Some people feel a need for blind listening tests.
I don't need it.
For one, I do know from a long time experience that I hear what I hear, not what I see, or think, or wish, or dream.
Second, most blind listening tests mask audible differences, thus they are useful only as a kind of a 'proof' that here are no audible differences between different components – even when those components measure differently to a great degree.
Please, stop this old and boring argue.
I'm sure i can provide more references about my "hearing" ability, than you can on your side.
(My 'Hearing ability' was the key for my living)
And i don't see how to be born in the country of Descartes or to had made some scientific studies can have a destructive effect on my tympanums.
Last edited:
You can't completely ignore R and G in a line, as they cause the attenuation which contributes to the bouncing pulse dying away (although terminations usually do most of the work). For most audio cables at most audio frequencies you can probably get a sufficiently good approximation by using just R, C and L. Instead of a sharp pulse bouncing, you can instead use the low frequency propagation parameters directly on the audio frequency components. For example, Z0 = sqrt(R/jwC) is close enough at low frequencies where R dominates over wL. (So Z0 = sqrt(R/2wC) x (1-j) ). Propagation speed is a bit more messy to calculate.
You are welcomed to hold on your views and beliefs.
The belief that some people have, that being born in a certain country make them superior to others, isn't new.
There is nothing unique about the way I hear.
Possibly, often the lack of trust in ones' hearing is a major barrier for discriminative listening.
Does it make one who was born in the same country as Descartes know any better, by the mere fact they were born in the same country?
If you'd like to discuss philosophies, I'm open to participate. Only, a different forum will be more suitable for such discussions.
As I said, inclusion of R and G slows down the response calculation. Losses in the line contribute to extending the response tail..it takes more transits before the load sees the current that is expected.
While claiming the model is insufficient is technically correct, you've not been able to state that the delay is not there...it is in fact worse than I've indicated.
edit: I use a step function because it includes all frequencies to DC. The longer you wait for the settling, the more LF content there is. Have you ever tried to measure a 10 microsecond delay in a 1Khz signal driving a 4 ohm load? By using a step function, all you have to do is look at the waveform to find the delay. Using a step function was taught me in engineering school back in '74.
jn
Last edited:
May i explain it very slowly ?
My text: "i don't see how to be born in the country of Descartes or to had made some scientific studies can have a destructive effect on my tympanums."
Means:
I don't see how using a methodical and scientific approach to analyse and try to understand some effect we can observe with one of our sense could destroy this sense.
Compute.
Last edited:
The delay is definitely there, I don't argue with that. I just regard it as irrelevant to audio. Although the delay is frequency-dependent, getting worse for lower frequencies, it is too small to do any harm because the line is short.jneutron said:
I'm not convinced that line attenuation extends the tail. By reducing the number of 'bounces' before the edge becomes too small it should shorten the tail. Perhaps not by much. In any case, the impulse method and the low frequency wave alternative must give the same answer for a low frequency signal if they are both correctly applied.
The line to load mismatch is frequency dependent. Meaning, at some frequencies there will be a match and no delay, while at other frequencies, it will happen. 10 uSec as a group delay is not an issue. As a differential, it is not dismissed so casually.
edit: getting worse with lower frequencies requires response. It is NOT the basic line impedance nor prop delay vs freq that I speak of....it is the response of the system as a consequence of the load impedance varying wildly from 1 or 2 ohms to hundreds. Speakers can be wild beasts in terms of impedance. What I discuss is the system delay as a consequence of that varying load impedance.
As I also mentioned, the terminal voltage of a speaker is v = LdI/dt + I dL/dt. The second term indicates a current dependent inductance.edit (remember, the second term component I've discussed is that of a 2Khz sine current in the VC, while it is travelling in the gap at a 20 or 50 hz rate at some large percentage of xmax..the eddy currents in the gap metal walls will be dependent on the velocity within the gap times the current within the VC. This means that the terminal voltage of the VC will NOT be as "orderly" as we would wish.
For a single full range driver, the hf impedance seen by the line will be velocity dependent and current dependent.
Then you will need another mechanism to explain how the load current rose so fast despite the transits required. Remember, each bounce contributes to the final value, if you attenuate the signal at each transit, the final value will take longer to achieve.
A many element RLCG model will produce the correct results. This has already been hashed on this very thread, scott ran a ten element.
As I asked, have you ever tried to measure a 10 usec delay of a 1Khz sine into a 4 ohm load?
You will find it is not so easy..Yes, you can get some waveform result, but you will not be able to trust it..
Bingo.
jn
Last edited:
It always surprised-me to see how people who live on their earing (instrument makers, enclosure makers, sound engineers, musicians, 'big ears' in sub marine boats etc...) are suspicious about their own sense, multiplying the experiences at various moments, various moods, ask other's confirmation etc.
While some others can be so confident.
Our listening lie on our attention and our memory, because when you hear a sound, it is already gone away.
Our earing is manipulated by our brain in several ways. We can focus our attention on some detail, filtering everything else. And we can even filter so much that our brain can correct a lot of obvious defects, or add imaginary ones.
Only reason why we can believe of any credibility in Hifi reproduction.
We are greatly fooled by our culture, as well. And more we focus, more we fool ourselves.
I could write hundreds of pages of funny stories about this.
I understand very well the position of sir Kgrlee about blind listenings.
Indispensable for those who believe so much in their hearing's feelings.
May-be less for those who don't ? ;-)
The low frequency equivalent of the constituent impulses bouncing back and forth is precisely given by the low frequency line impedance and propagation delay. You are speaking of it whether you realise it or not. As I said, the two methods when correctly used must give the same result as they use exactly the same physics.jneutron said:
Thanks very much for your help. I still don't see how this is a mechanism that wouldn't already show up as ordinary THD/IM. Maybe it's just over my head, but thanks anyway.
Chris
Um, I've been saying that for a rather long time now. As I said, we've already hashed this half a year ago.
And I said back then, it isn't possible to measure a 10 uSec delay on a 4 ohm load driven at 1Khz, nevermind 50 or 100 hz.
You can try. But you won't get accurate results. That's why I've developed current viewing resistors with equivalent inductance (low dB/dt trapping)below 100 picohenries.
So as I've said all along, the t-line model I've provided is sufficient to understand the issue. That you choose to fight it based on the lack of inclusion of R and G (which makes the delays even longer) is a moot point. I've stated multiple times, the model is providing a lower limit on the delay. Your arguments are leaning towards longer delays..I've stated that before as well.
jn
edit: A little clarification..a low dB/dt trapping current viewing resistor (CVR) is one which gets around the field created around the resistor as a result of current. If you try to simply measure the voltage across a .1 ohm resistor for example, you will measure both the IR drop across it as well as being affected by the rate of change of magnetic field (dB/dt) which surrounds the body of the resistor. This is a big concern when using low impedances and high rate of change currents. Three possibilities can solve this.
1. Return the viewing conductor through the geometric center of the resistor. There is no field at the geometric center of a symmetrical current. This solution doesn't reduce the cvr inductance however.
2. Return the current being monitored through the geometric center of the resistor. For example, a coaxial shell of resistors, with the return current through the center. The inductance can be easily calculated based on coaxial equations.
3. Run the out and back currents interleaved by using an array. Easiest to build btw... This can be modelled as a large quantity of pairs of current, all orthogonal by design, and is trivially scaleable.
Last edited:
- Status
- Not open for further replies.