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

[Tony Lauck]

This is not how computers work. Sorry.

 

Really, now. I know how computers work. I was one of the top engineers at a large computer manufacturing company for several decades and I am quite familiar with how computers actually work, especially in the weird cases that seem impossible until the explanation is finally found after many sleepless weeks. My peers and I used to exchange stories of our most difficult problems.

 

As to file activity, it is also the case that I have heard two bit identical files sound different when played and I traced down the cause of the difference. When one file was played there were audible noises, when the other was played there were none. It so happened that one file had been fragmented on disk and I could hear the file seeks from the disk drive.

[wgscott]

I can hear my disk drive when album art loads on my iPad. This does not prove anything.

 

If you can hear how fragmented your disk is via your stereo system's speakers, then there might be something very seriously wrong with your electrical isolation.

 

You might want to try an optical cable.

 

I hear the one with wolves on it sounds better.

 

By the way, I am really impressed you can hear differences in file name conventions. What font do you find sounds the best? Can you hear a difference between ascii and non-ascii characters?

[crenca]

 

As to file activity, it is also the case that I have heard two bit identical files sound different when played and I traced down the cause of the difference. When one file was played there were audible noises, when the other was played there were none. It so happened that one file had been fragmented on disk and I could hear the file seeks from the disk drive.

 

How does one "hear" file seeks (i.e. the electro/mechanical/time dely of data, etc.)? Not trying to be confrontational, honest question. In other words, in what way was the file seek *changing* the data (bit stream) so that the DAC *changed* it's analogue output? Some sort of time delay (a real pause in the output of the DAC) is the only thing that makes a lick of sense to me...

[Tony Lauck]

I can hear my disk drive when album art loads on my iPad. This does not prove anything.

 

If you can hear how fragmented your disk is via your stereo system's speakers, then there might be something very seriously wrong with your electrical isolation.

 

You might want to try an optical cable.

 

I hear the one with wolves on it sounds better.

 

By the way, I am really impressed you can hear differences in file name conventions. What font do you find sounds the best? Can you hear a difference between ascii and non-ascii characters?

 

As I expected from those of your ilk, your reading comprehension was somewhat lacking. I heard the seek noise acoustically, not through my speakers. I have a near field setup and all my equipment is within a few meters. The clicking noise is obvious with the volume muted, when running disk self-test, etc... It is possible to hear the acoustic noise during pianissimo passages of classical music. I could replace the clunking disk with a 4 TB SSD for $$$, but I instead I convert files on disk from FLAC to WAV and store the WAV in a RAM disk. Nothing moves, no noise.

 

I never said that I heard differences caused by file name changes. However, I am quite certain that it would be possible to code up player software that would produce subtle sonic degradation depending on the choice of file name. In addition, it is entirely possible that the differences might manifest in one computer and not in another one, or one time and not another time. If you know how modern processors work, you would appreciate how this might be possible. It is also possible, and more likely than the file name differences, that the same file, stored on the same place and played with the same software on the same computer might provide different sound on one play vs. another. The possibilities would be obvious to one skilled in the art of computer systems performance.

[mansr]

How does one "hear" file seeks (i.e. the electro/mechanical/time dely of data, etc.)? Not trying to be confrontational, honest question. In other words, in what way was the file seek *changing* the data (bit stream) so that the DAC *changed* it's analogue output? Some sort of time delay (a real pause in the output of the DAC) is the only thing that makes a lick of sense to me...

 

Aside from mechanical sounds emanating directly from the hard drives, any disk activity results in a burst of electrical noise within the computer, as does every other component (Ethernet, USB, etc.). I once had a computer (made around 1995) with a built-in sound card so poor that it picked up all manner of noise. SCSI activity was particularly bad, but even DRAM accesses (as opposed to cache hits) had their own signature. I'm guessing Tony had a similarly poor system when this experience of his took place. Modern sound cards, even on-board ones, are much better, and an external DAC of any kind is better still.

