The first mile: What power stations are best for audio?

[jabbr]

Have faith, close your eyes and walk...

 

 

Proving evolution one step at a time:)

[Jud]

I was unaware of his work. I just used the Pen as an example, hoping to avoid some of the trigger subjects that seem to raise so much ire here. Even so, I never heard any difference when I used the Pen and treated a number of my discs and when I and a group of my audio buddies compared treated and untreated copies of the same CDs in a semi-DBT, nobody could tell any difference, ever. At the time my CD player was the highly regarded Sony XA777ES, an Audio Research SP11 Preamp and a pair of VTL 140 mono-blocks in triode mode feeding a pair of Martin Logan Aeon-i speakers, so the resolving power of my system was not in question, especially for the time.

 

But that really doesn't matter. My question stands: Whatever the subject, when one staunchly maintains that a personally observed hypothesis is true when all empirical and scientific evidence says that the observed result cannot be, is the advocate remaining objective about his conclusions, or has he crossed the line into religious belief? Whether the subject is the existence of ghosts, little green men from Mars, or belief in a phenomenon that physics can't explain, it's all one with regard to that larger question in my opinion.

 

Sure. But two things I very much appreciate, since I particularly enjoy irony, are:

 

- Many folks, not just you, use the example of green pen on CD as the ultimate subjectivist religious belief, only to find it's a lovely example of a situation where solid engineering eventually bore out what the subjectivists claimed to hear.

 

- Many folks of an objective bent, not just you, then say either that they can't hear a difference, or simply maintain their skepticism after the engineering explanation is given to them. In other words, clinging to a favored belief system in the face of evidence to the contrary is by no means the exclusive province of so-called subjectivists.

[Jud]

One more thing about green pen and CDs:

 

I like to use that example because the thought of green pen making an audible difference was so obviously ludicrous to *me*. After being shown that even something I thought so ridiculous had a prosaic engineering explanation that requires no magic, just an open and curious engineering mind capable of exploring the possible causes of a reported phenomenon rather than dismissing it out of hand - well, I try to exercise a little more humility now. I think that's a good thing.

[wgscott]

What is the compelling engineering explanation?

 

The only thing I found when googling was this:

 

The most commonly offered explanation for the allegedly improved results produced by a "greened" CD is that the light from a CD player's laser reflects off the shiny inner rim and outer edge of the CD and enters the "eye" of the player, thereby altering the digital bit count and distorting the sound. Coating the edges of the CD with a colored marker supposedly reduces or eliminates the amount of stray light reflecting of the disc's edges, producing "better" or "cleaner" sound. (Green markers are used because the faithful believe that color most effectively "absorbs" the light from the laser's infrared beam.) As former Stereo Review and High Fidelity editor David Ranada pointed out, however, light travels so quickly that it would be reflected back to the laser from the edge of the disc while the laser was still reading the same digital bit and therefore could not produce a distorted reading. Ranada confirmed his assertion by connecting a digital error counter to a CD player to compare data errors produced during playback of both colored and uncolored discs. He found no difference between the two types of discs at any portion of their surfaces -- inner rim, outer rim, or middle. He also tried coloring only half the circumference of a disc and using an oscilloscope to analyze the signal picked up by the laser. The scope showed no difference between the patterns produced by the colored and uncolored halves of the disc.

 

Urban Legends Reference Pages: Music (Bewaring of the Green)

[sandyk]

In other words, clinging to a favored belief system in the face of evidence to the contrary is by no means the exclusive province of so-called subjectivists.

 

+1

Not that George or The Ill Tempered Audiophool would ever admit to it !

[wgscott]

+1

Not that George or The Ill Tempered Audiophool would ever admit to it !

 

It happens all the time to everyone.

 

The good thing about science is that it is self-correcting.

 

One also has to evaluate the so-called evidence and decide whether it is compelling, fraudulent, irrelevant, etc.

 

I am eager to hear what the evidence for the green pen CD thing is. I admit to clinging to the belief that this is BS, but would like to see the evidence in its favor.

[sandyk]

I am eager to hear what the evidence for the green pen CD thing is. I admit to clinging to the belief that this is BS, but would like to see the evidence in its favor.

