| mrfeedback |
Hi John, Thanks for the kind words.
Yes, to clarify, the lead in question is a 240V extension lead about 400mm long, with a standard plug at one end and standard in-line socket at the other.
"My comments may have gotten lost, but I hear a noticeable improvement using an AC power conditioner."
I hear changes with such condtioners also.
I have several 240V-240V isolation transformers, and I hear differences with these too that I regard as improvement.
"So it is where it is interesting to find news, such as yours, that aids in learning about a sound system. The rattling windows seemed like an AB test that just presented itself."
Yes, we were doing A/B tests wrt the 240V cable and mostly listening to the better clarity in mids and highs when surprise surprise the window rattled with the cable in place.
Actually this did not come as surprise to me as I have used this cable in a PA system, and the effect there is to quite dramatically clean mids and highs, and the low bass becomes stronger and better defined too, such that the soles of one's shoes buzz when standing on thinly carpeted cement floor.
Despite system eq's I am not able to achieve this without the cable supplying the system.
"So, what is your thoughts on measuring cables? Does it surprise you that the distortion measurements don't reveal anything? Or do you think we don't know how to make the proper measurement?"
Yes, I suspect that the test methodology is not able to replicate what Bruno, JC, myself and others are hearing.
Music is of course a non steady state waveform, and it may be that sine wave testing does not reveal decay products due to DA etc.
"(As far as my power conditioner, I guess I think it must be breaking some sort of ground loop. This is just a small point though. Something else comes to mind.)"
There are two types of conditioner - on line and off line, but as far as I know the earth wire connection is continuous through both types.
I think you would be surprised at how transparent supplies are to noise on the AC power lines.
AC power leakage to chassis is another factor, but I don't think that this is the whole story.
The power and signal cables that I have made cause a distinctive clarity and solidness in all systems that I am unable to get any other way.
"The sounds that cause windows rattling must be something that is measureable with microphones etc. You might even be able to record the difference. Anyway, though it's probably not categorically true, maybe there is too much focus on single component testing. Although, if your system does change sound with the power lead, it still seems that there would be a test that would reveal how it differs from a standard lead."
Yes, I tried increasing the volume by 1 dB without the lead installed but this did not excite the windows.
I do expect that a suitable microphone and analysis software would tease out the differences over a suitable time span.
My leads do not alter frequency responses per se, but they do alter dynamic behaviours interestingly and very pleasantly.
"Anyway, at this time, I mostly wish to point out that someone in addition to Millwood appreciates your forum contribution."
Yes, thanks again - it seems that Mr Millwood has some kind of preocupation with statistics, such that it precludes him from noting or understanding facts when presented to him. |
|
|
| fdegrove |
Hi,
| quote: | | it seems that Mr Millwood has some kind of preocupation with statistics, such that it precludes him from noting or understanding facts when presented to him. |
His friends are into hip-hop but he's into "folk".....
Sorry, Belgian joke...:D |
|
|
| millwood |
| quote: | Originally posted by mrfeedback
it seems that Mr Millwood has some kind of preocupation with statistics, such that it precludes him from noting or understanding facts when presented to him. |
so that it precludes me from buying into bogus "theories" of yours based solely on correlation. |
|
|
| fdegrove |
Hi,
Correlation and cause and effect aren't the same thing...
Eric showed causality.
Cheers,;) |
|
|
| millwood |
| quote: | Originally posted by fdegrove
Correlation and cause and effect aren't the same thing...
Eric showed causality.
Cheers,;) |
thanks for pointing that out for Eric.
Yeah. Don't you wish sometimes you could call things the way you want?
Too bad you could only do that in the cyberspace, not in the real world, :). |
|
|
| mrfeedback |
| quote: | Originally posted by millwood
so that it precludes me from buying into bogus "theories" of yours based solely on correlation. |
There are no theories presented - only observation of fact.
Eric. |
|
|
| fdegrove |
Hi,
| quote: | | Too bad you could only do that in the cyberspace, not in the real world, . |
Seems its real enough for him, so?
Cheers,;) |
|
|
| Steve Eddy |
So Frank, y'ever get hold of Dr. De Ceuninck? So far my EMail to him remains unanswered.
se |
|
|
| fdegrove |
Hi Steve,
| quote: | | So Frank, y'ever get hold of Dr. De Ceuninck? |
Not yet...
In Europe people usually take most of January to either visit or receive relatives for New Years wishes, etc.
So no, I've not even called him yet but I will.
As I now have both his work and private numbers back again I expect to call anytime soon.
Cheers and my best whishes for 2004, ;) |
|
|
| Steve Eddy |
| quote: | Originally posted by fdegrove
Hi Steve,
Not yet...
In Europe people usually take most of January to either visit or receive relatives for New Years wishes, etc.
So no, I've not even called him yet but I will.
As I now have both his work and private numbers back again I expect to call anytime soon. |
Okie doke.
| quote: | | Cheers and my best whishes for 2004, ;) |
Thanks. And the same to you and yours.
se |
|
|
| Steve Eddy |
Just received Bruno's measurements of the companion set of cables I'd sent to John.
He measured them at two different levels and two different source/load impedances.
Here they are at 30mV, 20 ohm source, 100k ohm load.
Freebie cable:
Old Radio Shack Gold:
New Radio Shack Gold:
RG-174:
Here they are again at 30mV, but with a 600 ohm source and 600 ohm load.
Freebie:
Old Radio Shack Gold:
New Radio Shack Gold:
RG-174:
Here they are at 13V, 20 ohm source, 600 ohm load.
Freebie:
Old Radio Shack Gold:
New Radio Shack Gold:
RG-174:
And finally, at 13V, 600 ohm source, 600 ohm load.
Freebie:
Old Radio Shack Gold:
New Radio Shack Gold:
RG-174:
Again, no signs of the high-order harmonics that John says he saw in the set that I sent him.
se |
|
|
| johnferrier |
Hi Steve,
Thanks for the update. Things look clean. And I remember something about John indicating that the RG-174 measured better than the others (except his reference cables).