 

The only conclusion we can draw from anecdotes like this is that sometimes noise sources within the computer can make their way into the analogue audio outputs. That's a far cry from proving that FLAC and WAV files sound differently as a matter of principle. If file fragmentation is making an audible difference, there's no way of predicting which files will suffer.

[crenca]

 

I never said that I heard differences caused by file name changes. However, I am quite certain that it would be possible to code up player software that would produce subtle sonic degradation depending on the choice of file name. In addition, it is entirely possible that the differences might manifest in one computer and not in another one, or one time and not another time. If you know how modern processors work, you would appreciate how this might be possible. It is also possible, and more likely than the file name differences, that the same file, stored on the same place and played with the same software on the same computer might provide different sound on one play vs. another. The possibilities would be obvious to one skilled in the art of computer systems performance.

 

Tony, thanks for the clarification - I also thought you were saying that heard a difference through your DAC.

 

I know very little about how compilers work, but I do understand that byte code and how a modern processor processes it can "vary". However, could you say more. I want to say that despite these variations in compiled byte code and the differing way's processors can process (depending for example load, what else is in memory, variations between models let alone batches/ manufactures/etc.) is not nearly enough to account for a difference heard through an audio chain that includes said processor. Otherwise, we would notice more variation in how our other software works (high leve, suc as a word processor, or even something much lower).

[mansr]

As I expected from those of your ilk, your reading comprehension was somewhat lacking. I heard the seek noise acoustically, not through my speakers. I have a near field setup and all my equipment is within a few meters. The clicking noise is obvious with the volume muted, when running disk self-test, etc... It is possible to hear the acoustic noise during pianissimo passages of classical music. I could replace the clunking disk with a 4 TB SSD for $$$, but I instead I convert files on disk from FLAC to WAV and store the WAV in a RAM disk. Nothing moves, no noise.

 

Of course the mechanical movement of the drive heads is audible if you're close enough. Why is this relevant here?

 

I never said that I heard differences caused by file name changes. However, I am quite certain that it would be possible to code up player software that would produce subtle sonic degradation depending on the choice of file name. In addition, it is entirely possible that the differences might manifest in one computer and not in another one, or one time and not another time. If you know how modern processors work, you would appreciate how this might be possible. It is also possible, and more likely than the file name differences, that the same file, stored on the same place and played with the same software on the same computer might provide different sound on one play vs. another. The possibilities would be obvious to one skilled in the art of computer systems performance.

 

Any processing of the filename will be long since completed by the time the first sample hits the DAC.

[crenca]

The only conclusion we can draw from anecdotes like this is that sometimes noise sources within the computer can make their way into the analogue audio outputs. That's a far cry from proving that FLAC and WAV files sound differently as a matter of principle. If file fragmentation is making an audible difference, there's no way of predicting which files will suffer.

 

I hear ya I agree, plain old fashioned RF/elctro noise can be passed on in diverse ways - but that is not a point of contention (at least not one with any basic understanding of computers, electronics, and audio gear). What is asserted that is much more radical is just what you note. I personally am struck by the claims about USB cables, how they "nail the mid range" and what not digitally. A I understand it, these folks are not claiming isolation from traditional noise - they are claiming a "digital" fidelity that "nails the mid range" over against one that "has an expanded sound stage"...in the digital realm of a bit stream (separate from DAC and it's interpretive/conversion/analogue functions).

[wgscott]

Tony, thanks for the clarification - I also thought you were saying that heard a difference through your DAC.).

 

Amusingly, I actually understood what he meant, despite what he then said in reply.

 

Likewise, I can hear my external drive make noise when my album art loads, if inam up close to it. I was suggesting neither observation is relevant. But maybe I just lack Tony's superior computer skills, as he suggests.

[gmgraves]

Yes! and here you get what you pay for, or to say another way, there is a reasonable price/weight!

 

I think my quote of Nelson Pass above was too subtle: the specific reason that he advocates only buying amps that have an appropriate weight for the price is precisely because it is well known that good transformers, chokes, capacitors etc are heavy! ... and these are the actual key to getting good power not magic wires.