 

My recollection is similar to that of Jud, in that it was found to increase "Jitter", which is perhaps why some thought it sounded a little better.

[firedog]

I don't have the link for reference, but yes, the Green pen CD was found to increase jitter, apparently in a way many listeners found euphonic.

 

Also not an unknown phenomenon in other aspects of digital audio - sometimes increased jitter of certain types sounds subjectively good. Not unlike the reaction to some of the distortions produced by some tubes. (Don't flame me, I'm not puttin down tubes in general. )

[Jud]

What is the compelling engineering explanation?

 

The only thing I found when googling was this:

 

Urban Legends Reference Pages: Music (Bewaring of the Green)

 

This post from our host has an embedded video where (at time 31:30) Keith Johnson gives the explanation and shows his scope measurements: http://www.computeraudiophile.com/f7-disk-storage-music-library-storage/do-apple-lossless-files-really-sound-same-aiff-15557/index6.html#post220485

 

It has to do with the disc player read head servos kicking electrical noise into the system and thereby creating jitter.

 

As the Snopes article you cite above and Keith Johnson's presentation show, while subjectivists were correct about a change in sound, their attempted explanation was bogus, and the effect was actually an increase in distortion rather than the subjectively perceived improvement. So both "objectivists" and "subjectivists" were wrong, and much vituperation and insult-throwing was wasted on both sides. (I hate to see a good insult wasted.)

 

As you've said, science (when it *is* science) is self-correcting. The reason it's self-correcting is that it's humble. It's not confident in an answer unless it's been demonstrated, and demonstrated again, and even then it keeps on asking questions to try to find holes in that answer. There's that lovely Isaac Asimov quote about the best words in science not being "Eureka!", but "Gee, that's odd...." It's hard to do real science if you already know everything. So we have to struggle against our natures that make us want to be first with the answer (think of grade schoolers frantically waving their hands wanting the teacher to call on them), and not think we have the answer until there's damn good data.

[mansr]

- Many folks, not just you, use the example of green pen on CD as the ultimate subjectivist religious belief, only to find it's a lovely example of a situation where solid engineering eventually bore out what the subjectivists claimed to hear.

 

Except it didn't. There has been one experiment showing the opposite effect. As far as I can tell, this has never been replicated. To believe it, I'll need to see at minimum an independent test coming to the same conclusions.

[Jud]

Except it didn't. There has been one experiment showing the opposite effect. As far as I can tell, this has never been replicated. To believe it, I'll need to see at minimum an independent test coming to the same conclusions.

 

Opposite to what? The objectivist claim was that no difference was physically possible; subjectivists claimed to hear a difference. They were correct. As I also mentioned in this thread, the subjectivists happened to be 180 degrees wrong about claiming the difference was an improvement. And thus (as I noted in my most recent previous comment), both objectivists and subjectivists had it wrong, and all the arguing was useless. This is what you get when people argue doctrine rather than data.

 

Regarding "one" experiment: Keith Johnson has a reputation of being an extremely careful engineer. I don't know whether he did just one single experiment on this particular issue, though that would be out of character. Are you a good friend of his, that you know he did only a single experimental run and was not careful to confirm his results?

 

When I previously gave you this link, you questioned whether any actual data at all was collected. Since that particular hurdle was crossed, you're now demanding an "independent" confirming experiment "at minimum" before you "believe it." I seriously doubt that'll happen. Barring that, you're certainly free to hang on to your faith that it's all BS.

[mansr]

Regarding "one" experiment: Keith Johnson has a reputation of being an extremely careful engineer. I don't know whether he did just one single experiment on this particular issue, though that would be out of character. Are you a good friend of his, that you know he did only a single experimental run and was not careful to confirm his results?

 

Enough with the appeal to authority. Also, if KJ did one or one hundred experiments doesn't matter, he's presenting a single result.

[Jud]

Enough with the appeal to authority. Also, if KJ did one or one hundred experiments doesn't matter, he's presenting a single result.

 

This reminds me of non-expert laypeople commenting in evolutionary biology blogs who believe in "Intelligent Design" and are thus skeptical of evolution, who start out saying there's never been an experimental demonstration of evolution; are provided information on Richard Lenski's long term evolution experiment (which I think might literally be worthy of a Nobel prize some day); and then say "That's just one experiment, and you're appealing to authority."

