or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Seeking education about those ultra-expensive interconnects
New Posts  All Forums:Forum Nav:

Seeking education about those ultra-expensive interconnects - Page 63

post #1861 of 2598
Quote:
Originally Posted by amirm View Post

Thank you for your answers Arny. Much appreciated.

I am a bit confused Arny. If I may replay what started all of this and now in its entirety so it doesn't look like I took anything out of context:

Amir, I have already admitted this this was an error in a reply to John Atkinson in post 1768.

That hasn't stopped you from bringing it up several times since then.

Amir, I guess your new strategy is to avoid responding to valid criticism of your new errors by replaying my old error ad nauseum.
post #1862 of 2598
Quote:
Originally Posted by stereoeditor View Post

I think it clear that your use of a Welch/Gaussian window for your FFT analyses is not optimal, Mr. Krueger, given the wide, window-related skirt either side of the central tone in your graphs.

John Atkinson
Editor, Stereophile

Perhaps you can illustrate what you mean by providing some graphs along with relevant commentary?
post #1863 of 2598
Quote:
Originally Posted by Chu Gai View Post

Perhaps you can illustrate what you mean by providing some graphs along with relevant commentary?

Here's a real world example:



Note the wide base.



Obviously nicer looking but in this case the message is the same. IOW, the height of the peaks and their location in the frequency domain is essentially the same.

Gaussian is an older form of windowing. Windowing provides a smooth shape to the block of data samples in order to avoid artifacts in the plot of the frequency analysis.
post #1864 of 2598
I'm sure Mr. Atkinson will illustrate what he means by a more optimal window and its relevance when he has the time to do so. BTW, thanks Arny.
post #1865 of 2598
Quote:
Originally Posted by stereoeditor View Post

"We"? Do you have a mouse on your pocket, Mr. Krueger?



Accuracy is such a precious thing, Mr. Krueger. It was not 5000ps but 7000ps. In any case, the sound quality of this device was disappointing in the extreme - independently reported by several listeners - which I felt was due to its sub-optimal rejection of jitter, given its generally good performance in other areas. Your suggestion that even this extreme level of jitter didn't have audible consequences is clearly wrong.

John Atkinson
Editor, Stereophile

Are you writing about the Panasonic DAT?
post #1866 of 2598
Quote:
Originally Posted by Chu Gai View Post

Are you writing about the Panasonic DAT?

I'm not sure, but he appears to have forgotten about the McIntosh MS750 media server. This from Stereophile:

"Whether the MS750 was playing back a FLAC-encoded, "CD-quality" file or a CD, the player offered a very high measured level of jitter: nearly 14ns, or two orders of magnitude higher than the best players I have measured. Significant spectral spreading can be seen in the spectrum of the McIntosh's output (fig.4), which suggests the presence of high levels of random low-frequency noise affecting the clock-generator circuit, as well as pairs of sidebands at high levels and low frequencies, and at higher frequencies but lower levels. The averaged level of the noise in this graph is around 12dB higher than the 16-bit noise floor, suggesting closer to 14-bit performance than the necessary 16-bit. As suggested in the review, the McIntosh needs to be used with a standalone D/A processor to give the best sound."
post #1867 of 2598
Quote:
Originally Posted by Chu Gai View Post

Quote:
Originally Posted by stereoeditor View Post

I think it clear that your use of a Welch/Gaussian window for your FFT analyses is not optimal, Mr. Krueger, given the wide, window-related skirt either side of the central tone in your graphs.

Perhaps you can illustrate what you mean by providing some graphs along with relevant commentary?

In FFT analysis, you are performing a mathematical manipulation of finite series of time-sampled data points. The assumption behind the discrete Fourier Transform is that if you wrap the final data point to the first and there is no discontinuity, you get a smoothly continuous waveform with a frequency equivalent to the reciprocal of the length of the series of time data. FFT analysis transforms this time-domain wave to a line spectrum in the frequency domain, with the lines (colloquially called "bins") equivalent to the harmonics of the assumed waveform and spaced at that frequency. A 1-second data sample will thus give you a spectrum with bins spaced at 1Hz all the way up to the half the sampling frequency of the time-domain data.