 

 

Of course, but in the pursuit of "better sound" many audiophiles (I know of one such, in fact) would opt for the belt-and-braces approach of a Furman power conditioner PLUS Shunyata Research power cable connected to a Pass amplifier. "Can't be too careful" is his reasoning.

[gmgraves]

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.

 

 

Yeah, but I'm not clinging to anything. I heard no difference 15 or so years ago, and never revisited the subject, and was not aware that any engineering explanation validating the pen existed. Someone else posting here agreed with you that someone did extensive tests on the pen and found that it did, indeed affect the reading of the CD, and it introduced jitter thereby giving a negative effect. I have to say, that not only did not hear any improvement with the pen, but I heard no degradation of the sound using the pen either.

[Tony Lauck]

Of course the mechanical movement of the drive heads is audible if you're close enough. Why is this relevant here?

 

 

 

Any processing of the filename will be long since completed by the time the first sample hits the DAC.

 

The discussion centered on what people were hearing. People hear with their ears, i.e. the acoustic output of the system. This includes the acoustic output of the loudspeakers, but it also includes other acoustic outputs, such as noises created by power transformers, cooling fans, motors of various sorts in hard drives, etc... All of these possible paths must be considered and carefully evaluated before reaching a conclusion that some audible difference is "impossible". Failure to do this amounts to a slip-shod approach, not even good engineering and certainly not science. Usually, it's the non-technical "subjectivists" who latch onto some "cause" posited by a product marketing type as to why they are hearing differences, and the "objectivists" ridicule their non-sensical explanation as evidence of delusional perception. Sometimes it is the "objectivists" who provide incomplete analysis of how things work and reach false conclusions that some perception must be "impossible" and hence could not have actually been heard. When this happens, it discredits the entire "objective" approach. Both camps can suffer from excessive confidence (ego) in their own abilities. Me, I've been (rightly) accused of being a "fence straddler" in these debates. I have learned through many decades that things are not always what they seem. This started in a high-school physics class when I discovered that my experiment that used a pendulum and electric stop watch to supposedly measure the acceleration had, instead, been measuring the power line frequency.

 

One would think, one would hope, that the file name would have no influence in the playback of music. But that may not necessarily be the case. As an extreme case, admittedly unfair, several years back a number of computer audio players had problems with some file names that used non US character sets (e.g. European character sets or Asian character sets). In some cases, certain tracks would not play at all, an obvious bug. Good software engineering practice would have isolated the file opening portion of the software with the musical playback, but if there is to be gapless playback, it will be sometimes necessary to lookup file names and open files for the next track in a play list while a previous track is playing. Here there may be differing code paths due to character sets, portions of file directories that are in memory vs. have to be paged in, processor cache collisions, and other factors that may not be obvious that are part and parcel of today's computing environments.

 

If one does not want to remain locked in epistemological dead-ends, one needs to retain an open mind and struggle persistently to rid oneself of dogmatic beliefs.

[gmgraves]

I can hear my disk drive when album art loads on my iPad. This does not prove anything.

 

If you don't mind me asking, what kind of an iPad has a mechanical disk drive? Do you mean one of the older iPods?

[mansr]

One would think, one would hope, that the file name would have no influence in the playback of music. But that may not necessarily be the case. As an extreme case, admittedly unfair, several years back a number of computer audio players had problems with some file names that used non US character sets (e.g. European character sets or Asian character sets). In some cases, certain tracks would not play at all, an obvious bug.

 

I've occasionally run into such bugs. In every instance it has been a case of the application somehow ending up requesting a file using a different encoding than the actual filename was stored with on disk. The result is that the application fails to open the file at all. There is no possibility of successfully opening a file and then having the audio contained within sound differently based on the filename.

 

If you believe the filename can affect the sound coming out of the speakers, there's a simple way to test that. All filesystems typically used on hard drives today (NTFS, anything on Linux/Unix, whatever Apple calls theirs) support hard links, a mechanism whereby the exact same data blocks are accessible through several different names. To test the proposition, take a music file and hard-link it to a bunch of different names, then play them and see if there's a difference. Since they refer to exactly the same blocks, fragmentation and such will not be a factor.