 

Yeah, I'm appealing to authority, if "authority" is translated as "someone who knows a hell of a lot about the subject, and is famous among peers for doing good careful work." But hey, as I said in my previous post, you can certainly choose not to believe the data due to a predisposition for a conclusion other than the one the data indicates, i.e., faith.

[wgscott]

I have to agree with mansr on this one. The single report of a difference in phase error, along with the difference being opposite in magnitude to what was claimed, is not compelling. What does it even mean to have the jitter increased? Is the difference in jitter audible? Does the jitter oscillate when only one half of the CD is marked? Is the result repeatable on a variety of CDs, in a variety of systems? Do different colors effect the jitter values differently? Does ripping a CD marked with a green pen produce a different file (why not?). What experimental controls were used? (Can one have differences in jitter of the magnitude measured using different unmarked CDs for example?)

 

To paraphrase Carl Sagan, extraordinary claims require extraordinary evidence, not just a single report of "there is a slight difference, but opposite to that predicted." What mansr or I believe or don't belive is irrelevant. What is hoped for is a robust, unambiguous and repeatable result. The experiment Snopes reports comes much more close to satisfying those criteria.

[wgscott]

I don't have the link for reference, but yes, the Green pen CD was found to increase jitter, apparently in a way many listeners found euphonic.

 

The link Jud posted said it was 10 times below the threshold of hearing.

 

Claim: Coating the edges of a compact disc with a green marking pen will noticeably improve its sound quality.

 

Keith Johnson's measurements very unambiguously refute this claim. He measured differences that are 10 times below the threshold of hearing.

[firedog]

The link Jud posted said it was 10 times below the threshold of hearing.

 

 

 

Keith Johnson's measurements very unambiguously refute this claim. He measured differences that are 10 times below the threshold of hearing.

 

No, unless you are using old experimental results that have been shown to be wrong. Jitter on that level is audible.

See paper by Morgan Hirosuke Miki, Yoshihiro Ohtani, John Kowalski. 10ns and 20ns can be audible.

 

I personally participated in a double blind test where I could repeatedly distinguish music with a difference of 30ns, others in the test could differentiate levels lower than me.

 

If you extrapolate the BBC jitter tests to real world conditions (listening to music in a system setup and correlated jitter) it's reasonable to think 50ns would definitely be audible.

 

BTW, I saw an interview where Bob Stuart of Meridian said that in in-house double blind tests, he and a few other experienced listeners could reliably hear 8ns of jitter when added to selections.

 

"Hearing Jitter" clearly depends on the type of jitter, the type of music, and the experience of the listener. Experience shows that listeners who know what the effect of jitter sounds like, and in the types of music where the differences are most readily heard (e.g., percussion instruments, plucked strings, acoustic music) can hear the very low levels of jitter that I've described.

[alfe]

Or more accurately, didn't work, as it increased "Jitter" as seen on the "eye" waveform, IIRC ?

 

If the eye pattern is affected then it's a defocus, hard to read a disc after that:)

[Jud]

The link Jud posted said it was 10 times below the threshold of hearing.

 

 

Keith Johnson's measurements very unambiguously refute this claim. He measured differences that are 10 times below the threshold of hearing.

 

Since that isn't what Keith Johnson thinks, I don't feel "unambiguously" is exactly the right word. Perhaps "Unambiguously to me, who knows as much about this topic as Jud does about the 'RNA world' origin of life." In other words, your expressed doubts about Keith Johnson's measurements and what they showed are exactly as meaningful as it would be for me to write a critique of your papers. How many scientists or mathematicians show up at evolution blogs or climate blogs saying essentially, "I am not an expert in the field but I am a smart scientist/professional in a science-related field, so my doubts should be taken seriously"?

 

This is not meant to be confrontational; how would you evaluate the question about who should be taken seriously on this topic, you or Keith Johnson? That's what I'm doing, same as I would evaluate the question of whether you or I should be taken more seriously regarding discussions of the role of RNA in the origin of life.

 

Academic papers and DAC designers say the threshold is as low as ~20 picoseconds. On the other hand, there are A/B/X tests showing thresholds three orders of magnitude or more higher than that. So there is a huge discontinuity between what the papers say should be the threshold and designers use in their work versus the results of A/B/X testing.

