrected by other circuit elements. Non linear distortion usually (but not always) can't.

Tony_Montana
07-25-2004, 03:18 PM
I think it's interesting that we can analyze the waveforms of millions of chemicals using infrared spectroscopy and identify virtually every single one using this method, we can set standards for the waveforms for broadcast and reception of radio signals up into the microwave region, we can analyze the waveforms of light and x-rays from outer space so faint they couldn't even be detected a few decades ago and deduce the structure of galaxies billions of lightyears away, but we cannot measure or understand the waveforms of audible distortion in speaker wires. Doesn't sound plausible to me.

I agree with your statement 100%. But the article said "with simple sine-wave measurements"...such as from an oscilloscope. Harmonic structure of signal can not be analyzed with such a instrument. But it can be analyzed with an Spectrum Analyzer :)


In a sighted listening, no matter how rapidly you switch, it is still has no meaning as it is still uncontrolled for bias.

There are too much generazalation in that statement. [IMO] The beauty of rapid switching is that it can be repeated as many times as needed easily (with just using a thump) until a person is satisfied that A sound better than B. So even if the person is slightly biased toward B because it is fancier, then repeated testing will either confirm his/her biasing or destroy it. There are no two way about it :)


What am I comparing? Compare what was played in the past with what is playing now? No, you cannot compare a running tape and have any meaning.

Then how would one conduct rapid switching if not a running tape? :D

skeptic
07-25-2004, 04:22 PM
"I agree with your statement 100%. But the article said "with simple sine-wave measurements"...such as from an oscilloscope. Harmonic structure of signal can not be analyzed with such a instrument. But it can be analyzed with an Spectrum Analyzer "

Then you disagree with Fourier. Fourier says that even non periodic waveforms can be described as the sum of an infinite number of sine waves of different amplitudes. A spectrum analyzer only indictes the relative amplitude of those sine waves within a given time window and within a given range which for audio is 20 hz to 20 khz. The specific equation which takes a waveform from the time domain to the frequency domain is f(jw)= integral from 0 to infinity of f(t)e exponent jwt times dt where w (omega) is 2 pi times the frequency and e is the base of natural logarithms. dt of course is an infinitesmal unit of elapsed time. The inverse function which converts from the frequency domain back to the time domain is similar. You will find it in any book on second year calculus or a beginning book on linear systems analysis. In fact it is the starting point and foundation for this kind of analysis.

BTW, a real spectrum analyzer doesn't have a bar graph type display of the type that you see in consumer units, it has a crt or other display like an oscilloscope and displays amplitude versus frequency continuously. It also uses a sweep frequency generator, not a pink or white noise generator as the test input. The sweep of the oscilloscope horizontal (x) axis trace is synched to the sweep frequency generator while the calibrated microphone of course provides the drive for the vertical (y) axis.