[Tony Lauck]

I've occasionally run into such bugs. In every instance it has been a case of the application somehow ending up requesting a file using a different encoding than the actual filename was stored with on disk. The result is that the application fails to open the file at all. There is no possibility of successfully opening a file and then having the audio contained within sound differently based on the filename.

 

If you believe the filename can affect the sound coming out of the speakers, there's a simple way to test that. All filesystems typically used on hard drives today (NTFS, anything on Linux/Unix, whatever Apple calls theirs) support hard links, a mechanism whereby the exact same data blocks are accessible through several different names. To test the proposition, take a music file and hard-link it to a bunch of different names, then play them and see if there's a difference. Since they refer to exactly the same blocks, fragmentation and such will not be a factor.

 

Personally, I have little interest in such investigations. That there might even be necessary is an indication that the DAC designer has not done a good job. Bits are supposed to be just bits, and making this so is the function of the DAC as this component is where the digital / analog boundary is in the playback chain.

 

So long as reviewers insist on using good transports to evaluate DACs there will be a continued lack of pressure on the DAC manufacturers to improve the input portion of their products. This strikes me as really stupid, because if they can successfully market a DAC that is source immune they will be able to capture a larger fraction of the audiophiles' budgets.

[alfe]

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.

 

You know that it's possible to make polycarbonate out of corn, chlorophyll increase the jitter

 

Reflected Near-Infrared Waves - Mission:Science

[Daudio]

Well, this thread seems to have seperated the wheat from the chaff ! (Jud and Tony vs. just about everyone else)

 

[Teresa]

Jud, I'm a big fan of Prof. Keith Johnson's recordings, he builds much of his own recording equipment and modifies recording equipment made by others. I feel the proof of his expertise is the results of his recordings which I find to be among the most realistic currently in the world. So if he says the green increases jitter, I tend to believe him.

 

I was a fan of the green pen at first, until I began thinking of how it behaved in my friends $300 mini-system, every disc she treated sounded almost twice as loud, I heard it with great surprise.

 

In my system I didn't notice any increase level, instead I noticed deeper and fuller bass which made the treble less anemic and less strident. Once I tried to equalize the levels the treated and untreated sounded more alike. I have no clue why the green pen would increase levels unless somehow it was restricting dynamic range. It's a real enigma.

[mansr]

I was a fan of the green pen at first, until I began thinking of how it behaved in my friends $300 mini-system, every disc she treated sounded almost twice as loud, I heard it with great surprise.

 

That's impossible.

[Teresa]

That's impossible.

 

I assure you it is not! The effect was quite real, and Kathy loved that she could finally get decent volume from her small mini-system. I am curious how the green pen accomplished this.

 

P.S. Based on the results with her mini-system I have a hunch that the green pen treatment appeared to sound superior due to the overall louder level since human ears select the louder sound as superior. That is why comparisons have to be level matched.

[Boris75]

(...) if (DAC manufacturers) can successfully market a DAC that is source immune they will be able to capture a larger fraction of the audiophiles' budgets.

 

Benchmark and Exasound are following this strategy, and I am sure there are others.

[Jud]

The personalities aren't relevant.

 

Completely agreed.

 

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.)

 

There's a distinction between appeals to authority and crediting specialized expertise in technical subject matter. Citing Linus Pauling about Vitamin C is the former. Citing Linus Pauling about molecular structures is the latter. What I'm asking here is not for us to believe Keith Johnson because he's a Smart Guy. I'm asking that we credit the fact that having designed a couple of the world's most highly thought of CD players, he knows how to run jitter measurements on a CD player. Did he do these particular measurements as a slapdash, one-off thing? That wouldn't match up with his reputation or the performance of the products he's responsible for, but there we *are* into appeals to authority; barring something I haven't seen in the rest of the video or elsewhere, I simply don't know how many runs he performed and what the protocol was.

 

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?

 

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?