[Tony Lauck]

Since that isn't what Keith Johnson thinks, I don't feel "unambiguously" is exactly the right word. Perhaps "Unambiguously to me, who knows as much about this topic as Jud does about the 'RNA world' origin of life." In other words, your expressed doubts about Keith Johnson's measurements and what they showed are exactly as meaningful as it would be for me to write a critique of your papers. How many scientists or mathematicians show up at evolution blogs or climate blogs saying essentially, "I am not an expert in the field but I am a smart scientist/professional in a science-related field, so my doubts should be taken seriously"?

 

This is not meant to be confrontational; how would you evaluate the question about who should be taken seriously on this topic, you or Keith Johnson? That's what I'm doing, same as I would evaluate the question of whether you or I should be taken more seriously regarding discussions of the role of RNA in the origin of life.

 

Academic papers and DAC designers say the threshold is as low as ~20 picoseconds. On the other hand, there are A/B/X tests showing thresholds three orders of magnitude or more higher than that. So there is a huge discontinuity between what the papers say should be the threshold and designers use in their work versus the results of A/B/X testing.

 

If there are objective measurements that change the audio output as a result of the green pen that are statistically reliable (compared to the noise in the results) and a reputable experimenter has made the experiments and reported the results, it strikes me as highly likely that there will be an effect (on the audio output). This removes any claims from the realm of nonsense, in my judgment. Were the spinning disk technology not utterly obsolete, there might even be reason to repeat KOJ's experiments.

 

As to hearing differences, experimental setups show thresholds for which audibility has been proven. A positive result establishes an upper bound on the real threshold. Even if this threshold is three orders of magnitude greater than the objective effects that were observed with the green pen, this does not provide any reason to doubt the possibility that there might have been an audible effect. The experimental tests are capable of showing positive results, but by their very nature they are incapable of showing negative results. They show that various subjects under various conditions did not apparently hear something under various conditions, but provide no evidence regarding other subjects, other conditions and other occasions.

 

As to a sonic change being an improvement or an impairment, this requires a reference for comparison. The very same sonic change in the CD player may be perceived differently with a different recording, a different listener, or a different playback system (e.g. amplifier - speaker wire - speaker -room setup).

 

Now if the claim had been that the quality of sound on an HDtracks.com download depended on the brand of Ethernet cable connecting the web hosting computer to the nearby router, then that would be an extraordinary claim.

 

Another claim that converting a WAV file to FLAC and then back to WAV results in WAV files sounding different also does not strike me as an extraordinary claim. Indeed, I have done these kinds of tests and found that while the second generation WAV files always had the same audio sample data in many cases the file contents were different (e.g. different headers, chunk structure). It is possible that these effects cause player software to operate differently, leading to a different noise environment in the computer system. It is even possible that two different WAV files that have the identical bit contents play differently. They will be stored in different physical locations on the the same (or different) disk volumes and they will have different file names. These are sufficient differences to allow for different activity in the computer. Without detailed investigation, it is an individual judgment whether people believing they heard differences were mistaken. (In this case, the person reporting audible differences is a well regarded recording engineer who obviously has superior listening skills otherwise she would not have a consistent track record of producing outstanding recordings.)

 

If one digs deep enough into these cases, it is usually possible to move past the "objective - subjective" split. Of course there are blowhards and fools on both sides of the debate, but such is the nature of humanity.

[crenca]

I meant "loaded", not "pejorative" - meaning just what I said - your definitions already answer your own question(s). Questions are like that, in that most questions are answered (or at least severely limited) by the presuppositions implied in the terms and the "direction" from which they were asked and the "direction" they are pointed to. Questions are much more like loaded guns than we usually give credit.

 

Again, "religion" (at least the classical 'big' religions that have rational adherents and apologists: Judaism, Christianity, Islam, Buddhism, etc.) are substantiated by "known means" to the adherents. The means are simply not known (and thus rejected) by those outside these religions...

[wgscott]

how would you evaluate the question about who should be taken seriously on this topic, you or Keith Johnson?