However, if there is a mismatch in an amplitude between the final and first data points in your series of time data, the resultant discontinuity in the assumed wave results in an inaccurate spectrum. So you need to "window" the time data with some kind of mathematical function so as to produce a continuous waveform. If, for example, before performing the FFT analysis, you multiplied the amplitude of each sample by the reciprocal of the number of datapoints between it and the centerpoint of the time series, you would have a waveform that smoothly dropped to zero at its beginning and end.

However, while it does the job of allowing a smooth wrap, such a "triangular" window is almost never used. Books have been written about FFT windowing functions - I recommend Bracewell's "The Fourier Transform" (1986) - but basically, different windows, to differing degrees all trade off accuracy in the spectral amplitudes against leakage between bins. A rectangular or "top hat" window has no leakage but cannot be used with time data that would produce a discontinuous wave.

So before performing FFT analysis, you need to choose an appropriate window for your data, which is what I do with my published jitter graphs. See, for example, fig.1 at http://www.stereophile.com/content/case-jitters , which has no "skirts" around the central spike.

The window used by Arny Krueger in the graphs he posted to this forum suffers from a large degree of leakage between adjacent bins, and is thus not optimal for displaying low-level sidebands close to a high-level signal.

Arny dismisses errors such as this as trivial, but they are actually very revealing. That Arny Krueger didn't use an appropriate window for his FFT graphs is, like his choral recording, evidence that he doesn't understand the theory behind the tools he uses. Just as he _still_ doesn't understand that Amir offered a complete answer to his question about the audibility of low-frequency wow.

John Atkinson
Editor, Stereophile
post #1868 of 2598
Quote:
Originally Posted by audiophilesavant View Post

I'm not sure, but he appears to have forgotten about the McIntosh MS750 media server. This from Stereophile:

"Whether the MS750 was playing back a FLAC-encoded, "CD-quality" file or a CD, the player offered a very high measured level of jitter: nearly 14ns, or two orders of magnitude higher than the best players I have measured. Significant spectral spreading can be seen in the spectrum of the McIntosh's output (fig.4), which suggests the presence of high levels of random low-frequency noise affecting the clock-generator circuit, as well as pairs of sidebands at high levels and low frequencies, and at higher frequencies but lower levels. The averaged level of the noise in this graph is around 12dB higher than the 16-bit noise floor, suggesting closer to 14-bit performance than the necessary 16-bit. As suggested in the review, the McIntosh needs to be used with a standalone D/A processor to give the best sound."

Indeed I did. Dreadful-measuring and dreadful-sounding.

John Atkinson
Editor, Stereophile
post #1869 of 2598
Quote:
Originally Posted by stereoeditor View Post

Indeed I did. Dreadful-measuring and dreadful-sounding.

And it didn't even have a big blue meter to offset its performance deficiencies.
post #1870 of 2598
Quote:
Originally Posted by stereoeditor View Post


Arny dismisses errors such as this as trivial, but they are actually very revealing. That Arny Krueger didn't use an appropriate window for his FFT graphs is, like his choral recording, evidence that he doesn't understand the theory behind the tools he uses.

Count on Gentleman John to write an extended personal attack like this when I post a FFT plot that accidentally was taken with a suboptimal windowing.

Perchance one fire up the wayback machine and sort through the nearly one thousand FFTs that I posted as part of my www.pcavtech.com web site, one finds essentially nothing but beautifully windowed plots.

Quote:


Just as he _still_ doesn't understand that Amir offered a complete answer to his question about the audibility of low-frequency wow.

Really? You mean all the time and bandwidth that Amir wasted pretending to laboriously correct a post that I made in error after I admitted that it was wrong and had posted a correction of my own?

There's at least one kind of double blind that John seems to like, and that's being double blind to the errors of people who argue against me. ;-)
post #1871 of 2598
Quote:
Originally Posted by stereoeditor View Post

Indeed I did. Dreadful-measuring and dreadful-sounding.

And of course John has all sorts of bias-controlled listening tests to back that up...

