Is Audiophiledom a confidence game?

[Tony Lauck]

On 10/13/2017 at 7:50 PM, wgscott said:

Ask SankyK under what conditions he would accept that his hypothesis that music files having identical checksums can sound different, depending on his past history, would be demonstrably wrong.  His answer is fairly telling.

 

I can very easily create two files with the identical bits in them that will pass all checksums, yet they will sound different.  This is a characteristic of my particular audio playback system which includes a computer, some hard drives, and a RAM disk.  The computer is located a few feet from my listening position and is under my right near field monitor.  If I have two bit-identical files, on one the RAM disk and one on the hard drive, when I play the RAM disk file I won't hear any acoustic noise from the spinning disk.  If I play the file on the hard drive I will hear noise from the (now) powered up disk drive.  This is unarguable and easily repeatable (however care has to be taken that the file system doesn't have the hard drive file contents cached in other parts of the RAM).

 

This one example completely demolishes the unrestricted claim that bit identical files must sound different and that people who hear differences between files with identical checksums are delusional. 

 

It would be more difficult to demonstrate differences were the computer with spinning disk outside of the listening room, but there is nothing, in principle, that would make this impossible.  The extent that the different electrical environment in the computer might affect the electrical signals to the speaker drivers would depend on details of the playback electronics.  (It would be easier to demonstrate measured differences in the analog output of the DAC under various conditions in the digital side of the DAC, as has been done with various USB regenerator devices, etc..)

 

[Tony Lauck]

8 minutes ago, Samuel T Cogley said:

Unless I'm misunderstanding, your argument is a total red herring.

It was a real-world example that shows the effect of imperfections of the playback chain  on the sound a listener hears.  I chose it because it was a blatant example of how differences in the representation of two files in a computer system can interact with defects in the playback system to affect what listeners hear, so blatant that there would be no calls for double blind testing, etc...  I could have chosen other examples, for example if I had two copies of the same file on the hard drive and one of them was badly fragmented it would be possible to hear the disk seeking.  (It was the seek noise that originally caused me to set up the RAM drive and play out of that.)

 

I admit this was a slippery slope argument.  IMO this was well deserved because the long history of some objectivists calling some subjectivists delusional over the very issue of playback of bit identical files.  (One of the earlier examples concerned the alleged difference between playing back WAV files vs. playing back WAV files that have been polluted by being converted to FLAC and then back to WAV.)

 

By the way, I am an objectivist when criticizing subjectivists reaching nonsensical conclusions from what they have heard and I am a subjectivist when dealing with dogmatic statements from objectivists, especially those who don't even understand the underlying technology they are using and therefore can be flat out wrong in their scientific conclusions in some cases.  I am more than OK if I sometimes irritate both groups.

 

[Tony Lauck]

9 minutes ago, Speed Racer said:

 

WAV and FLAC are containers that hold data. When done properly, converting from WAV to FLAC, the container changes but the data stays the same. The same is true when converting from FLAC to WAV. Explain how the data is polluted when changing the container.

 

The data is not polluted.  However, the result is a new file.  As in the example I gave earlier the new file will be stored on a different place in the computer system and may therefore sound different when played back.  (It shouldn't, but it might.)  In addition, many times the conversion software changes the output WAV file in the process WAV to FLAC to WAV.  Examining the audio samples in the two WAV files one will see that they contain the identical samples, but the two WAV files have different checksums.  This may be related to changing ways the metadata is stored or how the chunks are laid out on the individual file blocks.  Again, the meta data should not affect the playback of the identical audio samples, but who can say what software may do and how the software can interact with various aspects of the underlying hardware.

 

It's a separate question as to whether real-time playback of a FLAC file sounds different from converting the file to WAV and then playing back the resultant WAV file in real time. Again, this should not make any difference, but on some systems it does.

[Tony Lauck]

8 minutes ago, kumakuma said:

Alex is saying that he is able to generate two bit identical files that sound different when played under identical conditions (same playback chain, same media, same everything). His theory is that noise is somehow embedded in one of the files and travels along with it when it copied to a USB drive and sent to the UK or downloaded over the Internet.

 

Glad to hear any theories you have on how this is possible. 

As I recall, he provided insufficient details as to how these files were produced.  This makes it impossible for anyone to replicate the experiment.  As such, and after wasting a fair amount of time, I  consider this a waste of time.

 

I don't doubt that two USB sticks might sound different.  One could even have a virus on it that hacked my computer.  Again, this is an extreme case, just to show that solid controls are necessary as part of any scientific experiment.  In addition, a useful experiment would need to show when these differences were possible.  This could lead to producing better playback gear (that eliminates said effects if they are real) or alternatively results in studios producing better quality recordings.  But this won't happen if the people involve wear audiophile, music lover, objectivist or subjectivist hats.  They must put these hats away and put on their scientist hat, if they have one.

 

If someone actually gave me two bit identical files that sounded different after I controlled for location and other effects then I would waste no effort getting to the bottom of it, just as I did with the question of hard drive noise.

