| Columns Retired Columns & Blogs |
| Loudspeakers Amplification Digital Sources | |
Analog Sources Accessories Featured | |
Music Columns Retired Columns | |
Show Reports | |
Features Latest News Community | |
Resources Subscriptions |
© 2024 StereophileAVTech Media Americas Inc., USA
All rights reserved
Thank you for explaining this. Because this double standard strikes at the heart of my quarrel with these anti-audiophile crusaders. It is but one example that exposes the fact that they are driven by an agenda (one based in pseudo-scientific actions and beliefs), and not a love for high resolution music reproduction. This too is what eternally seperates audiophiles from pro-audio knob twiddlers, with neither a clue nor an interest in the pursuit of advanced high fidelity sound quality. Who whenever they go and mix with knowledgable audiophiles, become an abberation on our hobby. It is but one of many double standards I have witnessed from Arny Kreuger and his crew of pseudo-objectivists for at least 15 years. They include:
1) The sanctification of audio DBT's/ABX. They beat us over the head with the DBT litany, but when queried on how they evaluate their own equipment, I and others have often found it is rare that the DBT crowd actually perform DBT and ABX tests themselves. Let alone proper DBT and ABX tests. Let alone proper DBT and ABX tests that weren't conducted and/or evaluated by someone with an anti-high end agenda or belief system.
e.g. "Krabapple", an advocate of DBT/ABX tests and perhaps the most heinous of the Hydrogen Audio crew that recently set up shop here, who when asked recently whether he used DBT tests to evaluate his audio purchases, admitted he didn't even bother to listen to his hifi equipment before purchase! After all, the "religion" they belong to tells them it all sounds the same, so why waste time with silly listening tests? I have seen his story repeated many times over among the audio-objectivist parishioners. I call it "the self-fulfilling prophecy".
2) Preference is preference. But technical accuracy is what matters. As in JA's example, "technical accuracy" only matters when they are defending their anti-audio religion in an argument against a "non-believer". When confronted with either the lack of rigor in their own tests, or lack of technical accuracy in their own pursuits, then it switches to "a matter of preference".
3) Do as I say, not as I do. They claim we should take a "scientific", "objective" approach like they do, and not be swayed by looks, reviews, word-of-mouth, or even just listening to the equipment! Fine! Except that they are the least "objective" $#@!er's I have ever encountered, this side of a jihad. They are constantly making statements and drawing conclusions, whether it is about audio or about you, with absolutely no evidence but the evidence of their infinite prejudices. When pressed to provide evidence to support their assertions, most of the time they'll ignore you, and the rest of the time will say it's "common sense" or "logic". They "know" with their heart and read with their mind. But at the same time, will never hesitate to claim, " It's 'science' "! (reg. tm.)
I say if you're going to belong to an audio religion, at least belong to a religion that provides you with a hope of good sound. In the audio sector, "good sound" beats "good ideology" any day. Regressive, status-quo keeping ideologies may be good for political arguments on discussion forums and conferences, but historically, they have never been good for sound. If they were, the headache-inducing headrill sound of the first CD players, billed as "perfect sound forever", would still be with us today.
I think you have it upside-down, Mr. Sullivan. Arny Krueger even admitted recently on rec.audio.opinion that he deliberately trolls me in postings such as the one to which I was referring.
And _again_ you pretend that I ignore your questions, Mr. Sullivan! I offered you a response in message #70453, which I posted yesterday at 05:15 PM. Or are you going to claim again that you don't bookmark my postings so you could hardly be expected to remember what I wrote?
John Atkinson
Editor, Stereophile
Your name fetish has been asked and answered in the thread below, where you first repeated the "since I at last have your attention" phrase. So at this point, you are simply trying to be deliberately annoying, disrespectful and badgering JA from one thread to another with this. If you had any hopes of continuing to troll our fair forums, I advise you not to do that.
http://forum.stereophile.com/forum/printthread.php?Board=rants&main=70453&type=post
I note that in bringing your attacks from one thread into another, then back into the originating thread, you've gone from "mere" cross-forum bashing to cross-thread bashing on the same forum. That's some very sophisticated trolling, I must say. In any case, please stop "RAO-izing" every thread you foist yourself upon, Krabapple. If you miss your RAO days so much, note that the newsgroup is still there, waiting for your return.
I note that yours has been. But I guess that's okay when it comes to you. See my recent post about the double standards of the objectivist crowd.
I had *already* asked several times on THIS thread, that you chime in on the matter, with no response from you. I'd call that 'ignoring'. As for yesterday's post,I hadn't read that thead yet, when I posted here. Your post about JJ there -- which was noncommittal -- came AFTER you had already replied to another post of mine, on a different topic. I seized the opportunity to ask there, because you had been *ignoring* the question here.
*This*, after all, is the thread on which Frog and SAS are conducting their bizarre slander campaign against the man you know to be JJ -- charges you could refute in an instant but for some reason chose not to, either in that post or, so far as I've read, here.
