Audio Cable Tester Requirement

[jneutron]

John;
What frequencies are you talking about? Low frequency bass (where the most amplifier power is required) below 80Hz is omnidirectional and the ear cannot perceive directionality or lateral clues?

I am talking about the low frequency, high power bass affecting the damping factor and timing response of the amplifier to what is perceived by humans in lateralization..so I'm not talking about lateralization of bass, but rather, how the bass affects the lateralization of signals from about 1.2Khz to 12 Khz...

I suggest reading some of the white papers authored by Dr. Floyd Toole and Thomas Holman (formerly of THX). These guys are dead on about bass and I don't really follow you here.

Boy, you left yourself open with that one...but I've been in a nicer mood since I was upgraded from "audiophyte" to "enthusiast"..

From your statements, it is clear you did not follow me.....as I said, pics or drawings would have helped..

Also, beyond a certain minimal measurable damping factor, it becomes more of an exercise in academics rather than a real world concern.

In terms of amplitude, of course...and most speakers are designed nowadays with decent Q's (except of course, for those darn car subs that rattle the house windows..)

But, my investigation surrounds the way the amplifier is capable of controlling a complex load to the tune of 1.5 uSecs, while also dissipating lots of power in a complex way.

Cheers, John

PS...perhaps this discussion should be cut and pasted into a new thread?

 

[WmAx]

I am talking about the low frequency, high power bass affecting the damping factor and timing response of the amplifier to what is perceived by humans in lateralization..so I'm not talking about lateralization of bass, but rather, how the bass affects the lateralization of signals from about 1.2Khz to 12 Khz...

Seems that a low frequency sine wave, overlaid with periodic impulses would be a suitable test signal. Record/sample the resultant waveform from the amplifier at moderate to high amplitudes into a sufficient speaker. Overlay the digital test signal and the recorded signal, align/shift to account for latency/delay then observe/measure the impulses for deviation from the original signal file. Of course, a small tolerance of error may be needed in order to account for some DAC clocking errors depending on the length of your measurements/samples.

Does that seem like a sufficient test signal/general method?

-Chris

 

[jneutron]

Seems that a low frequency sine wave, overlaid with periodic impulses would be a suitable test signal. Record/sample the resultant waveform from the amplifier at moderate to high amplitudes into a sufficient speaker. Overlay the digital test signal and the recorded signal, align/shift to account for latency/delay then observe/measure the impulses for deviation from the original signal file. Of course, a small tolerance of error may be needed in order to account for some DAC clocking errors depending on the length of your measurements/samples.

Does that seem like a sufficient test signal/general method?

-Chris

Sounds like a reasonable method..this style of measurement was proposed at either an IEEE or AES convention a while back...I forget the details...but I think they decided to use FFT analysis to check the results...FFT will not see a time shift in the waveform, and I'm not sure if it will see asymmetrical time shifting..and they certainly didn't consider 1.5 usec things..

I can think of a lot of things that can cause errors...those will have to be worked out..

Another point to consider...1.5 usec timing is a big bear w/r to 8 ohm impedances...the field collapse voltages are large in comparison..so accurate representation of the voltage is an issue.

One more: since an electrodynamic speaker is a current based device, it may not be the voltage we wish to look at...but, the current..meaning a CVR capable of half a Mhz response...and very low resistance...perhaps .1 ohm or so...

So many questions...not enough time...

Buying some quickcrete today...for my basement lab...

Cheers, John

 

[wankerr]

analyzing speaker cable beyond 20khz

Why is it important to test a speaker cable's inductance, capacitance, etc.. beyond what people can hear (20-20khz)?

 

[jneutron]

Why is it important to test a speaker cable's inductance, capacitance, etc.. beyond what people can hear (20-20khz)?

hmmm..

graph 1..the calculation of the delay as a result of time of flight from a source, to each ear...ears spaced 6 inches apart (hopefully not variable), and distance from source ten feet. For a human to discriminate the location of the source to within 6 inches requires the ability to hear 20 usec level delays.