 

Sorry that the quoting mechanism didn't pick up the text of the assertion, but I don't think the assertion was framed correctly anyway. There are actually two assertions:

 

Objectivist assertion: It is physically impossible for green pen to affect the sound of CD playback.

 

Subjectivist assertion: Green pen makes CD playback sound better.

 

Keith Johnson's measurements demonstrate a small increase in jitter. Whether such an increase is audible is a subject of dispute. But now I submit we are out of the realm of "The universe doesn't work that way," and on to "Can people hear this?" Thus, to me, Johnson's results are an effective refutation of the assertion that it is ludicrous due to physical impossibility to think green pen might affect the sound of CD playback.

 

Johnson's results also, for me, effectively refute the subjectivist assertion. They say any effect would cause greater distortion rather than an objective improvement.

 

Note objecti​ve improvement. There are many production techniques used in recording that are actually forms of distortion which are known to increase subjective enjoyment of the sound by many people. (Jimmy Iovine, now of Beats, used to use a lot of something called the Aphex Aural Exciter back in his engineering days.) So yes, there are in fact devices used to increase distortion that make music sound subjectively better. As far as "devices sold that increase jitter" - well, that was apparently (though unintentionally) the green pen, eh? (Forum member barrows was present at PS Audio internal jitter testing where some test subjects liked slightly higher jitter better.)

 

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

 

Nah, of course not (well, maybe I wavered a little on "closed-minded").

[mansr]

What I'm asking here is not for us to believe Keith Johnson because he's a Smart Guy. I'm asking that we credit the fact that having designed a couple of the world's most highly thought of CD players, he knows how to run jitter measurements on a CD player. Did he do these particular measurements as a slapdash, one-off thing? That wouldn't match up with his reputation or the performance of the products he's responsible for, but there we *are* into appeals to authority; barring something I haven't seen in the rest of the video or elsewhere, I simply don't know how many runs he performed and what the protocol was.

 

If assuming Keith Johnson is infallible isn't appeal to authority, I don't know what is.

 

Note objecti​ve improvement. There are many production techniques used in recording that are actually forms of distortion which are known to increase subjective enjoyment of the sound by many people. (Jimmy Iovine, now of Beats, used to use a lot of something called the Aphex Aural Exciter back in his engineering days.) So yes, there are in fact devices used to increase distortion that make music sound subjectively better.

 

Comparing to production techniques is a mistake. Sometimes an artist will use a vocoder to obtain the sound he wants. That does not, as I'm sure you'll agree, imply that a vocoder will make all music sound better. Same for reverb, etc.

[Jud]

If assuming Keith Johnson is infallible isn't appeal to authority, I don't know what is.

 

 

 

Comparing to production techniques is a mistake. Sometimes an artist will use a vocoder to obtain the sound he wants. That does not, as I'm sure you'll agree, imply that a vocoder will make all music sound better. Same for reverb, etc.

 

Speaking of distortion: How exactly does saying someone knows how to run a jitter test on a CD player equate to saying he's infallible?

 

Speaking of distortion again: Where did I ever make a statement that a given form of distortion made "all music sound better"?

 

If you want to engage in constructive discussion as opposed to exaggeration for the sake of trying to make rhetorical points, I'll be right here.

[Tony Lauck]

 

Johnson's results also, for me, effectively refute the subjectivist assertion. They say any effect would cause greater distortion rather than an objective improvement.

 

I agree with your entire post, except this brief portion. You are making an unwarranted assumption that increased distortion will necessarily be perceived as an impairment. If you had spent many hundreds of hours doing audio restoration work as I have, or if you were a professional recording engineer, especially a mastering engineer, you would appreciate that sometimes adding "distortion" can sometimes subjectively improve sound.

 

It is even possible (albeit unlikely) that converting a file to MP3 and converting it back can subjectively improve it. (Beware, even the same listener with the same equipment may reach one conclusion immediately, only to realize on extended listening that this conclusion was incorrect. ) For this reason in any subjective vs. objective debates it's safest to confine listening tests to perception of differences rather than preferences.