Btw, a folie a deux takes two, and yours extends over years. I doubt anyone except you and Arny care at this point whether you or he are more guilty of 'trolling' for cross-forum responses to the other. Seriously.
I'm starting to get the idea that you simply don't know when to shut up. How long are you going to continue to insist on making a nuisance of yourself here, from thread to thread? Now you're making excuses for not having read JA's response to you on this matter, when you asked this question yesterday in the Lavorgna thread. But yet you're still demanding an answer from him in this thread, despite the fact that he directed you to his answer, and I went and gave you a direct link so you wouldn't have any excuses to hide behind. And you don't, you have been told that your question has been responded to. Oh you may not like your response, I can see that, but given your record of hostilities on this forum and that you hate everything it and JA stand for, why do you think anyone here cares whether you're a happy troll, or an unhappy troll? You should be truly thankful that your campaign of assault and abuse on this site has been tolerated for as long as it has.
You state this so-called "bizarre slander campaign" is relevant to this thread, so you're admitting you were crapping on the Lavorgna thread with this name fetish nonsense of yours, and had no call to raise it there. Furthermore, what you're calling a "bizarre slander campaign" was actually started long before this thread ever existed, by this "j_j" character (whoever it is), against me. When he demanded that people use their real names on the Stereophile forum, and accused me, without a shred of evidence as usual, of hiding behind a pseudonym like he and you do (and it's obvious why you both choose to hide behind pseudonyms. Both of you are cross-forum bashing trolls. That's what trolls do). But there's nothing "bizarre" or "slanderous" in someone pointing out the FACT that whoever is hiding behind the anonymous handle of "j_j" (and it appears has several other sockpuppet names he goes under, if he is who he says he is), has refused to state his name on this board. He has not even done so ONCE. Hence, the reason his identity will always remain in question, until he signs his real name and affiliation in his messages, as SAS does.
LOL! I'm more familiar with the exchanges between Krueger and JA than you are. And if you think JA's behaviour on the forums is in any way comparable to Arny's, that pretty much tells us how much your opinion is worth, and exactly where your head is lodged.
It is in example of what I was saying earlier, that those who claim the rational high ground, who claim to adhere to the logic of scientific deduction, like Mr. "Krabapple" Sullivan, can still be blinded to reality - in this case, Mr. Krueger's sick, unrelenting, trolling and antisocial behavior - by their prejudices and preconceptions.
John Atkinson
Editor, Stereophile
Can anybody dig up the Usenet posts in question?
Seriously. I'm not pulling anybody's chain. I'm all ears here.
Are you serious?? There are something like 10 years and 10,000 posts to wade through, if you're trying to compare JA's contributions with AK's. And the Krueger troll must have changed his email address at least half a dozen times, which makes it even more difficult to sort through his legacy. If you really don't have a clue as to who John Atkinson and Arny Krueger are, just ask anyone on forums outside of HA or any other Stereophile-bashing forum. Go and ask whoever's left on RAO about the differences between their behaviour. Or sort through the messages yourself, on Google Groups. JA's reputation is very well known and very well regarded. Arny? Well it's a miracle he hasn't been imprisoned yet. Over the length of his trolling career, and regardless of whether JA is present or not, Krueger has falsely accused him of everything short of global warming.
As always, the opinions you get in the audio community depend on who you ask. JA, as the figurehead of one of the most well known audio publications, has always been a lighting rod - attracting the world's most potent and mentally unstable anti-audiophile lunatics with internet access. All of whom appear to have been trained by Arnold Krueger. It really goes much farther than just "prejudice and preconception" on the part of these "lunatic fringe objectivists". It's a relentless, deliberate animosity towards JA, if not a strategic campaign, because of what he represents to you people. I can fully understand if he doesn't want to get in the crossfire, because whatever he says is always used against him by the LFO's.
As demonstrated recently by your friend Krabapple, with among other things, his stupid remark about how because JA has sometimes engaged Krueger, this somehow means he's a sick, obsessive troll like Krueger. Which is the same thing as saying the non-violent objectors in Iran who are as we speak getting their heads beaten in by the Basij, are just as guilty as these paramilitaries, for engaging in a "folie a deux". And of course, that no one in the world cares about the distinction between the two parties!
As it is with "Arny" and "Krabapple", the rest of the world disagrees.
It is bad behaviour for asking legal verification of one's identity and taking legal responsibility for one's post? Afterall, it was J_J who first requested we post our real names and insulted those who didn't.
By the way, any others stating X is YY legal name is purely heresay and not admissible in court. Not Krabapple's nor your contentions. It lowers the barriers by removing responsibility of what is posted.
Considering this thread is supposed to be about the science of audio, it seems to me rude and pointless and off-topic to spend dozens upon dozens of posts complaining about whether the site linked in someone else's profile is really his.
