Don't let me stop you ...!
I have not the experience to expertly comment on your points; however I do know that the ear/brain is extremely competent at sorting out the most subtle directional clues if the information is not too clouded by excess added distortion. Hence my approach is to try to reproduce the sound in the cleanest possible way, which, for me, certainly compensates for the possible "deficiencies" in speaker setups that you mention.
Also, something I was going to mention earlier, but forgot. Has a flash of inspiration: people have talked of "sterile" sound, usually of equipment that nominally, objectively performs very well, but I never quite groked the term. Holt in his audio glossary defines it as "Pristinely clean but uninvolving", which translates to: very low in audible distortion but boring. Well, I beg to differ, that apparent sterility is indeed a distortion, but a subtle, low level one. And it is precisely what the cable issues I've been talking about cause: realistic, involving sound steadily turns "sterile", slowly or quickly, depending upon everything, and that is one of the great scourges of audio replay ...
Frank
He didn't say that, and he hasn't. John refers to some published research which suggested audibility of 10us delays in localization. His area of expertise and accomplishment is running extremely high currents down wires, and trust me, he DOES know what he's talking about.
I'd still like him to generate some test signals digitally to demonstrate his hypothesis. I don't think he's correct, but the proof is in the listening.
Yes it does. But for high power amps, we have to care about power rating of such resistor. 40Vrms 300 ohm makes more than 5W. 40Vrms and 75 ohm makes more than 20W. And wirewound resistors are inductive.
Last edited:
Sorry for boring the rest of you. I started a conversation with jn, but then it turned into a reminder to others that you cannot terminate an audio cable. This is not esoteric, but it seems to be surprisingly not well-known.fas42 said:
Anyone offering you 8ohm speaker cables does not understand transmission lines as well as he thinks he does. It is a good example of a little knowledge being a dangerous thing. Ordinary people just think of speaker wires as wires - quasi-DC approximation. Then they hear of RF transmission lines and correct termination, so assume that this must be a good thing for audio. Unfortunately many textbooks don't give the full transmission line theory, only the high frequency approximation which is fine for RF. The full theory tells us that a typical audio cable does not behave like a simple transmission line at frequencies below about 5-25kHz (depends on cable details). Instead it has a rising and reactive characteristic impedance and slower speed at lower frequencies. It becomes dispersive, so phase and group velocity differ. All this was a major headache for telephone companies, but fortunately for us our wires are much shorter so we don't have to worry. It turns out (IMHO) that the quasi-DC approximation is good enough for the audio signal, but we might have to worry about RF behaviour for reasons of amplifier stability. People sometimes mix these two separate issues together, perhaps when they have a product to sell.
jn,
This discussion has turned into a waste of time. Those reading it are smart enough to make up their own minds.
Bob
Instead it has a rising and reactive characteristic impedance and slower speed at lower frequencies.
DF96 , I'm not at all a specialist of transmission lines, so, here, can-you provide figures of group delays ?
Looking square waves (my amp goes up to > 2Mz) i can see influence of the wires on the settling time or slew rate, and nothing else, and correctly terminated wire help to fasten the edges.
Of course, this is only interesting in regard with the degree of magnitude, comparing with our driver's performances, both on group delays and slew rates. (sarcastic smile)
DF96 , I'm not at all a specialist of transmission lines, so, here, can-you provide figures of group delays ?
Looking square waves (my amp goes up to > 2Mz) i can see influence of the wires on the settling time or slew rate, and nothing else, and correctly terminated wire help to fasten the edges.
Of course, this is only interesting in regard with the degree of magnitude, comparing with our driver's performances, both on group delays and slew rates. (sarcastic smile)
Fas42, on my point of view, math is NOT esoteric. "Sound of cables" is.
Last edited:
Well, I hear cable differences. My colleagues have been known to be VERY SENSITIVE to cables and such. All that I can really attest to is in the listening.
For cables to make any serious difference, the rest of the audio system has to be at just as high quality, or it is a waste of time. It would be like putting the highest quality tires on a VW Beatle. It would work, but the cost of the tires would not be offset by any real improvement. Put quality tires on a Porsche, and see what happens, especially if the tires on the vehicle were below Porsche recommendations for some reason. It can be the same with wires.
Personally, I don't like to do comparative listening tests with wires or just about anything else, but I have heard the differences when I attended a serious trial.
Who does these trials? OH, engineers, physicists and MD doctors in my general experience. The MD's because they are some of the VERY FEW people who can afford the best all the way down the line. I have met perhaps a dozen of them . They often became the CTC Blowtorch owners, not some engineer who would rather do it himself, or whatever.