Definitely, don't ask anyone to do this at this point, but silly me liked the idea of measuring a coat hanger (and I guess a second coat hanger for return). I'll use my imagination on this...uhmmm coat hangers look a lot like the above.
And just wondering, off topic, have you used your screen name Koyaan I. Sqatsi lately? (For others, this is another forum.) Just wondering how low a profile you're keeping these days. I found something about a quad-comp that was interesting.
JF |
|
|
| Pjotr |
Shhhh… finally, over 200 posts later. Nice work Steve, but I am missing a “null” measurement that shows the residual distortion of the AP itself :cool: Anyway it speaks for itself.
;) |
|
|
| SY |
| quote: | Originally posted by Pjotr
Shhhh… finally, over 200 posts later. Nice work Steve, but I am missing a “null” measurement that shows the residual distortion of the AP itself :cool: Anyway it speaks for itself.
;) |
Well, since you have to use a piece of wire to connect the input to the output, I think you actually ARE seeing the residual (the change and consistency with impedances is suggestive). The way to be certain is to halve the length of the wire and see if the distortion products scale similarly.
Thanks to Bruno for doing this work, and thanks to Steve for setting this up and getting wire to everyone! |
|
|
| Steve Eddy |
| quote: | Originally posted by johnferrier
Thanks for the update. Things look clean. And I remember something about John indicating that the RG-174 measured better than the others (except his reference cables). |
John only said that the RG-174 measured the best of the cables I sent him, not necessarily better than everything else he had measured other than his reference cables.
| quote: | | Definitely, don't ask anyone to do this at this point, but silly me liked the idea of measuring a coat hanger (and I guess a second coat hanger for return). I'll use my imagination on this...uhmmm coat hangers look a lot like the above. |
Hehehe. The problem with doing a coat hanger is termination. Steel doesn't solder well so you'd have to weld it to the connectors somehow. Or use some sort of compression contact.
| quote: | | And just wondering, off topic, have you used your screen name Koyaan I. Sqatsi lately? (For others, this is another forum.) Just wondering how low a profile you're keeping these days. I found something about a quad-comp that was interesting. |
Only place I use Koyaan I. Sqatsi is over on head-fi.org. Haven't posted there recently.
Not keeping any sort of profile. Just hasn't been much that I've found interesting lately.
se |
|
|
| Steve Eddy |
| quote: | Originally posted by Pjotr
Shhhh… finally, over 200 posts later. Nice work Steve, but I am missing a “null” measurement that shows the residual distortion of the AP itself :cool: Anyway it speaks for itself. |
Here's a residual from the AP brochure. I believe it was done at 2 volts, but don't recall what the source and load impedances were.
In any case, as you say, the plots ultimately speak for themselves. |
|
|
| millwood |
so what does this tell us?
a: our golden ear friends can hear distortions down to -150db?
b: our golden ear friends can hear improvements down to -150db?
c: it is sufficient if i heard it, you heard it and your buddies heard it?
d: none of the above?
thanks SE for all the work. |
|
|
| Steve Eddy |
| quote: | Originally posted by millwood
so what does this tell us?
a: our golden ear friends can hear distortions down to -150db?
b: our golden ear friends can hear improvements down to -150db?
c: it is sufficient if i heard it, you heard it and your buddies heard it?
d: none of the above? |
D. None of the above.
The purpose of doing these measurements had absolutely nothing to do with hearing or not hearing anything.
| quote: | | thanks SE for all the work. |
The thanks go to Bruno. All I did was buy some wire and pay some postage.
se |
|
|
| Pjotr |
Hi,
Such a null measurement should be done with the same source and load resistance as the actual cables measured but without the DUT (device under test). But is doesn’t matter much, I am quite sure the result will be the same as with a cable. I just mentioned it for completeness. Thanks Steve for all the work and also a lot of thanks to Bruno for all his work. Until now I didn’t see such an extensive cable test.
Cheers ;) |
|
|
| aw_dee_o |
If there has been anything learned throught his, it should be that the characteristics of the cable are somewhat non-linear and vary with amplitude of the signal fed into them. Sean
> |
|
|
| NEAR_SOTA |
Thanks for posting results.
As mentioned it looks rather clean and there would not be any discernable listening diifferences in cables based on the results shown unless you have bat ears which BTW I bet there are those who would say they have.
It does not change my opinion that Ratshack cables SUCK!LOL
I do use them in non critical area's and when I need an affordable cable without going a long way to travel.
Thanks SE & Bruno for the time it took putting this together.
Larry |
|
|
| Steve Eddy |
| quote: | Originally posted by NEAR_SOTA
As mentioned it looks rather clean and there would not be any discernable listening diifferences in cables based on the results shown unless you have bat ears which BTW I bet there are those who would say they have. |
Again, the purpose of these measurements has absolutely nothing to do with listening differences. That's a whole other issue which simply isn't going to be resolved just by measurements.
The purpose was simply to get at the truth with regard to harmonic distortion and try and duplicate John's results. That's it.
| quote: | | It does not change my opinion that Ratshack cables SUCK!LOL |
Hehehe. Even if the Radio Shack Golds were the best sounding cable I'd ever heard I'd hate them !@#$% things. Those ViceGrip RCAs are just horrid.
| quote: | | I do use them in non critical area's and when I need an affordable cable without going a long way to travel. |
I don't even use the Radio Shack Golds for purely utilitarian purposes anymore. For that I just use the cheapies with the molded plugs.
| quote: | | Thanks SE & Bruno for the time it took putting this together. |
You're welcome.
I guess Philips deserves some thanks too seeing as I'm assuming the System Two Cascade belongs to them. :)
se |
|
|
| fdegrove |
Hi,
| quote: | | That's a whole other issue which simply isn't going to be resolved just by measurements. |
Probably not by those measurements made but I'm hoping to get some measurements from my friend that are as conclusive as he claimed them to be.
| quote: | | Those ViceGrip RCAs are just horrid. |
Don't know those but, yes, connectors are very important.