;-)


Just speculating, but I suspect this is more like it: John runs through the staff cubicles at Stereophile shouting: "Guys you have to hear this - its got over a zillion picoseconds of jitter!"
post #1872 of 2598
Quote:
Originally Posted by arnyk View Post

Quote:
Originally Posted by stereoeditor View Post

Arny dismisses errors such as this as trivial, but they are actually very revealing. That Arny Krueger didn't use an appropriate window for his FFT graphs is, like his choral recording, evidence that he doesn't understand the theory behind the tools he uses.

Count on Gentleman John to write an extended personal attack like this when I post a FFT plot that accidentally was taken with a suboptimal windowing.

It wasn't _one_ graph, Mr. Krueger; it was _all_ of them. And yes, you did subsequently post a graph with a more suitable window but only after I had pointed out your error. "Accidentally," Mr. Krueger? Just like you accidentally said that your mono recording had excessive stereo separation? :-)

Quote:


Perchance one fire up the wayback machine and sort through the nearly one thousand FFTs that I posted as part of my www.pcavtech.com web site, one finds essentially nothing but beautifully windowed plots.

That link appears to be broken. :-(

John Atkinson
Editor, Stereophile
post #1873 of 2598
Quote:
Originally Posted by audiophilesavant View Post

I'm not sure, but he appears to have forgotten about the McIntosh MS750 media server. This from Stereophile:

"Whether the MS750 was playing back a FLAC-encoded, "CD-quality" file or a CD, the player offered a very high measured level of jitter: nearly 14ns, or two orders of magnitude higher than the best players I have measured. Significant spectral spreading can be seen in the spectrum of the McIntosh's output (fig.4), which suggests the presence of high levels of random low-frequency noise affecting the clock-generator circuit, as well as pairs of sidebands at high levels and low frequencies, and at higher frequencies but lower levels. The averaged level of the noise in this graph is around 12dB higher than the 16-bit noise floor, suggesting closer to 14-bit performance than the necessary 16-bit. As suggested in the review, the McIntosh needs to be used with a standalone D/A processor to give the best sound."

This is also from Stereophile:

"The real question is How good did the MS750 sound? As I said, I ripped all the music I stored on it as full-resolution FLAC files—results with lossy formats will necessarily suffer in comparison. In FLAC, I found it virtually impossible to detect any difference between the original CDs played on the MS750 and their stored FLAC versions. Well, I say virtually to salve my wounded pride. I wouldn't wager large amounts of cash on my consistent ability to discern the originals from the ripped copies."

I guess the question is which Stereophile does one believe? ;-)

And that my friend is why one does DBTs - you get more reliable, consistent results.
post #1874 of 2598
Quote:
Originally Posted by audiophilesavant View Post

And it didn't even have a big blue meter to offset its performance deficiencies.

Another interesting quote from John's "A case of the jitters"

http://www.stereophile.com/content/c...ess-cd-quality

"As suggested in the review, the McIntosh needs to be used with a standalone D/A processor to give the best sound."

The implication being that the standalone D/A processor will alleviate the Mac music server's jitter. This in spite of what John said earlier in this thread about once jitter gets into a signal, nothing can take it back out.
post #1875 of 2598
So, Mr. Atkinson, in the example that Arny posted where he used a different window, how does this affect the interpretation of the results? But let's not be too harsh for there must have been instances when FFT has confounded you too, no?
Further, finding PCAvtech using the wayback machine works. Perhaps you weren't aware of www.waybackmachine.org. It's a repository of websites long since gone as well as other things. An amazing amount of material is archived.
post #1876 of 2598
That McIntosh was pretty shaky. Whose product did they rebadge?
post #1877 of 2598
Quote:
I take the easy way out from time to time and consult Wikipedia to refresh my recollection.
So, wiki tells you that black and white are colours...even though the same page says they're not....but part of it supports your 'theory' so you run with it.
LOL.
post #1878 of 2598
Quote:
Originally Posted by arnyk View Post
Another interesting quote from John's "A case of the jitters"

http://www.stereophile.com/content/c...ess-cd-quality

"As suggested in the review, the McIntosh needs to be used with a standalone D/A processor to give the best sound."

The implication being that the standalone D/A processor will alleviate the Mac music server's jitter. This in spite of what John said earlier in this thread about once jitter gets into a signal, nothing can take it back out.
There is no conflict. There is no causal connection between the jitter in the McIntosh' analog output and that from its digital port.