For cables to make any serious difference, the rest of the audio system has to be at just as high quality, or it is a waste of time. It would be like putting the highest quality tires on a VW Beatle. It would work, but the cost of the tires would not be offset by any real improvement. Put quality tires on a Porsche, and see what happens, especially if the tires on the vehicle were below Porsche recommendations for some reason. It can be the same with wires.
Personally, I don't like to do comparative listening tests with wires or just about anything else, but I have heard the differences when I attended a serious trial.
Who does these trials? OH, engineers, physicists and MD doctors in my general experience. The MD's because they are some of the VERY FEW people who can afford the best all the way down the line. I have met perhaps a dozen of them . They often became the CTC Blowtorch owners, not some engineer who would rather do it himself, or whatever.
If you are looking at the edges of square waves then you will mostly be seeing the high frequency behaviour of the line, which follows the simple RF theory.
The characteristic impedance is fairly easy and has already been given in this thread by me and jn: Zo=sqrt( (R+j2pi f L)/(G+j2pi f C) ). When R dominates over wL (low frequency) the result is Z0=sqrt(R/(2pi f C)) x (1-j) - equal resistive and capacitive parts, scaling with 1/sqrt(frequency).
The propagation constant is more messy. It also involves a complex square root. I don't carry it in my head so I will have to refer you to the textbooks. It may be online somewhere. Speed slows at low frequencies. As it is frequency-dependent the phase and group velocities become different. However, all this is negligible when compared with what happens in loudspeakers.
My main point is: don't worry about it, but remember to smile when anyone offers you a 'matched' loudspeaker cable. A Zobel-like network to provide a good high frequency termination at the speaker might make sense for the immediate ultrasonic region where SS amps can sometimes come unstuck due to loop stability issues. Ordinary speaker cable could have an RF impedance in the 30-100ohms region, so a 30-100 ohms Zobel might be useful if the tweeter/crossover impedance rises sharply above the audio range.
The characteristic impedance is fairly easy and has already been given in this thread by me and jn: Zo=sqrt( (R+j2pi f L)/(G+j2pi f C) ). When R dominates over wL (low frequency) the result is Z0=sqrt(R/(2pi f C)) x (1-j) - equal resistive and capacitive parts, scaling with 1/sqrt(frequency).
The propagation constant is more messy. It also involves a complex square root. I don't carry it in my head so I will have to refer you to the textbooks. It may be online somewhere. Speed slows at low frequencies. As it is frequency-dependent the phase and group velocities become different. However, all this is negligible when compared with what happens in loudspeakers.
My main point is: don't worry about it, but remember to smile when anyone offers you a 'matched' loudspeaker cable. A Zobel-like network to provide a good high frequency termination at the speaker might make sense for the immediate ultrasonic region where SS amps can sometimes come unstuck due to loop stability issues. Ordinary speaker cable could have an RF impedance in the 30-100ohms region, so a 30-100 ohms Zobel might be useful if the tweeter/crossover impedance rises sharply above the audio range.
What is the difference between copper and silver, technically?
I have never measured cables but my silver cable sound different with my copper cable. Both products came from the same manufacturer. The build quality is very similar and I use 2m from amp to speaker, and I also use them inside speaker enclosures.
Inside my speaker I have decided to use copper on tweeter and silver on woofer. I couldn't decide which one I like more. The silver cable is like carbon composite, the copper cable is like metal film. I think I can understand why there is cable made from a mix between copper and silver.
I have never measured cables but my silver cable sound different with my copper cable. Both products came from the same manufacturer. The build quality is very similar and I use 2m from amp to speaker, and I also use them inside speaker enclosures.
Inside my speaker I have decided to use copper on tweeter and silver on woofer. I couldn't decide which one I like more. The silver cable is like carbon composite, the copper cable is like metal film. I think I can understand why there is cable made from a mix between copper and silver.
May-we suppose people who became CTC Blowtorch owners to be outrageous rich people, believing that more expensive, better it is ?
I do not make any criticism about the real quality of your preamp i never listened to, when you refuse to provide schematic, and when you refuse to provide measurements, so i'm just VERY suspicious about-it.
And, John, by the way, more you talk about burning cables, and black magic, more i'm suspicious about you.
At the end, the only thing I am sure about the Blowtorch, is that it is not worth the price, because its price is a nonsense to me.
This is not productive, on my point of view. Imagine a circuit witch will oscillate with a 100pf charge. Add a 100pf cable ? Conclude something about "hearing a cable sound difference" ?