That's why you use Redels and I happen to use Lemos, basically the same thing...
| quote: | | I guess Philips deserves some thanks too seeing as I'm assuming the System Two Cascade belongs to them. |
Yeah... "Let's make things better", that's their slogan over here...
Whether or not they actually do is another story. :D
Cheers,;) |
|
|
| Christer |
I have deliberately stayed out of this part deux thread so far
and have only skimmed through it briefly and thus may have
got things wrong. However, it seems quite clear from Brunos
plots that there the possibly visible differences between cables
are so small that they can probably discarded as measurement
errors. On the other hand, I seem to understand that John
did measure clear differences in distorsion between the same
cables on his setup. From a purely scientific point of view we
cannot really make any valid conclusion from this. We have
two conflicting results, and no clear evidence which one is
flawed of if both are. I am not familiar with any of the equipment
used and so cannot comment on which one should be most
reliable. Correct me if I am wrong, but it get the impression
that although John uses equpment he is very familiar with he
had to use it in a non-familiar way, which introduces a lot
of uncertainty factors. However, that is still no reason to
make draw a definitive conclusion from the results. That would
take further measurements by others, preferably on copies
of both equipemnts used and also analysis by experts on
measurement technology of the two systems used.
Note that I am not referring to what seems reasonable to
conclude, not what is reasonable to believe. |
|
|
| fdegrove |
Hi,
[QUOTE]From a purely scientific point of view we cannot really make any valid conclusion from this. QUOTE]
Couldn't agree more....
People seem to jump to conclusion though...
They shouldn't.
Cheers,;) |
|
|
| Christer |
Yes, experiments have to be repeatable and preferrably on
different equipment to rule out systematic errors. So far
neither result has been repeated by others. Of course, one
has to be more careful with repeating null results, since it
is very easy to get one by using equipment withouth the
proper capability to determine a non-null result if there is one. |
|
|
| Steve Eddy |
| quote: | Originally posted by fdegrove
Probably not by those measurements made but I'm hoping to get some measurements from my friend that are as conclusive as he claimed them to be. |
Conclusive of what exactly?
| quote: | Don't know those but, yes, connectors are very important.
That's why you use Redels and I happen to use Lemos, basically the same thing... |
Yeah. Except the Redels only use metal for the contacts.
se |
|
|
| fdegrove |
Hi,
| quote: | | Conclusive of what exactly? |
Huh?
| quote: | | Except the Redels only use metal for the contacts. |
Do they?
To the best of my knowledge, so are the Lemos or Fisher-Camacs...
At least the ones I use...
So, are we going to argue about those or just wait for conclusive measurements done on something more powerful than an AP?
Cheers,;) |
|
|
| Steve Eddy |
| quote: | Originally posted by fdegrove
Huh? |
You said you were hoping to get some measurements from your friend that you said were as conclusive as he said they were. I asked conclusive of what?
| quote: | Do they?
To the best of my knowledge, so are the Lemos or Fisher-Camacs...
At least the ones I use... |
Don't know which ones you use, but all of the Lemo and Fischer CAMACs that I've seen used metal for the shells, housings, etc. Most everything except the insulation. Which models use metal only for the contacts?
| quote: | | So, are we going to argue about those or just wait for conclusive measurements done on something more powerful than an AP? |
Again, conclusive of what exactly?
se |
|
|
| FLZapped |
| quote: | Originally posted by aw_dee_o
If there has been anything learned throught his, it should be that the characteristics of the cable are somewhat non-linear and vary with amplitude of the signal fed into them. Sean
> |
Really. And where do you get this idea from?
-Bruce |
|
|
| SY |
Steve, were all these cables the same length?
Well, we now have an idea as to how the residual noise and harmonics of the AP vary with source Z and load! |
|
|
| janneman |
| quote: | Originally posted by SY
Steve, were all these cables the same length?
Well, we now have an idea as to how the residual noise and harmonics of the AP vary with source Z and load! |
As far as I remember, System Two can be configured to internally connect gen out to analyzer in, effectively bypassing the external WUT wire (OK, there will probably be SOME piece of wire, internally). If Bruno could do the same level and termination tests with the internal connection, that would give us the residual of the system in each case. John might be induced to do the same.
/Speculation mode ON/
If we find that both systems measure the same without or with a WUT, that would be a strong indicator that the WUT has distortion products below both pieces of equipment.
/Speculation mode OFF/
That would still leave the explanation of the fact that John measures differences between WUTs. One interesting test for John would be to do a series of tests with a single WUT but with different cable runs/orientations, different times of day, with or without the TV etc on. IOW, John would be asked to use all his considerable mental and physical facilities to prove that he measured his environment instead of his cables. This is asking a lot, but is also the stuff true scientists are made of.
/Speculation mode ON/
If John would measure differences under these circumstances, that would be another strong indicator. If he would measure the same in each of these tests, we still have a problem to solve.
John, are you game?
Jan Didden |
|
|
| mrfeedback |
| quote: | Originally posted by Steve Eddy
Again, no signs of the high-order harmonics that John says he saw in the set that I sent him.
se |
Ummmm, when I spend 10 minutes scrolling back and forth between the graphs of all the loading and cable differences, I see quite substantial differences between both cable type and source impedence/loading.
This to me gives the first question - How are these graphs derived ? - by that I mean how are averaging times and peak hold/release times etc filtering the displayed test results ?.
According to the published graphs, there are variations, and some are quite substantial in the order of 1-2 db.
I regard changes like 1 dB as strongly substantial - in a live system changes of LESS THAN HALF a dB on a FOH or stage monitor graphic eq fader can make seriously substantial differences to the overall tonality of a system.
There is one absolute variable that is not being tested for here and that is 'possibly measurable' possible directionality in all these cables.
Q - If each cable was measured twice (or many more trials) in opposing directions, could this seperate out a directional bias factor, or isolate other differences if enough trials are measured ? - IOW could more experimental trials statistically improve resolution (S/N) of this AP test set ?.
IOW, could enough correctly conducted trials statistically separate out very fine differences according to cable type and direction ?.
The graphs show substantial variations - is this an artifact of the measuring equipment and method ......... or not?