John Atkinson
Editor, Stereophile
post #1879 of 2598
Quote:
Originally Posted by stereoeditor View Post
There is no conflict. There is no causal connection between the jitter in the McIntosh' analog output and that from its digital port.
Interesting. So you are telling me that the analog and digital outputs have nothing in common?

Two of everything from input to output?

Well, it is an expensive piece! ;-)

Back in the real world, the Mac's digital output is probably closely associated with the input to its internal DAC.

The timing of the digital output and therefore the input to the DAC were jittery, perhaps due to the fact that the master clock for the whole box was jittery.

I know of no published measurements that disprove this conjecture.
post #1880 of 2598
Quote:
Originally Posted by rock_bottom View Post
I'm rather partial to phony impulse responses myself.
Quote:
Originally Posted by SAM64 View Post
So, wiki tells you that black and white are colours...even though the same page says they're not....but part of it supports your 'theory' so you run with it.
LOL.
C'mon Sam! Get on the jitter and ultra expensive interconnects chat. You know WTF AS was talking about!
post #1881 of 2598
Quote:
Originally Posted by Chu Gai View Post
So, Mr. Atkinson, in the example that Arny posted where he used a different window, how does this affect the interpretation of the results?
In the example he reworked with a more appropriate window, there were no other sidebands revealed by the elimination of the windowing bin leakage. The interpretation doesn't change, therefore. But the point is that Mr. Krueger didn't realize the error until it was pointed out to him. There seems to be a double standard operating on this forum, where things that Amir and I are examined with a microscope yet errors and misstatements by, for example, Arny Krueger, go without question.

Quote:
But let's not be too harsh for there must have been instances when FFT has confounded you too, no?
Really? I hope not.

Quote:
Further, finding PCAvtech using the wayback machine works. Perhaps you weren't aware of www.waybackmachine.org. It's a repository of websites long since gone as well as other things. An amazing amount of material is archived.
Thank you.

Quote:
Originally Posted by Chu Gai
That McIntosh was pretty shaky. Whose product did they rebadge?
It was an Escient.

John Atkinson
Editor, Stereophile
post #1882 of 2598
Quote:
Originally Posted by arnyk View Post
Amir, I have already admitted this this was an error in a reply to John Atkinson in post 1768.
What??? You mean I could have saved all the effort that went into my last explanation? Darn it Arny .

Let's review post 1768:

Quote:
Originally Posted by arnyk View Post
I double checked the reference, and the quote from Villchur and Allison was correct and in context. It is also incorrect in the light of a number of later references including Zwicker and Fastel etc.

So that makes it two for you in one day, John. I was mislead by a seemingly authoritative reference, and came up wrong.
You say you double check the reference and that it was incorrect. It actually is not that at all. The table from Villchur and Allison starts 4 Hz peak and goes down as it should relative to Zwicker and Fastel's work. Here is your quote and my chart again:

Quote:
"On the Magnitude and Audibility of FM Distortion (JAES, 1982) by the well known audio pioneers Allison and Villchur is an old paper but one that is still referenced by modern writers. In that paper we find the following table of the audibility of FM distortion versus frequency:

Flutter Frequency [Hz] Weighting Factor
4 Hz 1.0
10 0.79
30 0.37
60 0.21
100 0.14
200 0.07
Now the Zwiker and Fastl reference again:



The declining numbers from table above agree with declining right side of the graph. You referenced the book where my graphics came from.

You do own the book and have read it. Right? If so, why did you not notice the graph I posted above? It directly contradicts your original statement.

Please note that you provided no public reference for us to check for ourselves. We had to take your word for it. I hope you appreciate that it is hard to do that in the future if you can be misled so easily and drastically.

Quote:
That hasn't stopped you from bringing it up several times since then.
I haven't brought it up several times Arny. Until this post, it still looked very much like you were disagreeing with me. I mean this last argument has lasted 12+ pages and you still going at it. Why didn't you stop folks beating me up to answer your flutter challenge and say to them you were mistaken? That would have stopped the discussion then and not put us through all this work.