 

The personalities aren't relevant. Appeals to authority are not compelling reasons to accept or dismiss an assertion. I have no doubt he knows more about jitter than I do (as I know next to nothing). It is however quite possible I know more than an audio engineer about how to test a scientific hypothesis. (It is also quite possible I don't, which is why I am saying appeals to authority aren't useful.)

 

What is relevant is if the data are (a) a test of the hypothesis, (b) are statistically significant, and © are reproducible.

 

You are right. I am by no means an expert on this. But taking the results at face value, do they either support or refute the following assertion?

 

Claim: Coating the edges of a compact disc with a green marking pen will noticeably improve its sound quality.

 

Most people would consider a tiny measurable difference in the opposite direction to what one predicts on the basis of the assertion to be a refutation. Sure, you can introduce a special pleading argument that jitter sounds better, not worse, but that further undermines an already weak argument.

 

Why don't you see audio devices sold that increase jitter to make the music sound better?

 

I remain unconvinced. Sorry if that makes me an ignorant close-minded arrogant fool in your eyes. Science is a disbelief system.

[Tony Lauck]

The personalities aren't relevant.

 

What is relevant is if the data are (a) a test of the hypothesis, (b) are statistically significant, and © are reproducible.

 

You are right. I am by no means an expert on this. But taking the results at face value, do they either support or refute the following assertion?

 

 

 

I'm unconvinced.

 

If one is relying on results provided by another person, one needs to consider the possibility that the person providing the results might not be skilled, careful or honest. This will be true even if one relies only on experiments one has performed oneself, since one will have to rely on the equipment and theories behind the experiments and these, too, came from people subject to human foibles.

[wgscott]

Another claim that converting a WAV file to FLAC and then back to WAV results in WAV files sounding different also does not strike me as an extraordinary claim. Indeed, I have done these kinds of tests and found that while the second generation WAV files always had the same audio sample data in many cases the file contents were different (e.g. different headers, chunk structure). It is possible that these effects cause player software to operate differently, leading to a different noise environment in the computer system. It is even possible that two different WAV files that have the identical bit contents play differently. They will be stored in different physical locations on the the same (or different) disk volumes and they will have different file names. These are sufficient differences to allow for different activity in the computer. Without detailed investigation, it is an individual judgment whether people believing they heard differences were mistaken. (In this case, the person reporting audible differences is a well regarded recording engineer who obviously has superior listening skills otherwise she would not have a consistent track record of producing outstanding recordings.)

 

If one digs deep enough into these cases, it is usually possible to move past the "objective - subjective" split. Of course there are blowhards and fools on both sides of the debate, but such is the nature of humanity.

 

The claim that two wav files with identical md5sums stored on the same device can have audible differences when played back though the same playback system is the perfect poster child for an extraordinary claim. The files themselves are bit-identical, so the differences in header content you invoke don't even exist. They playback software strips off the container. You are now left with the idea that different sectors of the drive will sound different, or different filenames have audible differences that are magically carried over to the playback software even after being stripped away. These are all prima facia extremely unlikely to matter, and if they did matter, one wonders how it is at all possible to do any computer-based computations without having this variability sabotage the results. The appeal you thrown in to authority at the end of the list of speculations is similarly laughable.

[wgscott]

If one is relying on results provided by another person, one needs to consider the possibility that the person providing the results might not be skilled, careful or honest. This will be true even if one relies only on experiments one has performed oneself, since one will have to rely on the equipment and theories behind the experiments and these, too, came from people subject to human foibles.

 

Too bad. Find a different person or lab or whatever to try to reproduce the results. Non-reproducible results simply aren't compelling.

[mansr]

Another claim that converting a WAV file to FLAC and then back to WAV results in WAV files sounding different also does not strike me as an extraordinary claim. Indeed, I have done these kinds of tests and found that while the second generation WAV files always had the same audio sample data in many cases the file contents were different (e.g. different headers, chunk structure). It is possible that these effects cause player software to operate differently, leading to a different noise environment in the computer system. It is even possible that two different WAV files that have the identical bit contents play differently. They will be stored in different physical locations on the the same (or different) disk volumes and they will have different file names. These are sufficient differences to allow for different activity in the computer. Without detailed investigation, it is an individual judgment whether people believing they heard differences were mistaken.

 

This is not how computers work. Sorry.