Graph 2..measured human lateralization capability..note that at 200 hz, 15 usec is the best one can hope for, while, unjittered, 5 uSec is capable..Jittered, 1.5 uSec is realized, from 2Khz up to 12 (this particular graph only goes to 8Khz, but subsequent graphical results go to 12Khz)

This graph clearly shows what human sensitivity thresholds are.

Graph 3: simple depiction of sin(x), cos(x) at 10% level, and sum of the two..it is of note, that addition of the cos component only shifts the wave..it does not change the amplitude, as when sin(x) is peak (90 degrees), cos(x) is zero. This means that any reactance in the wire will shift the signal in time..the right side of this graph zooms in on the zero crossing point..note the shift of the zero crossing point..

If this shift as a result of reactance, is within human discrimination levels, it will affect the soundstage, or virtual image, of the reproduction..while, not affecting the amplitude to a huge degree..

Cheers, John

 

[WmAx]

While the lateralzation can be detected under special test conditins, using different timing pulses in each channel; what is the implication for a steady rate change in phase as frequency rises? You see to be referring to a symmetrical phase difference in each channel in this reply; i did not read the entire previous thread, so perhaps I missed something from before.

Listening tests have been performed to test the effect of quite severe phase angle deviations into upper audible bands using varied low pass filters, and the thresholds of detectivity were suprisingly high -- demonstrating typical severe rate of phase difference in the top octave as produced by the typical RBCD player to be inaudible among the tested sample groups. Again, I may have misunderstood your implication.

-Chris

hmmm..

graph 1..the calculation of the delay as a result of time of flight from a source, to each ear...ears spaced 6 inches apart (hopefully not variable), and distance from source ten feet. For a human to discriminate the location of the source to within 6 inches requires the ability to hear 20 usec level delays.

Graph 2..measured human lateralization capability..note that at 200 hz, 15 usec is the best one can hope for, while, unjittered, 5 uSec is capable..Jittered, 1.5 uSec is realized, from 2Khz up to 12 (this particular graph only goes to 8Khz, but subsequent graphical results go to 12Khz)

This graph clearly shows what human sensitivity thresholds are.

Graph 3: simple depiction of sin(x), cos(x) at 10% level, and sum of the two..it is of note, that addition of the cos component only shifts the wave..it does not change the amplitude, as when sin(x) is peak (90 degrees), cos(x) is zero. This means that any reactance in the wire will shift the signal in time..the right side of this graph zooms in on the zero crossing point..note the shift of the zero crossing point..

If this shift as a result of reactance, is within human discrimination levels, it will affect the soundstage, or virtual image, of the reproduction..while, not affecting the amplitude to a huge degree..

Cheers, John

 

[jneutron]

While the lateralzation can be detected under special test conditins, using different timing pulses in each channel; what is the implication for a steady rate change in phase as frequency rises? You see to be referring to a symmetrical phase difference in each channel in this reply; i did not read the entire previous thread, so perhaps I missed something from before.

Listening tests have been performed to test the effect of quite severe phase angle deviations into upper audible bands using varied low pass filters, and the thresholds of detectivity were suprisingly high -- demonstrating typical severe rate of phase difference in the top octave as produced by the typical RBCD player to be inaudible among the tested sample groups. Again, I may have misunderstood your implication.

-Chris

It seems the listening tests you are talking about are either one channel only, as in monophonic reproduction, or the exact same change being applied to both channels. That is not what I'm talking about.

Cheers, John

 

[WmAx]

Ok. Then I'm lost.

Are you suggesting some audible degree of assymtery between two pieces of the same calbe used on each channel? Regardless of the LCR parameters any any frequency, if the same model/length cable is used on each channel the effect will be symmterical. Or are you suggesting the dynamic difference betwwen the 2 stereo channels will result in potential audible assymetry on specific cable? I don't think you mean this -- so could you please clarify as to what association you mean in reference to lateralazation discrimination vs. cable measurements?

-Chris

It seems the listening tests you are talking about are either one channel only, as in monophonic reproduction, or the exact same change being applied to both channels. That is not what I'm talking about.

Cheers, John

 

[jneutron]

Ok. Then I'm lost.