Most of us agree that cables differences are based on the interaction of their electrical characteristic on the active devices around. And , when we want to tune something a cable can modify, it is better to do-it active devices side. So simple.
And, please, don't suppose we are all deaf and running poor systems. You could be VERY surprised, and may-be shamed.
Last edited:
I will find it for you, it has been a while.
Yabut, don't bend the fiber too much..tight bends and some of the bits may fall out, they are going pretty fast you know...
Sy below has it rather accurate. There is literature out there which actually provides 1.5 uSec testable and repeatable ITD discrimination out to 12 Khz with dither, and 5 uSec to about 1.5khz without. Not suggests, but also lateralization not localization.
I do not have the necessary materials to do this digitally.
I harped on you for several reasons (my apologies).
First, I decided last night to stop tiptoeing around the elephant in the room. You are one of participants here that I set the standards bar very high for as a consequence of my high regard for your expertise. My expectations of you do not include diversion or obfuscation.
Second. The t-line analysis I have presented is not rocket science. In point of fact, it was an analysis presented to the students in the second lecture of my E/M field theory course. Anybody who took this course and paid attention knows how to do this. Personally, the first two lectures were extreme eye openers for me, as they were my first introduction to signal propagation in wires. DF96 correctly points out the more refined t-line equations, but inclusion of those elements makes the problem WORSE, not better. IOW, it doesn't invalidate the result, but sets the result as a lower limit.
I have extended that analysis in several ways. A t-line will carry a signal at it's prop speed ONLY with the signal V/I relationship of the cable, and that propagating signal will have equal inductive and capacitive energy storage in the wake of the leading edge. A 100 ohm wire pair does not instantly know that the load is 8 ohm, so the signal travelling down the line cannot support 8 volts per amp, it supports 100 volts per amp.
Your "speaker tolerances are far more important and swamp any such delays" is either obfuscation on your part, or you simply do not understand. Read the next paragraph slowly please.
I have a pair of cabs with eminence delta pro 12's, selenium D205TI's, first order LP to woofer at 3Khz, 3rd order 5khz HP to tweeter, 5% tolerance all caps/inductors.....and when I run both cabs off the same amp channel, I "see" a central image which my brain interprets as the source of the sound. This despite absolutely NO tolerance matching, no driver matching, no phase concerns, no level matching..nothing. These speakers were slapped together with NO regard to any level of quality and are abused as PA speakers. And yet, they are absolutely capable of providing images despite having little engineering involved. The only engineering I used was the use of an overly tight cabinet such that the very small back volume helped prevent woofer overexcursion during a 145+ dB SPL pulse from a hydrogen/oxygen explosion. I say 145+ because it was measured 20 feet from the source, whereas the speakers were 10 feet away. And I made sure the amp (QSC RMX 1450) was on and connected so that it also stiffened the cone against the pressure wave.
You are welcome to run off claiming it's useless, I harped on you to give you that "out". Or, you could continue contributing. Your choice.
If you choose to continue, I recommend you learn enough t-line theory to understand the analysis. I unfortunately lost the testbook I used back in '74 or '75, perhaps somebody else can chip in with a reasonable text...certainly not Jackson or Becker...
Read above.
jn
Last edited:
There are other, more rational explanations.
The amplifiers driving these cables are intolerant of the load for one.
http://www.belden.com/docs/upload/Precision-Video-Cables-Part-1.pdf
If you look at page 1, they show a graph on the right hand side.
The article is written about coax because they make coax. The derivations still apply to parallel line t-lines, but some entities like capacitance and inductance require different equations due to geometry.
Note in the graph, which I assume is about a 75 ohm cable (can't tell exactly from the vertical log axis), at .001 Mhz, or 1Khz, the impedance is about 200 or 300.
My derivation assumes "75" constant, DF96 is trying to include the rise with LF.
The analysis I provide considers the load mismatch, so if the load is 4 and MY line is 75, lots of reflections will be required before the 4 ohm load sees the current the supply is supposed to provide. If we adopt DF96's numbers, the mismatch is no longer 75 to 4, but 250 to 4. That requires even more reflections. What is worse, if we frequency limit the signal, the transit speed of the line is too fast with respect to slew rate, so we cannot see the effect clearly. I already pointed out the difficulty of seeing a 10 uSec delay in a 1Khz signal especially in a low impedance system.
Use of a step transition to examine system time response is a standard engineering technique. I use it whenever I need to program or tune any of my motion control platforms.. With digital and analog circuits, it's not so clearcut as the hf component violates digital and can broadcast analog.. With wires and simple resistive circuits, it's great.
jn
Last edited:
- Status
- Not open for further replies.