--Ethan
JA's referring to an extremely specific event involving a lawsuit. I guess I could try to dig it up myself but I'd imagine that somebody who actually read it as it happened would have a better idea of what to search for, as opposed to me who would in fact have to scrounge through all 10,000 posts.
It's a fascinating allegation, because I have firsthand read plenty of posts accusing Arny of pedophilia on RAO, and Arny in the past has given a pretty convincing explanation in the past for how that particular variety of smear came about. And more tenuously, the idea that nobody has sued Arny for all this time is hard to believe. So I'm interested to hear how JA doth object to Arny's engaging in the same behavior while not objecting to others.
But besides that one episode I'm not even sure I care anymore about who did what and this and that. If you really want to play that game, I've got plenty of ammunition against whatever follies Arny has engaged in. It doesn't matter. What matters the most is how one actually behaves, in the present, and there is plenty of blame to go around there - particularly at you and Steve, but also at krab and jj and arny (and probably me) too. I'm pretty sure I've schooled y'all several times over; unless I see good evidence to the contrary there's just no point in squirting any more bug juice on this fire.
As Stephen has probably already realized I'm trying harder to avoid actually flaming people out and instead just mashing the "Report" button. Dunno if it will be useful, but....
You know... you have a good point.
And ironically all of this is coming from a guy who doesn't sign his own posts with his name.
Getting back to the subject, if sasaudio can stick to it, and stop calling names, etc, I will repeat what I pointed out in the other thread:
Distinguishing two pulses separated by 5 microseconds, either in ONE or TWO channels, is not surprising, and the cues exist within the 20kHz bandwidth.
Sasaudio has, over there, refused to realize that the filters are simply part of the process, and that when I show an additional effect from putting two pulses in one channel, that that passes through the filters that remain the same filters, and are thereby filtered OUT of the experinent,.
The point?
You can extract the timing information in several ways, from either interchannel or intrachannel signals. There is simply no magic here, and the reason you can do this comes about specifically because of the AMPLITUDE resolution of the signal.
Any signal that has more bandwidth than the PCM system can handle will get spread out.
So, back to the hypothetical system that shows this, and the question:
Start with a 1 Hz sampling rate. Make the samples be taken on the seconds, i.e. 0 1 2 3 4 seconds, etc. Put the relevant .5Hz antialiasing and antiimaging filters in the system. Use a 16 bit quantizer.
Now, what is the output when you put in a rectangular pulse from .3 to .5 seconds?
What is the output when you put in a rectangular pulse from .4 to .6 seconds?
As you can determine quickly with either simulation or hardware building, the results are:
1) Non zero
2) Different
You have, in this experiment, now distinguished, inside of a PCM system, 1/10 th of a sample instant in onset time.
Now, WITHOUT dithering (yes, it's necessary!), calculate the time difference that will result in the same PCM values for the two samples.
Reduce the bit depth to 8 bits, and try it again.
Whoops, there we have it, plain, clear evidence that:
1) amplitude resolution matters
2) 1/5th of a sample in time differential is easily recognized in a PCM system, either in one or two channels.
For extra credit, what is the difference between one and tw channels in terms of minimum time differential?
N.B. If you want to try this, the easiest way is to use a Guassian pulse about DC with a sigma of 5kHz or so, because that has insignificant energy above 20khz,and thereby you won't see much filter effect at all.
In fact, Sasaudio, that's an interesting point to make to you, since you're worrying about the filters, now, what happens when you put a purely in-band signal through a constant-delay filter?
Does it change shape?
You note, now, we're doing this with purely in-band signals, with no ultrasonic energy ever having been inserted into the system. And we still distinguish enormously less than 1 sample in time resolution at 16 bits, and even at 8 bits.
For a hoot, try 1, 2, 3, 4, 5, etc, bits, and see what happens.
Now, after all that, insert dithering, Now what happens? (Hint, the results now have a random component, but even more information is present regarding attack time.)
The facts are simple.
No anti-image filter means LESS time resolution either interchannel or intrachannel, because it doesn't apply the expected "smoothing" and allows nonlinear effects to bite you.
No dithering reduces the time resolution of a system, but by a much smaller extent.
And that's the fact, jack.
You can argue all you want about Dr. K's comments, but the comment ATTRIBUTED to him about amplitude resolution having no effect on time resolution is just simply, provably, and obviously wrong. I don't know if it's the attribution, or his error, or what. I'm just telling you the facts about what has appeared here.
Bizarre, innit? Of course, it was SAS who voiced the 'complaint' in the first place and continues to do so; Frog jumped on the bandwagon as he must; and it's long past time that someone they presumably respect -- Mr. Atkinson, or Mr. Meijas the nominal moderator -- could have nipped that whole pointless, ignorant, and increasingly vicious sideshow in the bud with one sentence affirming what they know to be true.
And now Frog (and JA?) are trying to divert this discussion of Kunchur's work into another chapter of the Arny vs. Atkinson saga (/rolleyes)
Good restatement jj.