I guess you are saying I should have noticed your shifting position:

Quote:
Originally Posted by arnyk View Post
Finally, here's what happens when I try to draw a maksing curve for the carrier over the worst case scenario:

....

Here we go again, folks. By leaving out critical information including the well-known Carson's rule, a false impression was created.
The entire argument started with me saying masking makes some distortion LESS audible. Let me repeat, here is Amir trying to say what is *less* audible. I got a ton of arguments, all of which centered on masking not be appropriate psychoacoustics method to predict audibility of jitter. And here you are, using masking to argue back with me? And in light of making a gross error say I am creating a "false impression?"

Quote:
Amir, I guess your new strategy is to avoid responding to valid criticism of your new errors by replaying my old error ad nauseum.
I don't know what "valid criticism" is left Arny. You said flutter audibility was one thing, I showed it was another. I said Dunn supported the notion of masking reducing audibility of jitter. You complained about his work yet adopted that very reference to make your point on Miller charts. I said masking is useful to figure out what is audible wrt to jitter. You said it was not yet you are using it now.

In other words, you have thrown our your position and now use everything I said as your arguments moving forward.

So please, after 12 pages of this don't say we are the guilty party. We don't owe you an apology at the end of this thing. We really don't even if we mentioned this a few more times to make up for all the typing we had to do to address the original, incorrect assertion.
post #1883 of 2598
Quote:
Originally Posted by arnyk View Post
Count on Gentleman John to write an extended personal attack like this when I post a FFT plot that accidentally was taken with a suboptimal windowing.
It is not accidental Arny. I have never seen you use the correct one. I always cringe looking at your FFTs. I said this earlier in this very thread (about a month ago!) over my frustration of you not following the ABC of FFT analysis. Please pardon old Amir's French :
Quote:
Originally Posted by amirm View Post
FFT? What the hell does FFT have to do with. Do you understand windowing issues with FFTs? I assume not.
Quote:
Perchance one fire up the wayback machine and sort through the nearly one thousand FFTs that I posted as part of my www.pcavtech.com web site, one finds essentially nothing but beautifully windowed plots.
It has nothing to do with beauty. I had to do with knowing the trade offs the each windowing function represents and using the right one for the application at hand.

Yes, the larger message may remain even with a different window. But one loses confidence when incorrect measurements are performed to make a point.
post #1884 of 2598
Quote:
Originally Posted by amirm View Post
It is not accidental Arny. I have never seen you use the correct one. I always cringe looking at your FFTs.
Amir, my records show that up until 7/15/2011, every FFT I posted was made by someone else - taken from a book, a paper, a web site, some taken from your posts.

So Amir, knock yourself out criticizing all those other people's work!
post #1885 of 2598
Quote:
Originally Posted by stereoeditor View Post
In the example he reworked with a more appropriate window, there were no other sidebands revealed by the elimination of the windowing bin leakage. The interpretation doesn't change, therefore. But the point is that Mr. Krueger didn't realize the error until it was pointed out to him. There seems to be a double standard operating on this forum, where things that Amir and I are examined with a microscope yet errors and misstatements by, for example, Arny Krueger, go without question.
You're right. I'm not savvy enough to have noticed it. I looked at a few graphs from his defunct website using Internet Archives and found this error to be absent. To me, Arny's error here was likely due to posting too quickly and not double checking first. IOW, his error was not a result of general ignorance or incompetence.

Quote:
Really? I hope not.
I know of no person who hasn't made errors except for my youngest. A quick Google Group search turned up the following post from Richard Pierce written in 1995.