Eric. |
|
|
| Christer |
| quote: | Originally posted by janneman
As far as I remember, System Two can be configured to internally connect gen out to analyzer in, effectively bypassing the external WUT wire (OK, there will probably be SOME piece of wire, internally).
|
But if I understand you right, we also bypass the connectors
in that case, and they should probably be considered part of
the test equipment since we are interested in the wires, or?
There are always potential sources of errors whatever one
does. Maybe one should do both what you suggested and
do a test with the shortest possible piece of solid cupper wire,
or whatever, between the connectors? |
|
|
| SY |
Scaling is one way to isolate it to the wire. If you double (or halve) the length, what happens to the distortion spectrum? This is one of the things I want to try when I go visit John Curl.
(Hey, John, I've been trying to call you for days- turn off the FAX machine!) |
|
|
| Steve Eddy |
| quote: | Originally posted by SY
Steve, were all these cables the same length? |
No. The freebie was about 8 feet, the old and new Radio Shack Golds were 12 feet, and the RG-174 was about 3 feet.
| quote: | | Well, we now have an idea as to how the residual noise and harmonics of the AP vary with source Z and load! |
:)
se |
|
|
| Steve Eddy |
| quote: | Originally posted by janneman
As far as I remember, System Two can be configured to internally connect gen out to analyzer in, effectively bypassing the external WUT wire (OK, there will probably be SOME piece of wire, internally). If Bruno could do the same level and termination tests with the internal connection, that would give us the residual of the system in each case. John might be induced to do the same. |
The AP can do that.
But the purpose of these measurements was to try and verify the high levels of high order distortion that John has been measuring in order to determine whether that distortion was being produced by the cables themselves.
We already know that the System Two Cascade is measuring more than 20dB below where John's measuring and there are no signs of distortion at the levels John is measauring distortion.
So I don't see what purpose measuring system residuals would serve.
| quote: | /Speculation mode ON/
If we find that both systems measure the same without or with a WUT, that would be a strong indicator that the WUT has distortion products below both pieces of equipment. |
But again, the purpose of these measurements was to verify the levels of distortion that John was measuring in order to determine whether or not they were being produced by the cables themselves.
| quote: | /Speculation mode OFF/
That would still leave the explanation of the fact that John measures differences between WUTs. One interesting test for John would be to do a series of tests with a single WUT but with different cable runs/orientations, different times of day, with or without the TV etc on. IOW, John would be asked to use all his considerable mental and physical facilities to prove that he measured his environment instead of his cables. This is asking a lot, but is also the stuff true scientists are made of. |
I'll leave you to ask him to do that. :D
se |
|
|
| Steve Eddy |
| quote: | Originally posted by mrfeedback
Ummmm, when I spend 10 minutes scrolling back and forth between the graphs of all the loading and cable differences, I see quite substantial differences between both cable type and source impedence/loading.
This to me gives the first question - How are these graphs derived ? - by that I mean how are averaging times and peak hold/release times etc filtering the displayed test results ?.
According to the published graphs, there are variations, and some are quite substantial in the order of 1-2 db. |
The plots use time domain synchronous averaging. 256 averages in this case. And like any averaging, there will be small variations between runs. If you flip a coin 100 times and note the number of times it comes up heads and how many times it comes up tails, if you flip the same coin another 100 times, you may get a different tally than the first time.
| quote: | | I regard changes like 1 dB as strongly substantial - in a live system changes of LESS THAN HALF a dB on a FOH or stage monitor graphic eq fader can make seriously substantial differences to the overall tonality of a system. |
Great. But what you seem to be overlooking here is that any of the changes going on here are going on at greater than 130dB below the fundamental.
| quote: | | There is one absolute variable that is not being tested for here and that is 'possibly measurable' possible directionality in all these cables. |
It wasn't tested for because that's not what the purpose of the measurements were.
| quote: | Q - If each cable was measured twice (or many more trials) in opposing directions, could this seperate out a directional bias factor, or isolate other differences if enough trials are measured ? - IOW could more experimental trials statistically improve resolution (S/N) of this AP test set ?.
IOW, could enough correctly conducted trials statistically separate out very fine differences according to cable type and direction ?. |
Ben Duncan already did some measurements with regard to directionality. But like some of his other measurements, they weren't very rigorous and the differences in directionality he measured aren't conclusive in terms of establishing that the directionality is due to the cable itself.
| quote: | | The graphs show substantial variations - is this an artifact of the measuring equipment and method ......... or not? |
Dunno. What they don't show is any high order distortion products at levels John has been measuring which was the sole purpose of these measurements in the first place.
I think I've asked more than enough of Bruno already so if you want anything else done, ask him yourself.
se |
|
|
| mrfeedback |
"The plots use time domain synchronous averaging. 256 averages in this case. And like any averaging, there will be small variations between runs. If you flip a coin 100 times and note the number of times it comes up heads and how many times it comes up tails, if you flip the same coin another 100 times, you may get a different tally than the first time."
So in other words these graphs are not correlated by enough trials to derive meaningful data.
"Great. But what you seem to be overlooking here is that any of the changes going on here are going on at greater than 130dB below the fundamental."
I'm saying that changing the level by 0.5 db over a 1/3 octave bandwidth can be clearly audible - that is a very small proportion of the overall signal.
"It wasn't tested for because that's not what the purpose of the measurements were."
I'm saying that I reckon that the cables need to be tested multiple times in both directions to filter out possible variation due to direction.
"Ben Duncan already did some measurements with regard to directionality. But like some of his other measurements, they weren't very rigorous and the differences in directionality he measured aren't conclusive in terms of establishing that the directionality is due to the cable itself."
I heard the effect plain as day on a friends system the other night.
That none of us has measured directional variation conclusively leads me to think that the testing is not precise enough to be conclusive.
"Dunno. What they don't show is any high order distortion products at levels John has been measuring which was the sole purpose of these measurements in the first place."
At this stage I don't think that these results are sufficient to fully prove or disprove JC's measurements.
The graphs given by Bruno are all different, and it seems that some of these differences are due to loading of the AP test set - IOW is the AP test set really fully up to the task?