Are you suggesting some audible degree of assymtery between two pieces of the same calbe used on each channel? Regardless of the LCR parameters any any frequency, if the same model/length cable is used on each channel the effect will be symmterical. Or are you suggesting the dynamic difference betwwen the 2 stereo channels will result in potential audible assymetry on specific cable? I don't think you mean this -- so could you please clarify as to what association you mean in reference to lateralazation discrimination vs. cable measurements?

-Chris

Wrong..right..wrong...

I am talking about a dynamic difference between the channels..

Timing skews in a single channel audio stream at the 10 to 20 uSec level, I don't think there's a snowballs chance that it would be audible..unless there are gross errors as a result of a wire, nobody's gonna detect it.

The issue is: can the spectral power content in a channel cause the amp/wire/load combination to time shift some of the signal, the region from about 500 hz to 12 Khz..

Hence the amp damping factor question..and the attention being focussed on inductances that should not cause a 20 to 20Khz audible change..

I've only seen the scientific audio establishment worry about FR and phase shift from 20 to 20K, but yet we can "see" 1.5 uSec time shifts, which is quite a ways outside the 20K realm. So, what I describe is beyond normal test regimes..

Hey, why would I bother looking in the same ol' box? So many have been there...why would I expect to uncover something there?

But, 1.5 uSec timing shifts??now there's uncharted territory...that's the fun stuff..

Cheers, John

 

[WmAx]

The issue is: can the spectral power content in a channel cause the amp/wire/load combination to time shift some of the signal, the region from about 500 hz to 12 Khz..

Hey, why would I bother looking in the same ol' box? So many have been there...why would I expect to uncover something there?

So, you are essentially investigating something that has not been established as being an issue of concern in measurements actally relating to 'known' audibility or suspected due to results from specific controlled listening tests of various amp/wire combinations isolating this variable? If so, happy testing/investigation.

-Chris

 

[jneutron]

So, you are essentially investigating something that has not been established as being an issue of concern in measurements actally relating to 'known' audibility or suspected due to results from specific controlled listening tests of various amp/wire combinations isolating this variable? If so, happy testing/investigation.

-Chris

Absolutely 100% correct...

I have started a line of investigation, in response to anecdotal accounts of people "hearing" differences in cables..where, based on accepted knowledge, none should exist.

From this investigation, I've seen that the accepted knowledge base is in essence, frequency response and phase shift, from 20 to 20K. That base covers the entire range of frequencies we humans are capable of, so the current thinking has been "that is all there is to it", and that any differences that are heard, are just a figment of the imagination.

And then, in comes lateralization..humans are capable of discerning 5 uSec timing shifts right to left, and test data indicates as low as 1.5 uSec..

None of the currently established testing regimes look at that level of timing errors..none..it is the equivalent of half a Mhz, and is a factor of 25 faster than we can hear..current regimes ignore this area..in single channel streams, as well as stereo ones. In single channel streams, that level of shift is unimportant..but the relation between two streams is..

So, that is where I look..Gene has been kind enough to test up to a Mhz, so at least a basis of measurement can be started.

Continued posting of my efforts....I think this may be the better forum for it, as Gene allows pics, does tests, and may put up a sticky thread if I ask..

Not to mention, the hostility I get over at AA from several people isn't really worth my time..there are, of course, others there who are adamant cable believers, and provide excellent descriptors and feedback which I value..Ted Smith gave me some really good descriptors when I sent him some 10 nH/288 pf cable to try. I'm trying to get it back to send Gene, but I seem to have fallen out of good grace with Ted as a result of a difference of opinion...so my cables may be gone forever..oh well, I'll just make some more..

Cheers, John

 

[WmAx]

Absolutely 100% correct...

I have started a line of investigation, in response to anecdotal accounts of people "hearing" differences in cables..where, based on accepted knowledge, none should exist.

None of the currently established testing regimes look at that level of timing errors..none..it is the equivalent of half a Mhz, and is a factor of 25 faster than we can hear..current regimes ignore this area..in single channel streams, as well as stereo ones. In single channel streams, that level of shift is unimportant..but the relation between two streams is..