Quote:

Several years ago, Stereophile presented a conclusive set of measurements
demonstrating the measurable difference accounting for the audible
differences between two versions of a power supply for a high end
turntable. As "proof" of the differences, spectra were presented that
showed significant sidebands in single-tone sources from LPs played on
that truntable. In a discussion that I had directly with John Atkinson,
it became QUITE clear that ALL of the results he obtained were due to
incorrect measurement procedure. Here's why:
His procedure consisted of putting on an LP with a continuous, low
distortion 1 kHz sine wave, sampling the data with a MLSSA system, then
performing an FFT on the data to obtain the spectrum. All such
measurement system depend upon the correct windowing of the data. In one
case, because of the speed of the turntable, he was able to window
exactly an integral number of cycles of the signal in the measurement
window. In the other case, because of NOTHING MORE THAN A VERY SLIGHT
DIFFERENCE IN SPEED, he was not.
The result was a slight phase discontinuity at the end of the window. The
FFT methodology presumes the signal exists for infinite time, and views
the window your looking at as a slice of that time period. Because, now,
there was a slight discontinuity at the ends of the window, the FFT
algorithm treated this as real data. The result was that instead of
narrow line spectra at the fundamental and the harmonics, we saw
significant spreading. It was this sigificant spreading that Stereophile
said was "proof" of the differences between the two units. They pouinted
to the resulting spectrum and claimed that it was "mechanical jitter" or
some such, rather than what it was: bad measurement procedure.
I was able to PRECISELY duplicate their measurement results, simply by
improper use of the equipment. Had they either adjusted the speed of one
OR the other so they were the same, OR had they used a chirp-z transform
instead (which allows them to window and analyze data whose length is
something other than an integral power of 2 samples long) and adjusted
the window length properly, they would have gotten markedly different
results.
Quote:
Thank you.
Might be some old recordings in the public domain of interest.


Quote:
It was an Escient.
I guess even flagships can be sunk.j
post #1886 of 2598
Quote:
Originally Posted by stereoeditor View Post
There seems to be a double standard operating on this forum, where things that Amir and I are examined with a microscope yet errors and misstatements by, for example, Arny Krueger, go without question.
If I were you John, I wouldnt put myself in the same category as Amir.

Virtally every technical post of substance he makes has at least one very serious flaw. I'm not talking about suboptimal choice of windowing, but major goof-ups such as this morning's statement by Amir that the effect of changes in modulation index over the range involved in our discussions was to significantly increase the bandwidth of the sidebands. Along the way he completely blew off the effects of modulating frequency.

In your case you've written and published so much that as targets go, you're about the size of the side of a barn. ;-) But as a rule your technical posts withstand a lot more scrutiny.
post #1887 of 2598
Quote:
Originally Posted by Chu Gai View Post
You're right. I'm not savvy enough to have noticed it. I looked at a few graphs from his defunct website using Internet Archives and found this error to be absent. To me, Arny's error here was likely due to posting too quickly and not double checking first. IOW, his error was not a result of general ignorance or incompetence.
Thank you.

John knows from practical experience on Usenet that my posts are intended by me to be casual conversation. Every post is not a filing for a PhD thesis!

I've told John this many times, many different ways, and when he gets under a lot of pressure, he always seems to forget.
post #1888 of 2598
Quote:
Originally Posted by arnyk View Post
Amir, my records show that up until 7/15/2011, every FFT I posted was made by someone else - taken from a book, a paper, a web site, some taken from your posts.

So Amir, knock yourself out criticizing all those other people's work!
How sure are you of that Arny?

Here is a set of graphs you and I have debated to death with you: http://home.comcast.net/~arnyk/pcavtech/lynxtwo/

No date on them but the linking page has this on it:

"Please let The Webmaster know of any problems you encounter on this website.
This Page Last Revised 11/06/2005 (abk)"

Pretty sure those Lynx charts were not made on 7/15/2011.

I am happy to dig up your statements on when you created those spectrum charts. Would you like me to or shall we add this to the list of incorrect things said and move on?
post #1889 of 2598
Quote:
Originally Posted by Chu Gai View Post

Quote:
Originally Posted by stereoeditor View Post

In the example he reworked with a more appropriate window, there were no other sidebands revealed by the elimination of the windowing bin leakage. The interpretation doesn't change, therefore. But the point is that Mr. Krueger didn't realize the error until it was pointed out to him. There seems to be a double standard operating on this forum, where things that Amir and I are examined with a microscope yet errors and misstatements by, for example, Arny Krueger, go without question.

You're right. I'm not savvy enough to have noticed it. I looked at a few graphs from his defunct website using Internet Archives and found this error to be absent. To me, Arny's error here was likely due to posting too quickly and not double checking first. IOW, his error was not a result of general ignorance or incompetence.