Modern radio astronomy is reliant on data sifting to pick out signals BELOW the noise floor - this is not being done in this mode of testing presented here, but is required imo.
Fred did say a while back that he has irrefutable evidence for directionality in digital interconnects - any level of directional characteristic would imply a harmonic distortion behaviour in my understanding.
Eric. |
|
|
| janneman |
| quote: | Originally posted by Steve Eddy
The AP can do that.
But the purpose of these measurements was to try and verify the high levels of high order distortion that John has been measuring in order to determine whether that distortion was being produced by the cables themselves.
We already know that the System Two Cascade is measuring more than 20dB below where John's measuring and there are no signs of distortion at the levels John is measauring distortion.
So I don't see what purpose measuring system residuals would serve.
But again, the purpose of these measurements was to verify the levels of distortion that John was measuring in order to determine whether or not they were being produced by the cables themselves.
I'll leave you to ask him to do that. :D
se |
Steve, IIRC the original purpose was to determine whether cable distotion existed (could be measured) at all. The focus then went to the reliability of John measurement tools. Now if with the AP we can look at the residuals we may be able to answer the original question (with some caveats), kill two cables with one stone, so to say.
Jan Didden |
|
|
| SY |
| quote: | | Modern radio astronomy is reliant on data sifting to pick out signals BELOW the noise floor - this is not being done in this mode of testing presented here, but is required imo. |
On the contrary- that's precisely what's being done here. We call that "signal averaging." At 256 samples, you're reducing the residual noise floor to pick out signals by a factor of 16. And the residual floor with that number of averaged samples is an order of magnitude below what John is reporting so is certainly more than adequate to see any phenomena at the level he reports.
The results don't scale with length, though that should be repeated (and I hope to do so with John's setup) with the same cable at different lengths. |
|
|
| Pjotr |
| quote: | Originally posted by SY
On the contrary- that's precisely what's being done here. We call that "signal averaging." At 256 samples, you're reducing the residual noise floor to pick out signals by a factor of 16. And the residual floor with that number of averaged samples is an order of magnitude below what John is reporting so is certainly more than adequate to see any phenomena at the level he reports.
|
To add to that SY, what sets the noise floor is measurement bandwidth. Concerning the AP it is sample frequency versus FFT length. Although the kind of windowing plays also a role (makes the effective measurement bandwidth somewhat wider in general).
The advantage of synchronous measuring AND averaging is that it lowers (averages out) all non-signal related noise, including mains hum. In the case of 256 averages indeed 16 times. It has a same effect on noise floor as extra narrowing of measurement bandwidth.
Cheers ;) |
|
|
| Bob Wortman |
I followed a link on AA to this thread and ended up lost of over an hour reading this and other previous ones on the same topic. I really don't understand the passion and anger aroused by whether or not miniscule distortions do or don't exist in cables except for the fact the no one likes to be wrong and will defend their position even in the face of evidence to the contrary. I personally don't believe wires cause non linear distortions but I haven't spent any time trying to prove it. Most if not all DBT has shown that cable effects are insignificant or do not exist at the level of audibility. I had a heated discussion at AA with Mr Curl about Bybee devices that ended up with agreeing to disagree and respecting each others position. Anytime we are measuring near the limit of the measuring equipments capabilities the measurements are suspect and must be taken carefully and repeated by more than one person's setup to be considered factual IMO. I think it is entirely possible that John's measurements reflect distortion in the test gear that are influenced by the cable loading and appear to be in the cable. It is also possible that his measurements are real. At no time in human history have we possessed the complete knowledge of anything. All passive devices exhibit behavior that departs from the theoretical ideal. It is unlikely that wires are different. It is also likely that current knowledge about such behaviors is not complete. Whenever I refuse to believe in the possibility of things that are contrary to my current understanding, I have stopped learning.
"There is a principle which is a bar against all information, which is proof against all arguments and which cannot fail to keep a man in everlasting ignorance -- that principle is contempt prior to investigation."
H. Spencer |
|
|
| SY |
| quote: | | I think it is entirely possible that John's measurements reflect distortion in the test gear that are influenced by the cable loading and appear to be in the cable. |
Hi, Bob! Thanks for putting in your thoughts. FWIW, I've chatted with John a few times about this and though he believes that what he's seeing is real, he admits this possibility also. That's a reasonable open-mindedness. |
|
|
| Steve Eddy |
| quote: | Originally posted by mrfeedback
So in other words these graphs are not correlated by enough trials to derive meaningful data. |
Correlated to what?
| quote: | | I'm saying that changing the level by 0.5 db over a 1/3 octave bandwidth can be clearly audible - that is a very small proportion of the overall signal. |
Where has it been established that changing hte level by 0.5dB over a 1/3 octave bandwidth is clearly audible? And please, no amusing anecdotes.
| quote: | | I'm saying that I reckon that the cables need to be tested multiple times in both directions to filter out possible variation due to direction. |
What variation?
| quote: | | I heard the effect plain as day on a friends system the other night. |
Sorry, this thread isn't for empty claims and amusing anecdotes.
| quote: | | That none of us has measured directional variation conclusively leads me to think that the testing is not precise enough to be conclusive. |
And what has the fact that no one to date has yet demonstrated actual audibility of cable direction lead you to think?
| quote: | | The graphs given by Bruno are all different, and it seems that some of these differences are due to loading of the AP test set - IOW is the AP test set really fully up to the task? |
Yes, it's fully up to the task of measuring high order distortion products at the levels John has been measuring. In fact, it's more than up to that task as again, the AP's measuring a good 20 dB below where John's able to measure.
| quote: | | Modern radio astronomy is reliant on data sifting to pick out signals BELOW the noise floor - this is not being done in this mode of testing presented here, but is required imo. |
Stuart has already set you straight on this one.
| quote: | | Fred did say a while back that he has irrefutable evidence for directionality in digital interconnects - any level of directional characteristic would imply a harmonic distortion behaviour in my understanding. |
But this "irrefutable evidence" has yet to be presented.
Talk is cheap.