I'd also like to point out/restate that Dr. K's proposed mechanism of audibility, due to nonlinear mixing (I like to call it the ear's internal distortion but whatever), is a frequency-domain phenomenon, not time-domain. That is, it doesn't even really have anything to do with time resolution to begin with: what's necessary to reproduce his setup is a system with good ultrasonic response, and not necessarily one with good "time resolution" (although the implications of what jj is saying mean that you're going to be getting more than enough of that anyway).
His result depends on having a steady-state signal with strong ultrasonic content. It is unclear if the same 5us figure could be achieved for analog signals that have negligible ultrasonic content. Ultimately this means that even if jj was wrong and CDDA cannot encode 5us of time resolution (which it can), the paper is not a bulletproof justification for the real-world inferiority of CDDA. We'd need to test with real music, as opposed to square waves, to be sure of that.
I have signed my name twice, steve sammet. However, axon, when have you signed your name? You pulling another scam job axon?
It is amazing how you can manipulate with your claims. Slick trick J_J. Unfortunately for you the photos clearly demonstrate the inaccuracies of your claims.
So consumer, he makes slick claims EXCEPT claims of accuracy, which Dr. Kunchur and I have been harping on for some 47 pages now. The photos clearly shows and demonstrates the distortion that J_J does not want you to see.
End of story J_J. I suggest you raise the white flag.
Repeating a falsehood does not make it true.
"distortion" is nonlinearity. Which nonlinearity are you referring to?
Or are you referring to the broadening of the pulses, which, of course, is the means by which the resolution is maintained? Perhaps you've missed that basic fact?
Once again, you refuse to cope with the facts.
So, fish or cut bate, SASAUDIO, fish or cut bait. What is the answer to the questions above?
SHOW US YOU CAN FIGURE IT OUT.
Show us how YOUR graphs make MY point.
Now do it, or stop making accusations and admit you have no idea how the process actually works.
If you continue to avoid answering the basic question posed above, I will have no choice but to assume your failure to perform is intentional, and that any ignorance on your part is also intentional.
Cease misteaching, and cease it now.
Dang! Did I wander in to the "Interesting Papers" thread, or the "Michael Lavorgna" thread?
Help a brother out:
When you post in the "Interesting Papers" thread, can you just end each post with "Bitch" so we can avoid consfusion? It would also help maintain the mood.
Examples:
"Repeating a falsehood does not make it true, bitch."
"Distortion" is nonlinearity. Which nonlinearity are you referring to, bitch?"
"Or are you referring to the broadening of the pulses, which, of course, is the means by which the resolution is maintained? Perhaps you've missed that basic fact, bitch."
"Once again, you refuse to cope with the facts, bitch."
"So, fish or cut bate, SASAUDIO, fish or cut bait. What is the answer to the questions above, bitch?"
"SHOW US YOU CAN FIGURE IT OUT, BITCH."
"Show us how YOUR graphs make MY point, Bitch!."
"Now do it, or stop making accusations and admit you have no idea how the process actually works. Bitch."
"If you continue to avoid answering the basic question posed above, I will have no choice but to assume your failure to perform is intentional, and that any ignorance on your part is also intentional, bitch."
"Cease misteaching, and cease it now, BITCH!"
(Sorry, just goofing off. I plead guilty to having fallen in the same trap in the past and in the future.)
_____
One other thing: "There is no dark side of the moon?"
Are you implying that 100% of the Moon is illuminated at all times? You'd think that at given given moment, about half the Moon could be described as the dark side.
Is there a dark side of the Earth?
On the night of a full moon, what percentage of it can you see?
I think J_J's signature line is an impossibility.
Pretty simple J_J, read the article. I know J_J will provide yet another excuse. That is right, everyone is wrong except J_J of course.
Below is the link again viewers. Check under "Example" for the photos and explanation of how antialiasing distorts. Just exactly what Dr. Kunchur, his paper, and I have been stating from the beginning. But J_J keeps finding excuses to hide behind, ever since page 2.
http://en.wikipedia.org/wiki/Anti-aliasing
The photos are quite accurate and you got caught with your pants down. Shame on you trying to dupe the public.
Stay on target man. Avoid steve's cheap shots.
.... although I would like to apologize and say that I did forget that steve does sign a few of his posts.
Let me attempt to lower the thermostat a bit and posit that we and you have a different idea of the meaning of "distortion"?
jj's arguing specifically about linear vs nonlinear distortion, which is as usual a little obtuse of him. There is such a thing as linear distortion, which an antialiasing filter of course has scads of.
So what? What specifically do you see in these plots that is distortion? Why specifically is this distortion objectionable (EDIT: in the context of audibility and in the context of Kunchur's results)?
Err, linearity addresses only the following.
