Such is the side effect of none face to face communication and that's why redundancy is a requirement.
Really, easier than L/R ? Besides, you can't mismatch lumped LCR and you acknowledge that TL and LCR models agree for audioband risetimes. Then you can't mismatch a TL in this context either........and this is absolutely true - no matter what 'terminates' the audioband cable its impedance appears at the source verbatim. It's meaningless and pointless to use terms such as 'mismatch' here, a conceptual mistake which you still don't get, JN.
@John: I guess that's not what I think of as "latency." It's looks more to me like construction of the signal via, in effect, the use of a basis set of sine waves. You might think that's easier that way than a lumped model as used in Greiner and Davis's papers, and you're certainly entitled to do so if it gives a correct answer, but I suspect most would find more ease and clarity from the simple lumped model- which also gives a correct answer.
As far as I can see, the only thing the lumped model doesn't account for is the constant 10-15ns (or so) delay because of the wire length. But that's constant and an entirely negligible one.
As far as I can see, the only thing the lumped model doesn't account for is the constant 10-15ns (or so) delay because of the wire length. But that's constant and an entirely negligible one.
Sorry, you won't be getting it from me, I'm not interested in this sort of thing 🙂 - I effectively short circuit any of these sort of behaviours by hardwiring and using decent cable - bigger fish to ...
Back to topic ... to recap, IME it's not the lifting, it's the separating that's key - lifting is just an easy way to achieve separation ... once the effect is heard of having separation vs. not having separation then it would be easy to experiment to see what parameters are critical.
Why the wife in the kitchen thing does work, is because she is hearing the overall patina of the sound, and the quality of that is what registers as changing - these "bizarre" sound altering mechanisms mainly effect this "patina". A good example - sorry, marce 😀 - is seeing the image of a working TV in the program on your TV - it's obvious when the colour balance of that filmed TV is wrong, there's typically a blue haze over the whole picture. Studios, as they should be, are vastly better - when they use a TV as a backdrop the colour correctness is close to being spot on. And a "blue haze" over the sound is what the wife hears, and when the "colour" of that changes it registers quite strongly .
Last edited:
I love how you and df keep on stating "you acknowledge", as if I've been saying something else. It's a deflection, a strawman you use.
I have never said the impedance is any different at the source, another statement of fiction on your part.
As I recall, you were attempting to measure the delay from one end of the cable to the other??? An obvious lack of conceptual understanding. As I also stated, the current at each end of the line can only be different by one transit time.
Honestly, learn what is being discussed.
And yet, you acknowledge that the t-line with mismatched ends matches the LCR model exactly.
You cannot have it both ways.
The step function shows the fastest rate the system can respond. With both ends severely below the line Z, the system can only respond with an exponential rise which as scott has demonstrated, is exactly equivalent to the LCR model he used.
Given that I've modelled using t-line, and produced exactly the same response a large scale LCR model shows, what seems to be the problem?
jn
Let's see, if I use normalized units, then the baseball travels in a potential field of V = Me/y. So -1/2del2Ψ - MeΨ/y = i∂Ψ/∂t. Solving the differential equation...
Do you really consider a simple model of an easy t-line example to be that difficult? It is trivial to make an excel spreadsheet that exactly produces the effects of a good LCR model.
As a side benefit, it shows what line impedance produces the minima.
If I modelled using a 200 element LCR, varying L and C to simulate many different cable inductances per foot, and then presented the output given various loads, it would be easy to show that when sqr(L/C) equals the load, the response delay is minimum. And would duplicate exactly what I've presented so far.
So, is it that the t-line model is so incredibly complex that it cannot be understood, or is it the fact that some cable vendor uses prop velocity in it's ad copy without any understanding of what they are talking about?
Seriously, how many people are going to accuse me of consorting with the enemy?? How many are going to continue to attack my understanding of the problem because the feeling is that I'm giving comfort to the enemy?
Personally, I don't live my life like that. If what I speak of can be construed or misconstrued to support the arguments of the "enemy", I do not care.
jn
Last edited:
Really impressive thread. So much opp for the big guns to show off their expertise post after post and boy did they deliver. 😀
Yup, why not ? Cable behaves as lumped inductance, naturally forms LP filter with load R, current has phase angle with source voltage - voila, there's measurable latency between ends of the cable for an audioband risetime......as I've shown.
I acknowledge that for small wavenumbers the cable does not behave as a TL so such terms are meaningless. That is why the models match .........'mismatched' does not mean anything.jneutron said:
There aren't really two ways......
I've done it, you're wrong for latency of audioband risetimes. Effect of C on latency is negligible, for realistic cases latency simply ~L/R where L is cable inductance R is load resistance, waveshape is conserved. Just do it, JN, you'll see.
More difficult than just calculating a simple lumped LCR (which I can solve with pencil and paper, no need to program a spreadsheet).
hmmm.. ok, let's see.
What is the difference in latency between a 1 ohm load and a 30 ohm load with a 20 foot long belden #12awg as used by lucky?
What real cable will minimize this difference?
I can see how daunting the math can be using transmission lines, so can understand why you, lucky, and df96 are so against it.
Addition, subtraction, multiplication, and division can be so confusing.. Especially that count by two's starting at one..
Luckily for me, I found this app that can assist me in that. Turns out, almost every phone has a built in calculator, who'da thunk. And, excel, man, just so confusing...It's to the point where I have to ask my grandchildren how to use it.
As to question number 2, which is the crux of the issue? 4 paralleled zip cords of #18 awg will serve nicely. That didn't even require I try to use that dang calculator thingy.....
Wow, can life get any easier?
Lucky, face it..your assertions have been toasted..thanks for the sims which proved my points.
jn
In my head, about 5uS + change for an audioband risetime. 1R & 30R is a bit extreme though. No need for calculator even.
Easy - a cable with lower inductance. Be it shorter, or having lower inductance per length.jneutron said:
Really ? It's pointless and misapplied.jneutron said:
Well IMO its hard to keep track of what your points might be, JN. But my simulations confirm what I've been posting all along if you read them. Great if you've come round to it. I don't get what stops you running your own simulations with audioband risetimes BTW, then you'd see. It's easy.jneutron said:
Last edited:
So, not only do you need to argue with Davis, you need to argue with Otala and Huttunen. Man, you just keep on disagreeing with everybody.
You first stated that there was NO latency. What was the word you used? 5 uSec was geologic? Yah, I remember your words as if they were from last week.
Oh wait, they were.
Now, you re-use the numbers I've been presenting since 2011 as "wow, I don't even have to use a pencil" for that one, it's so simple...
Interesting how far we've come, isn't it? Now, you've totally accepted that which you previously called "geologic".
You've progressed. Slowly, but you're getting there..
jn
- Status
- Not open for further replies.