With respect, you are bending over backwards to give Arny Krueger the benefit of the doubt.

Quote:
Originally Posted by Chu Gai View Post

Quote:
Originally Posted by stereoeditor View Post

Quote:
Originally Posted by Chu Gai View Post

But let's not be too harsh for there must have been instances when FFT has confounded you too, no?

Really? I hope not.

I know of no person who hasn't made errors except for my youngest.

I wasn't claiming _not_ to have made any errors, only that I hoped I hadn't. Why don't I get the same benefit of the doubt you give Arny Krueger?

Quote:


A quick Google Group search turned up the following post from Richard Pierce written in 1995.

To save bandwidth, I have not included Dick's Usenet posting in this response, but it was posted on rec.audio.tech under the subject: "Re: Directionality of cales" on 1995/08/27. The Message-ID is .

I am extremely skeptical of Dick's claim that the FFT was synchronous in one case but not in the other. I still have the MLSSA capture files and it appears that the use of a rectangular window was sub-optimal in all cases. Mea culpa.

BTW, Dick was referring to measurements I had published in January 1991. You can find these at
http://www.stereophile.com/content/l...y-measurements . Yes, I incorrectly used a rectangular window. But contrary to what Dick wrote, I don't think I claimed anything that wasn't supported by my use of the suboptimal window in the measurement: "What can I draw from these measurements? Certainly nothing that would indicate why the Linn/Lingo appears to dig down an extra octave in the bass. But two things do strike me. One, the tonearm low-frequency resonance appears to dominate measurements of a turntable's speed stability, though the Lingo-powered LP12 is excellent, nevertheless, in that it produces a clean fundamental peak in the spectrum."

Nothing very controversial or misleading there. I repeated the work in 2003, this time using Adobe Audition's FFT tool (which is what Mr. Krueger uses) and a more appropriate Blackman-Harris window. You can compare the graphs at http://www.stereophile.com/content/l...m-measurements with those from 1991.

So yes, I did make an error in choice of FFT window more than 20 years ago. I don't think it significant as I appeared both to have not claimed anything that wasn't supported by the data and to have learned from my error.

John Atkinson
Editor, Stereophile
post #1890 of 2598
Quote:
Originally Posted by arnyk View Post

That would be an oops, and in this case an oops that is irrelevant to the point being made.

WOOHOO! You can do it arny, you can DO it arny, you CAN do it arny!!

And and and and??? waiting for the very brave very tiny very little next step!!




What's this arny? An inkling into what ails your inner soul, the source of the personal turmoil?

Quote:


There's at least one kind of double blind that John seems to like, and that's being double blind to the errors of people who argue against me. ;-)

Where's AJ when you (finally and unexpectedly) need him? He's not here so I will have to paraphrase myself..."oh, playing the victim game again are we?"




I asked you just before what was this sh1t?

Quote:
Originally Posted by arnyk View Post

Here we go again, folks. By leaving out critical information including the well-known Carson's rule, a false impression was created.

Questions? Comments?

So, according to YOU, when others make what YOU perceive to be an error, it is deliberate, cunning, calculated and fully known, that it was done due to an attempt to deceive.

Well then, what is THIS sh1t??

Quote:
Originally Posted by arnyk View Post

Count on Gentleman John to write an extended personal attack like this when I post a FFT plot that accidentally was taken with a suboptimal windowing.

Quote:
Originally Posted by arnyk View Post

Thank you.

John knows from practical experience on Usenet that my posts are intended by me to be casual conversation. Every post is not a filing for a PhD thesis!

Oh man, what a frickin hypocrite! When YOU make an error, there are all these extenuating circumstances, when someone else does it it is duplicitous conduct.

HECK! WE are all aware that as it is NOT a PHD thesis sometimes things get said that are incomplete, poorly worded or whatever. We don't expect any sort of self flagellation as a result, we know that it can happen easily. Those of us that do it readily and promptly go 'oh yeah, how silly, what I meant was..', or maybe even all we do is say 'oops'.

why? because it happens.

You poor old bugger, you have lost the basic human decency along the way.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Seeking education about those ultra-expensive interconnects