Does f( a + b) = f(a)+f(b)
"Yes" for all a and b and a+b in range means that f() is linear over that range. That's not obtuseness, that's the definition. (grumble)
In that sense all (edited to add "frequency shaping") filters are linear.
Putting something through a filter is not a nonlinearity.
It puts something through a transfer function, but not a nonlinear one, generally, or so one usually ensures.
And, yes, putting a pulse through a filter changes the shape of the pulse. This is due to changes (presuming a constant delay filter for simplicity) in the frequency content of the pulse.
This is not distortion, it is filtering.
And, since you can distinguish between the single pulse and the two pulses after filtering, again my point is made, and sasaudio is again shown to be unjustifiably uttering serious accusations.
As to obtuseness, I think it's entirely and deliberately obtuse of sasaudio to continue to make an appeal to his own ignorance rather than simply do the work.
I'm not making an extraordinary claim. The idea that you can resolve well below one sample time is well known and accepted, your cell phone works, your HDTV receiver works, your cable modem works, your cell phone voice coder works, AAC works (within the bounds of bit rate, of course, there is a lot of nonlinearity there), etc.
Were you unable to resolve below one sample in a PCM stream, some essential part of all of the above would fail outright.
That is pretty funny Buddha. Besides that, can you imagine all the phd contacts Dr. Kunchur made, all the university professors, physics etc, the three national organizations, more than a dozen referees, during his 5 years on the project are all wrong because they concurred/agreed with Dr. Kunchur's findings, 16/44 is not high enough.
Besides that imagine all those college students these flawed professors taught, phds etc, who were taught wrong according to J_J. And for how many years before Dr. Kunchur contacted the PHD duds.
Buddha, can you imagine all those graduates out there who know nothing and are faking it out in the world. This should cause a major collapse of science since none of them have been taught correctly by these PHDs who concurred with Dr. Kunchur's findings.
Quick, find some sparsely populated area, gather food, shelter before civilization comes to an grinding halt because of all the incompetence.
Oh, good, produce such a "flawed" professor in the realm of DSP or Fourier analysis.
You have now claimed they exist. Produce one or admit you're appealing to an authority that doesn't even exist.
Got beef, sasaudio?
By the way, what is the answer to my question regarding the simplified sampling system that I've asked you now about 4 times.
How about it?
So basically what I hear you saying is, you would have to be halfway insane to believe the meanderings of a random forumer on an internet web forum, who has admitted he has never done any of these studies, or even read them, against a respected working phd'd scientist across numerous disciplines, who is a university professor and has the backing of numerous organizations and respected working scientists and professors. I believe in science, so call me crazy, but I think I'm going to have to go with Dr. Kunchur here.
The only people I can see supporting the irrelevant strawman arguments of the random web forumer on this, are his or her friends from the HA forum. Who all have the same anti-audio agenda. Dr. Kunchur's 5-year long study is clearly a serious work, and is not intended to have any association with high end audio. But it seems equally clear that the motivation for the random web forumer and its acolytes from HA to decide to try to tear it apart (and they have certainly not succeeded), despite not having the math, the research, or the expertise to do so, is what has always motivated these people to fight against any kind of progress in audio: an irrational rejection of the aspect of sound quality in audio. Pursuing this ridiculous agenda where they try to convince intelligent people that "everything sounds the same" in audio, and therefore, "why bother to make anything sound better?". As Jan pointed out earlier, "why NOT make things sound better"? It seems they are unhappy to see others owning hifi systems that might not be as crappy as theirs. Of course, if they had any ability to engineer anything themselves in audio, they might be spending all of this energy they expend here on arguing against sound quality, and getting a life working to advance it instead. Much easier to tear things down, than to build them.
When you stand with the audio regressionists, you stand with the Luddites, the Philistines, the Flat Earthers, the Naked Emperors and the Tin-Foil Hat Crew. So be vigilante, always!
Stop it, you're killing me!
Or better put:
Stop it, you're killing me, BITCH!
Well, sorry about that, but sasaudio is misteaching on an issue that encourages a common misunderstanding. I don't find it funny.
That is interesting since I am agreeing with Dr. Kunchur, all those backing him, in which he stated, including this quote:
Dr. Kunchur:
Of course plenty of physics and math experts from different disciplines. All the PHDs back Dr. Kunchur, not you.
Now you and Ethan Winer can buddy up and play your cute misleading games, but you have yet to provide any backup/ peer review for your claims. And you have two major forms of conflict of interest, fellow of AES and employer a financial sponsor of AES. Why you have not bowed out is quite evident since you and Arny Krueger (also belongs to AES) attacked Dr. Kunchur from page 1 of this string.
J_J:
Yet Dr. Kunchur has all the backup/peer reviews and you/Ethan have none.
Audience I suggest you start at the beginning and read the string yourselves as this guy changes positions at least three times in 30 some odd pages. It is hard to know what J_J actually believes.
Honestly I'd love to hear what the peanut gallery has to say about this too, but I'd be pretty shocked if you or Dr. Kunchur actually came out ahead in that evaluation.