So, that is where I look..Gene has been kind enough to test up to a Mhz, so at least a basis of measurement can be started.

In order to increase efficiency/evaluate feasibility of the extent of this research -- don't you think it would be beneficial to organize DBT or ABX method testing cmoparing a cable that is linear at high frequency(500kHz) and one that is non-linear at high frequency, level matched, to evaluate the audibility before focusing on a specific target of research? Or, this may be of further interest then simply audibility for you? Acedemic value?

Not to mention, the hostility I get over at AA from several people isn't really worth my time..there are, of course, others there who are adamant cable believers,

I admit I do not read AA except on rare occasion. However, I do seem to remember a popular amplifier designer named 'John Curl' with whom you were conversing a few months back in the Prop Head forum; it seemd from the few posts I bothered to read, he was being irrational and accusatory in a discussion concerning capacitor dielectrics. Is that the kind of hostility to which you are referring?

and provide excellent descriptors and feedback which I value..Ted Smith gave me some really good descriptors when I sent him some 10 nH/288 pf cable to try. I'm trying to get it back to send Gene, but I seem to have fallen out of good grace with Ted as a result of a difference of opinion...so my cables may be gone forever..oh well, I'll just make some more..

Did such test take place under DBT/ABX level matched conditions? If so, do you have the test data to publish?

-Chris

 

[jneutron]

In order to increase efficiency/evaluate feasibility of the extent of this research -- don't you think it would be beneficial to organize DBT or ABX method testing cmoparing a cable that is linear at high frequency(500kHz) and one that is non-linear at high frequency, level matched, to evaluate the audibility before focusing on a specific target of research? Or, this may be of further interest then simply audibility for you? Acedemic value?

I work in the field of e/m theory..so this avenue is of tremendous academic interest to me..

Since DBT stuff has not seen a difference, where so many swear there is, I question using that method until more is known about what we are looking for.

When one considers what is being looked for here, I can't get excited about DBT...yet...

I'll give a brief explanation of that. Soundstage imaging is not a hardware entity...it is a construct entirely of the mind..we are attempting to fool our brain into believing that something is producing sound where no physical object is...as a result, it is necessary for the brain to learn by experience, how to interpret what is being heard, and construct the virtual sound field pattern for our enjoyment..

This is the exact same thing that happens when you look at one of those 3-D posters, with the whale, or bird, or whatever, encrypted into the picture..you have to stare at it, drifting the eyes out of normalcy, to derive an image.. Once learned, it is easier and easier to form the image..some people, however, are unable to do it no matter what..

Now, back to audio..how long does it take for the brain to re-establish the virtual image of soundstage, if something in the system, cables for example, causes a shift of the temporal relations we are using for building that image? If it takes an hour of listening, how is rapid shifting gonna help discern a change? If it takes ten days??

I admit I do not read AA except on rare occasion. However, I do seem to remember a popular amplifier designer named 'John Curl' with whom you were conversing a few months back in the Prop Head forum; it seemd from the few posts I bothered to read, he was being irrational and accusatory in a discussion concerning capacitor dielectrics. Is that the kind of hostility to which you are referring?

Yep...also, one of the moderators plays fast and loose, spouting all kinds of derogatory insults to anyone who disagrees with his point of view.....and occasionally, one or two others will chip in to defend the poor behaviour of these two..It bugs me when they badger others also, and I make my feelings clear in that regard...hence the fallout with Ted..

Did such test take place under DBT/ABX level matched conditions? If so, do you have the test data to publish?

It was not a controlled scientific test..it was Ted just putting the wires in his system, and telling me what he heard..

His comments were very well constructed, to give me an interesting indication as to what he felt to be the differences..it has led me towards further research into how we visualize sounds in 3 dimensions, and how we try to fool our brain into thinking something is there via two sound sources. Where I can, I glean from published papers, and where I can't, I do the tests myself..

I gotta tell ya, this downtime to build my lab is totally killing me...