Wake me up when someone has something more than empty claims and amusing anecdotes to offer.
se |
|
|
| fdegrove |
Hi,
So, it is about audibility now?
Calm down, Steve....
Cheers,;) |
|
|
| Steve Eddy |
| quote: | Originally posted by janneman
Steve, IIRC the original purpose was to determine whether cable distotion existed (could be measured) at all. |
Nope. That wasn't the original purpose at all.
The original purpose dates back to February 2002 when I first saw John's plots of what he was claiming to be distortion produced by cables.
I'd never believed that cables couldn't produce any distortion at all at any level. What I had doubts of was that the distortion John was measuring, at the levels he was measuring it (higher than some amplifiers) was being produced by the cables as he claimed it to be.
When I questioned John's measurements, he told me:
Steve, you are an insolent fool. You have NOT shown anything! Either seriously attempt to duplicate my results, or find another sandbox to play in.
So what this has been about is seriously attempting to duplicate John's results.
While exploring possible cable distortions in general is interesting, that is not what either of these two threads which I started were intended to address.
So if we want to discuss that topic, let's open a different thread.
| quote: | | Now if with the AP we can look at the residuals we may be able to answer the original question (with some caveats), kill two cables with one stone, so to say. |
That's fine except for the fact that that wasn't the original question. :)
se |
|
|
| Steve Eddy |
| quote: | Originally posted by fdegrove
So, it is about audibility now? |
Actually what it's always been about is whether or not the distortion John's been measuring was being produced by the cables.
se |
|
|
| Christer |
From a scientific point of view (once again :) ), I am inclined to
agree with Eric that the test should ideally be done with cables
tried in both directions. Some people claim cables to be directed
so if that is actually the case, we cannot rule out the distorsion
being different depending on direction until we have measured
whether it is or not. The tests so far has probably not been
extensive enough to rule out the statistical possibility of Bruno
measuring all cables in the "right" direction and Johns measuring
them in "mixed" directions.
SY,
your previous comment on scaling seems reasonable, but it does
rely on the assumption that distorsion is proportional or at
least a monotonic function of cable length. Although it seems
reasonable to assume that, it may still be an assumption we
shoudln't make. Since we don't know what causes the distorsion
if it is there at all, we also don't really know if that assumption
holds. One possibility could be that the distorsion is caused
in the connector-cable junctions and different cables give
different distorsion here. |
|
|
| fdegrove |
Hi,
| quote: | | Actually what it's always been about is whether or not the distortion John's been measuring was being produced by the cables. |
As if I didn't know....
Vendetta Research sounds ironic all of a sudden....
Anyway, time will tell....eventually...
Cheers,;) |
|
|
| Steve Eddy |
| quote: | Originally posted by Christer
From a scientific point of view (once again :) ), I am inclined to
agree with Eric that the test should ideally be done with cables
tried in both directions. Some people claim cables to be directed
so if that is actually the case, we cannot rule out the distorsion
being different depending on direction until we have measured
whether it is or not. The tests so far has probably not been
extensive enough to rule out the statistical possibility of Bruno
measuring all cables in the "right" direction and Johns measuring
them in "mixed" directions. |
Huh? Even if John's measuring them in "mixed" directions, John's measuring distortion in ALL of the cables I sent him at levels more than 20dB above where the noise floor is in the AP plots. Higher even than the AP's third harmonic residual. So even if Bruno measured the cables in the "right" direction, if the distortion John's measuring even with the cables in the "right" direction is being produced by the cables, we should see high-order distortion products well above the AP's third harmonic residual.
se |
|
|
| Christer |
| quote: | Originally posted by Steve Eddy
Huh? Even if John's measuring them in "mixed" directions, John's measuring distortion in ALL of the cables I sent him at levels more than 20dB above where the noise floor is in the AP plots. Higher even than the AP's third harmonic residual. So even if Bruno measured the cables in the "right" direction, if the distortion John's measuring even with the cables in the "right" direction is being produced by the cables, we should see high-order distortion products well above the AP's third harmonic residual.
se |
Hm, yes you're right, but suppose John measured them all
in the "wrong" direction, then it could be the explanation.
Or maybe I am just too tired. Ought to have tucked in long
ago. |
|
|
| SY |
| quote: | One possibility could be that the distorsion is caused in the connector-cable junctions and different cables give different distorsion here.
|
Absolutely valid point. So, if true, that means that whatever the mechanism, it's not the wires. IOW, if it doesn't scale, it's some other factor, whether the test set, the plugs, the jacks, external noise, whatever. If it does scale, it's more likely the wires, assuming one does proper controls to exclude increased noise pickup from a larger or smaller loop.
Now go to bed! |
|
|
| Steve Eddy |
| quote: | Originally posted by Christer
Hm, yes you're right, but suppose John measured them all
in the "wrong" direction, then it could be the explanation.
Or maybe I am just too tired. Ought to have tucked in long
ago. |
What ho! what ho! this fellow is dancing mad!
He hath been bitten by the Tarantula.
--All in the Wrong.
se |
|
|
| fdegrove |
Hi,
| quote: | | Hm, yes you're right, but suppose John measured them all |
Nope...That's based on assumption albeit a possibly correct one.
When I mean directional here I mean cable that's directional by geometry, not just wire directionality.
How about the cables that were sent off for testing, are there any asymmetric ones involved?
Cheers,;) |
|
|
| Steve Eddy |
| quote: | Originally posted by fdegrove
How about the cables that were sent off for testing, are there any asymmetric ones involved? |
Nope. They were all symmetrical coaxial geometries.
se |
|
|
| mrfeedback |
"Where has it been established that changing the level by 0.5dB over a 1/3 octave bandwidth is clearly audible? And please, no amusing anecdotes."
The other night I did exactly this to the FOH 31 Hz graphic eq fader, and my colleage responded immediately by asking what I did - " That's sitting much better now ! - What did you just move ?"
Go and do some live sound mixing and you will find out quick smart.
While you are at it, try changing foldback eq settings by 0.5 dB and see what happens.
"What variation?"
I said 'possible variation'.