Produce one, please, of these "all" experts.
Sorry, no, you can't. There's a simple reason.
For once, take responsibility for your claim and provide the evidence.
It is telling that you can not tell the difference between three different versions of the same issue, and three different positions.
All of your "positions" show exactly the same elementary fallacy in sampling theory and fourier analysis.
Once again, does your cell phone work? If the answer is "yes" then you have to concede you're wrong.
As to the question of the simple, 1Hz sampling example, what's the answer, now, do tell me. You are the one claiming to show a concrete example, but you won't actually deal with one? What's the problem here?
Sneaky way of calling Dr. Kunchur a liar. Naughty naughty little boy. Read his email again that I quoted from. Please try to be honest with the public J_J.
Just keep misleading the public. The subject is whether Dr. Kunchur is correct or you. He has backup (see my previous post) and graphs/photos have been provided, you do not have anything. This has been gone over page after page. Let's see how many more times you can change positions, like a chameleon.
This string should be stopped as J_J is simply going around in circles and cannot provide any proof.
The argument stems, I believe, from the proponents looking at the situation in two different lights. Dr. Kunchur, you, and others are looking at the implications of the sampled data in isolation. J_J is looking at the sitation after the sampled two-channel data has been used to recreate a two-channel analog signal by the necessary reconstruction filter. As I understand it - and my understanding, as I said in an earlier posting, is incomplete - is that the spatial information encoded in that reconstructed two-channel signal can include timing differences between the channels that are less than the 22.676 microsecond resolution of the 44.1kHz-sampled data.
Please correct me if I am wrong either in this understanding or in my interpretation of the dispute.
John Atkinson
Editor, Stereophile
Hi JA,
You are correct, anyone can use a pulse in each of two channels, or one external reference channel and one pulsed channel and produce, reproduce 5us shifts with accuracy. Dr. Kunchur, by stating that higher sampling/bit rates are needed is clearly not using two channels or one external reference channel, but he is referring to one channel.
Yet J_J is using a different system in order to accuse Dr. Kunchur of being "fundamentally ignorant of the basics of sampling theory". Notice J_Js quote indicates Dr. Kunchur is ignorant of basic one channel sampling while he uses both two channels and one reference channel (see from page 3 to page 30 or past). This appears to be not only misleading but he makes a false accusation as well based upon a different premise.
Take care.
The claim I am addressing was a quote, attributed to Dr. K, that argued that one could not maintain this time resolution in a 16/44 system, which, of course, for either one channel OR two channels, is simply wrong, with or without any external reference OTHER THAN the existance of two pulses. One pulse, one channel, has no meaning here.
Another quote was then attributed to Dr. K, insinuating that amplitude resolution did not affect time resolution in a sampled system.
As quoted, this statement is also incorrect. Please note "as quoted here". I am well aware that people can and do misquote others, both by design and by accident.
Furthermore, as quoted here, the argument does not prove the need for ultrasonic response, as the interaural time difference, for 2 channel material, or the frequency response change (for two pulses in one channel) both provide cues, within the 20khz bandwidth, that one would expect to be distinquished by the ear, via standard understanding of the ear's function.
If you are referring to the data in the "digital" domain as "in isolation", of course the actual time delays can be calculated from that data, as well, which is obvious from the fact that one can reconstruct the signal. If the information were not encoded, it would not be available to reconstruct, of course.
The extraction of the data from the samples is not particularly difficult, but does require substantial understanding of the meaning of bandwidth limitation, and the resulting impulse response that will appear inside the sampled system. (linear phase, minimum phase, some of each, whatever, the statement still holds)
Moving the discussion in that direction is the goal of asking the question about a 1Hz sampled system with an input of duration less than 1 second, with the "on" and "off" points entirely between two sampling instants.
In order to answer the question, one must consider what the response of the antialaising filter that is presented to the sampler will be. This will show that as one moves the input signal earlier and later within the sampling instance, the data captured by the sampler will in fact change, thereby encoding the time delay via use of amplitude.
And that is a simple fact of any sampled data system.
I am not addressing the question of any need for ultrasonics in any setting beyond these. While I may or may not have some other reservations on the work, they are not for this forum.
I will say that if anything comes of such testing, air nonlinearities stand out as the potential source. Just looking at the spectrum of a rimshot as one moves away from the drum is, well, very interesting to say the least.
The question is one of level in that case. Instantaneous pressure change from mean is the key question for determining the linearity of sound transmission in air.
As has been explained to you at least twice in this thread, putting two pulses in one channel causes a clear, simply calculated, measurable, and audible FREQUENCY RESPONSE CHANGE within the 20kHz bandwidth. (for someone of excellent hearing in a very good setting, btw, not for someone in an averate situation)
This frequency response change, and the resulting impulse response changes, also, not so surprisingly, encode the presence of one or two pulses, just as expected.