Cheers, John

 

[Mudcat]

John, Chris,

I've been following this thread for a while even though most if not all is way (I mean stratospheric) over my head, I have an observation/question.

We are aware that an unterminated cable will send reflections back down the wire. To what degree/phase/amplitude/whatever I do not know, just what I've read. Would a crimp style termination do the same, and would it effect the phasing/lateralization/dohicky?

Would a good solder termination with good silver solder eliminate this (if there is anything to eliminate)?

Just an ME trying to keep up appearances.

 

[jneutron]

John, Chris,

I've been following this thread for a while even though most if not all is way (I mean stratospheric) over my head, I have an observation/question.

We are aware that an unterminated cable will send reflections back down the wire. To what degree/phase/amplitude/whatever I do not know, just what I've read. Would a crimp style termination do the same, and would it effect the phasing/lateralization/dohicky?

Would a good solder termination with good silver solder eliminate this (if there is anything to eliminate)?

Just an ME trying to keep up appearances.

For the frequencies I am talking about, and even for the freq's steve nugent is talking about (up to 100 Mhz) there will be no effect of choosing either solder type, or termination type...w/r to what we are discussing.

I have seen effects of this nature, where the solder joint size, using a lug or crimp, or how the wire is bent going to the joint. However, it was when I was doing trr testing, with a 50 ohm source cable driving 100 volts in 250 picoseconds...at that rather high current slew rate, I found that 1/8 of an inch of wire extending past the solder joint of the 1 ohm current viewing resistor would cause reflections and slight ringing at the cvr. But for extremely slow rise times (10 to 12 nanoseconds), lugs and joints and lead extension didn't have any effect. I did note a delay of a nano or two when the 250 ps signals had to go through an aluminum wall, where the hole was 3/8 of an inch dia and the wire #14 solid..but I believe it was an impedance mismatch thing, the slew was delayed by the image currents in the aluminum requiring some field energy to set them up. (it forced me to replace the aluminum with plastic..the aluminum was just a test fixture wall, there to keep the operator from touching the hot BeO microwave resistors while holding the diode.)

The solder joint will affect it if it is bad..but, for the purposes of this discussion, the main entity is the impedance mismatch.

Cheers, John

 

[gene]

Of course at frequencies where cable termination methods matter for speaker cables there is no amplifier gain available for it to become an issue, not to mention the brickwall filter in digital playback systems ensures signal attenuation greater that 100dB after 22kHz

 

[Wayne Kerr Man]

Audio Cable Analyzer

As promised, I have the 1st version of the specification of a audio cable tester. Should I post here for comments or should I welcome you to visit our new website : www.waynekerrusa.com? I can post there as content of my bi-weekly news letter. Ideally I would like to get a survey to see how interest Audiholics audience may have to own such an equipment.

 

[gene]

By all means post the spec here. I am certain you will get some good feedback from all the cable loonies (like me) participating on this forum

 

[Wayne Kerr Man]

Audio Cable Analyzer

Wayne Kerr Audio Tester Gear Preliminary Specification
Audio Cable Tester
Purpose: The purpose of this test gear is to measure cable metrics for speaker leads, audio and video interconnects.
i. Measurement
i. Rs, Total Resistance
ii. DC Resistance (precise milliohm meter)
iii. AC Resistance (Skin effect DC to 500KHz)
iv. Loss
v. Phase Change (degrees)
vi. Ls / Lp (DC to 500KHz; Series Inductance/Parallel Inductance)
vii. Cs/ Cp (DC to 500KHz; Series Inductance/Parallel Capacitance)
viii. RLC measurements with DC bias up to 1AMP
ix. RLC measurements with 20 mV to 10V sourcing
ii. Connector Interface
i. RCA Connector
ii. Banana Connector
iii. BNC Connector/with custom binding post adapter
iii. User interface
i. Ability to download data to computer interface with software complaint to Windows 2000/XP
ii. Sweep mode for selected frequency range with ability to import graphs to computer software
iii. Interface via USB with Parallel ports for printing.
iv. LAN connection would be useful for networking to a remote server.

 

[Wayne Kerr Man]

no response?

oops, no response ? No interest or ?