"Sorry, this thread isn't for empty claims and amusing anecdotes."
How about you try this for yourself - the fact that I have encountered this effect many times with speaker wires shows me that there is something going on at system amplifier/speaker level at least.
"And what has the fact that no one to date has yet demonstrated actual audibility of cable direction lead you to think?"
This leads me to believe that the appropriate tests have not been performed correctly yet.
I believe Frank, you and I are co-operating on this, but of course this is not possible until we get the sample set of wires/cables concerned - have they been posted to Frank and I yet ?.
"But this "irrefutable evidence" has yet to be presented."
I assume that this finding has shown up in the TDR testing that Fred and Jocko get into.
You should politely ask them to share what they have found.
The AP test set is showing load dependency according to the graphs.
These test loads are purely resistive, and not properly representative of a typical stereo system that is exhibiting changes according to cable type and direction.
Secondly sine-wave THD testing is a steady state single frequency tone and is not fully representative of dynamic and directional music waveforms.
"Talk is cheap."
And auto nay saying is beneath this - Steve, several of us are saying that we are hearing cable sonics and directional differences in systems.
That the testing described is not showing these up merely leads me to believing that the testing is not fully appropriate to show up these observed differences.
Eric. |
|
|
| Steve Eddy |
| quote: | Originally posted by mrfeedback
The other night I did exactly this to the FOH 31 Hz graphic eq fader, and my colleage responded immediately by asking what I did - " That's sitting much better now ! - What did you just move ?"
Go and do some live sound mixing and you will find out quick smart.
While you are at it, try changing foldback eq settings by 0.5 dB and see what happens. |
:yawn:
Amusing anecdotes are a dime a dozen, Eric. You're on no firmer ground than the amusing anecdotes of photographs in freezers.
| quote: | | I said 'possible variation'. |
Ok, possible variation of what?
| quote: | | How about you try this for yourself - the fact that I have encountered this effect many times with speaker wires shows me that there is something going on at system amplifier/speaker level at least. |
Great. So go start a thread about it instead of trying to hijack this one.
| quote: | | This leads me to believe that the appropriate tests have not been performed correctly yet. |
Great. So go start a thread about it instead of trying to hijack this one.
| quote: | | I believe Frank, you and I are co-operating on this, but of course this is not possible until we get the sample set of wires/cables concerned - have they been posted to Frank and I yet ?. |
Frank's had his for weeks now. As for yours, ask SY. He's the one who's sending them out to you.
| quote: | I assume that this finding has shown up in the TDR testing that Fred and Jocko get into.
You should politely ask them to share what they have found. |
It's your assumption. You ask them.
But what's it have to do with the subject of this thread?
| quote: | | The AP test set is showing load dependency according to the graphs. |
What load dependency are you talking about?
And what's it have to do with the subject of this thread?
| quote: | | These test loads are purely resistive, and not properly representative of a typical stereo system that is exhibiting changes according to cable type and direction. |
Exactly how are the loads any different?
And again, what's this have to do with the subject of this thread?
| quote: | | Secondly sine-wave THD testing is a steady state single frequency tone and is not fully representative of dynamic and directional music waveforms. |
So? What's that have to do with the subject of this thread? The distortion John was measuring he was measuring with steady state sinewaves.
| quote: | | And auto nay saying is beneath this... |
What auto nay saying are you speaking of? I've not done any naysaying, auto or otherwise.
And again, what's this to do with this thread?
| quote: | | - Steve, several of us are saying that we are hearing cable sonics and directional differences in systems. |
So what? That has absolutely nothing to do with this thread.
| quote: | | That the testing described is not showing these up merely leads me to believing that the testing is not fully appropriate to show up these observed differences. |
For the last time, these tests weren't intended to do anything other than try and replicate John's results. Which is what this thread is all about.
Please stop trying to hijack this thread. If you want to discuss those other issues, start a new thread.
se |
|
|
| johnferrier |
So Steve,
What remains to be done here? Does SY's examination of John's setup fit within the scope of this thread on Cable Distortion Measurements?
JF |
|
|
| Steve Eddy |
| quote: | Originally posted by johnferrier
So Steve,
What remains to be done here? Does SY's examination of John's setup fit within the scope of this thread on Cable Distortion Measurements? |
I suppose that could be considered some sort of "closure" in a peripheral sort of way. Though not something I've any particular interest in. My interest was in trying to determine whether or not cables were routinely producing distortion on the order of what John was measuring and attributing to the cables.
It's become rather apparent that they're not. Why John's 1700B is behaving the way it is and producing the levels of distortion that it is is John's problem, not mine.
se |
|
|
| mrfeedback |
There are multiple differences between those graphs, and saying "Again, no signs of the high-order harmonics that John says he saw in the set that I sent him." is simply not true.
One thing that GLARES at me out of those graphs is WHY are the 4 kHz and 6 kHz harmonics not present when there is harmonic activity at 2kHz, 3kHz, 5 kHz and 7 kHz an all the 13V graphs ????
I this to be expected for a perfect piece of wire ????,
or is this an artifact of the measurement sytem ????.
Eric. |
|
|
| Steve Eddy |
| quote: | Originally posted by mrfeedback
There are multiple differences between those graphs, and saying "Again, no signs of the high-order harmonics that John says he saw in the set that I sent him." is simply not true. |
It is true. The level of the high order harmonics that John has been measuring have been above -125dB relative to the fundamental. The AP measurements are measuring some 20dB below that (around -145dB) and there's no sign of any high order harmonics at -125dB.
| quote: | | One thing that GLARES at me out of those graphs is WHY are the 4 kHz and 6 kHz harmonics not present when there is harmonic activity at 2kHz, 3kHz, 5 kHz and 7 kHz an all the 13V graphs ???? |
There's 4th harmonic visible in the 13 volt plots. Not as evident in the jpgs as in the pdf though.