What is more, great care must be taken to have the same energy in a single pulse vs. two pulses adjacent, otherwise simple, and even more obvious loudness differences will suffice for distinction.
I have said this before, several times. In the future, it would be more ethical of you to admit that this issue has been fully, completely, and incontrovertably addressed.
Your stipulation that one can resolve this in two channels, by the way, shows that you also understand that this can be done in one channel, even if you don't realize it.
You could (these are linear systems, remember?) just add the two channels to get one channel, IN WHICH BOTH PULSES CAN BE RESOLVED VIA THE APPROPRIATE MEANS.
The appropriate means are not so simple as just saying "here's a peak, there's a peak". You're dealing with bandlimited systems, which are not that simple.
As you have sidestepped for a dozen times, but not accurate reproduction. 1.5us rise time and fall times at 20khz (-.15db), total of 3us and that is not the complete rise and fall times by any means. Check the graph below for what happens with a 3us pulse.
(16/44 is miserable, notice the distortion. 5us is not much better.)
Try 4 times this time, or more, for the total/entire rise and fall times. 5us shift doesn't cut it. You can't break the law of physics J_J. The photos (See also link below) are quite accurate and you cannot wiggle out of that fact. Here is the link to antialiasing and the photos he doesn't want you to see. See under "Examples" for explanations.
http://en.wikipedia.org/wiki/Anti-aliasing
The rest of your post is an attempt to cover yourself. Don't forget your conflict of interest problems you tried to keep covered.
And, again, you fail to try to add together two of your waveforms separated by 5 microseconds.
You don't get it. It's a linear system.
Enough. Give it up, or go get the books and study.
Or just set your trace width to a reasonable size, and put 1 vs. 2 pulses through your system with the right spacing.
You will see a difference.
Then, take the one captured pulse vs. the combined captured pulses, and take a spectrum.
You will clearly see the droop at 20kHz due to the comb filter you have introduced.
Finally, learn the difference between distortion and filtering.
Distortion does not obey linearity criteria, i.e.
f(x) + f(y) != f(x+y)
Filtering is a linear operation, and
f(x) + f(y) = f(x+y)
Do you dare attempt this to sufficient accuracy? Dare you post the spectra of the two vs. 1 pulse. Dare you address the issue of total energy? Dare you address the issue of DC energy in the pulse?
Enquiring minds want to know.
The first graph clearly demonstrates the distortion produced by the 48k player with a 3us pulse fed into the machine.
Notice the original 3us pulse widens to 5us or more, but less and then none in the DSD reproduction. Anytime you change the shape from the original, distortion is created, period. Why because the original does not show any 5us information like J_J would have you believe. To simply call the change simple "filtering" and not distortion is misleading to say the least. You have introduced 5us wide information that is not originally part of the 3us width pulse. Does the original 3us pulse have any portion that is 5us wide? Of course not. You better stop refusing to provide weblinks so we can check your equations J_J.
The link to anti-aliasing:
http://en.wikipedia.org/wiki/Anti-aliasing
has photos that very clearly demonstrate the blurring/smoothing that occurs, the loss of musical inner detail. The proof is right there in front of your eyes. This is not theory but real world J_J. The graphs/photos vividly portray what is actually happening in the real world.
Give it up J_J. Your manipulation and conflict of interest clearly shows. Dr. Kunchur is correct, 16/44 is not sufficient. By the way, I would suggest you do some actual ultrasonic testing for yourself instead of sitting at a desk relying on other's reports.
Wrong. The original pulse does have a portion that is 5us wide, as I mentioned a few days ago. Counterintuitive, but physically true.
And AFAIK, you never actually responded cogently and technically to that line of argument. Only with appeals to authority.
Also, why do you refer to the width of the pulse as "information"? It's not anything the ear is sensitive to. It's not anything that is well preserved in analog systems, anyway.
That is, even if I accept the line of argument that the Pyramix plot shows distortion, you never actually proved that anybody should care about it. As it stands, there are extremely good reasons why we shouldn't.
Like from Dr. Kunchur, I'd really appreciate a response that made use of that degree you have.
The analog 3 microsecond pulse sounds better to me than the others.
Which of those rigs would you want handling your big Bach 12 inch?
Once again, you make serious accusations without any foundation.
It's clear that you are impervious to being actually educated in the technical issues.
I believe Axon has pointed it out to you already, but let's try this...
Go look up the "sinc" function (it's sin(2 pi f)/(2 pi f) ).
(Alternatively in the time domain, sin (w t) / (w t) where 'w' is omega, i.e. 2 pi f)
Your own graph shows very clearly a very nice sinc function (you appear to have good filters).
Now, what you won't (or can't) do is put in two pulses to that filter, and look at the output, and compare it to ONE pulse.
If you want them to be as similar as possible, I'd use two pulses (you can use your 3 us pulse) of half the amplitude as the original pulse, so as to have the same DC component, and thus wind up with the closest match. (otherwise, your scale will be off by a factor of 2, more or less, depending on your scaling accuracy)
You've been told the experiment, now go do it. I could run it in matlab, but you've made it clear already that you don't believe in mathematics, so go try it in the real world IN YOUR SETUP so you know nobody cheated.
JUST GO TRY THE EXPERIMENT.
As to "distortion" I will once again advise you to use the term correctly. Distortion is NON_LINEAR modification. Filtering is linear. Distortion does not obey the equation of f(x) + f(y) = f(x+y). Linear processes (like filters) do.
Go show this to yourself, since you won't take the word of a world-class expert here, TRY IT YOURSELF.
You keep appealing to authority, now it's time to appeal to my authority, son, I outrank most of the people on the planet in this subject, and now a real by damn authority is telling you what to go try.
As to your utterly INSANE paranoia about "conflict of interest", get a grip. Your tinfoil hat is showing.
Oh, and while you're at it, take the fft of both the single and double pulse, and you'll see exactly the droop on the double pulse that I've specified a few pages back. And that, my child, is a perfectly clear illustration of how a good listener in a very good setting can distinguish between the two inside of a 20kHz bandwidth)
Seen in the loo:
Please Wegel Handel, it won't Wegel Bach.
Let's see, Axon.
Let's say we allow 3us. Look at the graph again.
So now instead of the 48khz "signal" being canceled (48khz data pulse), it is present. The DSD on the far right demonstrates the 48khz "signal" being cancelled. Notice it matches the reference input 3us pulse. So by using 44khz sampling, you have introduced a signal that should have been cancelled. Pretty easy to understand.
Now the ear can detect to at least 5us right.
The fact that no ultrahigh frequency testing has actually been performed on your part (and your friends) leaves you at a distinct disadvantage, as Dr. Kunchur's paper reveals. Simply doing some math does not reveal the whole story, as we shall see next.
The RCA Radiotron Designers Handbook is always written with below 20hz and above 20khz in mind. It is a high fidelity engineering handbook.
From the RCA Radiotron Designers Handbook, written by minimum of 26 engineers, 1960, page 365.
Exact quote with bold by RCA discussing damping (above 20khz).
Three issues being discussed.
1) degree of damping of transients at ultra high frequencies.
2) Both sides discuss uptake (rise time) etc.
3) Both sides agree that damping at ultrasonic frequencies affect the uptake, which affects the sound. So they are dealing with ultrasonic frequencies (which affect below 20khz as well).
By the way, does the 48khz signal present (was not cancelled anymore) change the ultra high frequency FR, the uptake?
The total speed or sluggishness of the "uptake" is determined by the degree of damping in the ultrasonic region and also by the high frequency response as well. This has been known for at least 50 years gents. Notice this is in relation to musical reproduction. Now go argue with RCA.
I do have an advantage of working in this area for some period of time, and I wholeheartedly suggest you gentlemen do as well before attacking everyone, including Dr. Kunchur. You need to understand the whole picture.
You mean "filtering".
Nobody doubts that a narrower bandwidth pulse is narrower than it's own wideband width. This is basic Fourier analysis.
No, both sides don't agree it "affects the sound" due to the ultrasonic components. And what do you mean by "uptake"? You mean impulse response?
What 48kHz signal do you refer to? I see none in your plots, and I'm not sure which 48kHz signal you refer to.
"damping" is a misnomer here. It's all filtering, unless you're proposing nonlinear effects. And, stil, nobody that I know of is in any doubt that narrowing the bandwidth of a pulse makes it wider. Again, that's basic mathematics.
Don't need to argue with RCA, you're confusing measurements with auditory perception. So please get your terms straight. I still have no idea what you mean by "uptake". The only form I'm familiar with it being used in is in chemical uptake in biology.
I suggest that you try a simple test to determine if it is possible to determine, within the 20kHz bandwidth, if you have one 3us pulse or 2 3us pulses with the onsets differing by 5 us. Put it into your system. Look at the two outputs. They will be different. Take the spectra of the two outputs, they will be different, in particular there will be a droop in the 2 pulse case of something like .9dB at 20Hz. This is trivially calculated from the 2-tap FIR filter you are implimenting when you add the second pulse.
And, please, find out what technical term you mean by "uptake".
I have no idea why you think your appeal to RCA has any meaning, by the way, it seems very nearly incoherent.
Once again, two channels, one channel, doesn't matter either way, both ways are distinguishable inside of a 20kHz bandwidth.
I suggest you re-read my post and Axon's post again, and if necessary again. If you don't understand what uptake means, then there is no hope for you. You are willingly being ignorant.
Then Go argue with RCA.
Ok. I am out of here.
"uptake" is not a technical term in this context.
The reason that technical terms are carefully defined would be obvious by this point, I trust.
Just go ahead and try the experiment. First one, and then two, pulses. Measure the output.
Just do it.
Pages