This behavior is rather characteristic of complimentary circuits which cancel even order distortion. You're just seeing the even order distortion products disappear into the noise before the odd order distortion products.
| quote: | I this to be expected for a perfect piece of wire ????,
or is this an artifact of the measurement sytem ????. |
It's just the residual distortion of the AP's generator.
se |
|
|
| Christer |
Morning guys, last night I posted late and was tired. Now it's
morning and I am tired, so maybe I am still wrong. :)
I am not sure what the supposed problem with my comment
on cable direction was. I don't believe the direction matters but
I cannot really know, so if we want to be scientific about this then
direction is a variable which should
be controlled, whether we believe it makes a difference or not.
On the other hand, if we want to be scientific there is a lot more
that would have to be done, having more people repeat the
measurements etc. etc. I am not saying we should or have to
do that, only that we cannot really know in a scientific manner
what the answer to the original question is. What we believe
is another thing. Although I don't know anything about the
test equipment used, the AP2 seems to be accepted as a
reliable piece of equipment so it seems reasonable to me to
assume that Brunos measurements are correct and John has
some measurement error, but that is a belief not a claim or
a truth. Anyway, let's hope SY gets a chance to do his
scaling experiments on Johns setup, then we will know more.
Eric,
correct me if I am wrong but it seems to me you are talking about
the audibility of an 0.5dB deviation in frequency response. I don't
think that has anything to do with audibility of 0.5dB variations
in the distorsion products. I am not claiming anything about
what is or isn't audible, just that there is no obvious connection
between audibility of the two things. It seems like apples and
pears to me. |
|
|
| janneman |
| quote: | Originally posted by mrfeedback
"Where has it been established that changing the level by 0.5dB over a 1/3 octave bandwidth is clearly audible? And please, no amusing anecdotes."
The other night I did exactly this to the FOH 31 Hz graphic eq fader, and my colleage responded immediately by asking what I did - " That's sitting much better now ! - What did you just move ?"[snip]
janneman: Wait a minute! We are talking about real minuscule, if at all, cable distortion. Even if your FOH stuff is objectively true, which is not proven, what does that have to do with cable distortion audibility, if any etc?
[snip]"And what has the fact that no one to date has yet demonstrated actual audibility of cable direction lead you to think?"
This leads me to believe that the appropriate tests have not been performed correctly yet.
janneman: Believe is the right operative word here. Another possibility, that I believe, if you will, is that nobody has demonstrated it because it isn't there.
[snip]
The AP test set is showing load dependency according to the graphs.
janneman: this is expected, as different levels and termination affects the noise level and therefor the dynamic range. If we would not see this, we would conclude that something is really wrong or the cables do have distortion and/or noise.
[snip] |
Jan Didden
(Hope I haven't messed up this text...) |
|
|
| johnferrier |
...but then beyond the possibility that John was always measuring in the wrong direction and Bruno was always measuring in the right direction, you've then got temperature, relative humidity, phase of moon, and you name it. This isn't a government funded "scientific" investigation. And in that case, the results sometimes become forever inconclusive. Maybe people think we are there (but then why waste more time and effort?).
However, for a diyAudio investigation, this seems to be as thoroughly investigated phenomena that I've seen (in my brief interaction with diyAudio). Maybe the topic of another thread, but if there are other examples of very thoroughly investigated subjects at diyAudio, I'd be most interested in a link or two...
One thing to remember is that, beyond the recent batch of four cables, there was also the batch of three cables in the prior thread. So the possibility of the all right/all wrong directionality goes down even further. Further, neither Bruno nor John make mention that they are aware of any directionality in there use of the equipment. Certainly, if John, Bruno, Charles, anyone else with an analyzer, a cable mfg, etc. has distortion measurements indicating directionality, then a new thread on measured directionality would swipe interest from Steve's thread here.
And as others state, I do agree that these measurements are separate from whether people perceive audible differences (again another thread). There may be other measurements that better flush out such phenomena. However, that is up to the people with a vested interest in directionality to show.
JF
*EDIT: errors* |
|
|
| john curl |
Folks, let's not quibble here. If Bruno is correct, perhaps there is something wrong with my measurement. I don't know what it is, but Sy is welcome to come by and advise me, or just check out what I do measure. If this is indeed correct then I would branch into another direction as to why the test equipment itself is sensitive to different cables. This would be just as important as the wires themselves, because the input stage of the ST is very much like any analog input stage. This would make many audio input stages suspect of being susceptible to the wire they are connected to.
For the record, and to reduce some speculation: I have recently put a .033 RT styrene cap at the output of the audio oscillator to ground, before the cable. This shorts out the 600 ohm source at high frequencies, and reduces the harmonic distortion contribution from the oscillator. It helps a little, but not much. The cable differences still remain. |
|
|
| mrfeedback |
Originally posted by john curl
Folks, let's not quibble here. If Bruno is correct, perhaps there is something wrong with my measurement. I don't know what it is, but Sy is welcome to come by and advise me, or just check out what I do measure.
Hi John,
Yes, it seems that a closer look is needed to work out what is going on, and causing these experimental differences.
If this is indeed correct then I would branch into another direction as to why the test equipment itself is sensitive to different cables. This would be just as important as the wires themselves, because the input stage of the ST is very much like any analog input stage. This would make many audio input stages suspect of being susceptible to the wire they are connected to.
SE excepted, there are a whole bunch of us (millions?) who observe audible differences according to cable type on typical systems.
Cable electrical parameters differences are accepted as the explanation, but I agree that output and input stages are part of this equation, and require closer investigation of their behavioral changes according to cable type.
I still feel that there is more than just this going on based upon experimentation with the modified cables that I have cooked up - I get the same sonic signature regardless of the cable type (different makes of balanced and unbalanced) that I am using, and the equipment connected.
Curiously this applies to power cables and audio multi-core cables also.
The magnitude of sonic difference may change acccording to cable or systems, but not the nature of the sonic difference.
For the record, and to reduce some speculation: I have recently put a .033 RT styrene cap at the output of the audio oscillator to ground, before the cable. This shorts out the 600 ohm source at high frequencies, and reduces the harmonic distortion contribution from the oscillator. It helps a little, but not much. The cable differences still remain.
So you at least have consistent measurement results.
I am interested to get to the bottom of this also.
Eric. |
|
|
|