 

[Tony Lauck]

Just now, Samuel T Cogley said:

 

I think you mean a non-ideal listening environment.  A noisy listening room proves nothing.

The noise was from the audio equipment.  I would agree with you if the noise came elsewhere.  I see no difference between noise that comes through the air from a component other than the speaker or if the noise originates elsewhere in the system end comes out through the speakers.   My example is no different from problems I recall ages ago, such as noisy tape reels when playing reel to reel tape,  needle talk from an LP cartridge, etc...

 

[Tony Lauck]

On 10/18/2017 at 8:07 PM, kumakuma said:

 

Copying the file to memory for playback would also eliminate any "embedded electrical differences".

 

This would depend on the particular memory technology involved.  However, some memory technology with which I am familiar has the potential to represent a digital one (or zero) with more than one distinct physical representation.  We are not at the level of using individual quantum states to encode bits, and even if we did we would need to use error correcting codes to make the system reliable.  And these permit multiple representions (e.g. as bits) of the same signal.

 

Bits as bits appear in the domain of mathematics.  Zeros and ones are a platonic abstraction of the underlying physics with is continuous and analog (as far as we know).  If you don't understand both the physics and the actual engineering involved then I would say you would need to restate your sentence, to read "would also probably eliminate".  Whether this was true or not would depend on details of the implementation, as engineered and manufactured.

[Tony Lauck]

On 10/18/2017 at 8:14 PM, lucretius said:

 

I don't know about your CDs, but there is no electricity hidden in mine.

Actually, your CDs have various bit errors encoded into the stream of pits.  This is almost certainly the case.  There are various levels of errors that can occur.  Some are eliminated by the first order error correction, C1 errors, some are eliminated by the second level of error correction, C2 errors, and if your disk is defective or damaged some may not be correctable and result in masking the error (by interpolating a bad sample).  If a disk is really damaged then playback will audibly glitch or even stop.

 

I have tools that work with my CD player to assess the quality of CD-R disks that I have burned.  This is basic QC that mastering engineers must do when they deliver physical disks to the pressing plant.  (This has largely been superseded by sending the bits to the pressing plant over the Internet, during which, FWIW, the encoded samples will be subject to various stages of compression and expansion.)

[Tony Lauck]

On 10/18/2017 at 8:45 PM, Speed Racer said:

 

Of course I have. In digital electronics, a 1 is a 1 and a 0 is a 0. You are talking about how these 1's an 0's are represented on the wire which can vary quite a bit and still be read as a 1 or a 0. You are confusing analog with digital here. The reason computers work is because the signals representing 1's and 0's can vary in amplitude and STILL be identical as far as the digital electronics are concerned.

There is a big difference between introductory study of a technology and getting into the actual theoretical and practical details involved in engineering cost effective products with state of the art performance.  Based on reading your few posts, my conclusion is that you don't know what you don't know.  A little knowledge is a dangerous thing...

[Tony Lauck]

On 10/19/2017 at 2:46 AM, marce said:

Not really we all know the real claim, Tony's answer is not based on the real claim, so is nul in this instance, even though it stands on its own.

My point was to illustrate the necessity of extreme precision in what is being claimed in these discussions.  It is only after my post that people dissed it, saying of course we don't want to count acoustic noise coupling, only electrical noise coupling.  And then, that is not enough either, because is it electrical noise at the speakers or electrical noise at the output of the DAC. 

 

Amirm has given an example where one particular USB "cleaning" device affected the measured output of a DAC, when the DAC was connected to his particular test equipment.  (In this case the device was measured as actually degrading the output signal of the DAC, not improving it.)  As it happened, this DAC had two output modes, single ended and balanced.  The degradation was observed only in the case of the single ended outputs, not when balanced outputs were measured.  So even something as specific as "output of the DAC" is not sufficient to define the experimental conditions. 

 

The subjectivists are right when they focus on how the system sounds.  This is the end goal.  The narrow minded objectivists who focus on one component at a time will never succeed in getting to the bottom of the problem because they have not properly defined the problem they are solving.  (And this certainly includes all of the people who post on either side, just for the sake of argument.)

[Tony Lauck]

1 hour ago, crenca said:

 

It is not a "view" caused by or circumscribed by a technical reality, it is a culture & a confidence game (perhaps "confidence market") which begins with a "art & wine" radical subjectivity.  The technical reality is simply something to be manipulated, affirmed, ignored, etc. at the pleasure of other things...

The culture and confidence game, with regard to digital audio, began with the adoption of the 44/16 PCM standard for CDs.  It was already known empirically that this format was inferior to formats with higher sampling rates, such as the 50 KHz rate used by the Soundstream recorder.   And yet we got the lying marketing, "Perfect sound forever".

 

In the all analog world everyone knew that analog devices were imperfect.  The objectivist vs. subjectivist split became possible once the technician quality engineers latched on to the phrase "bits are just bits" and started calling the people who listen and heard differences delusional.

 

There were scammers from the very beginning of audio.  And since the very beginning of people trading goods and services.  Nothing new here.  The most common technique used to be arranging the equipment being pushed to play back music at a louder volume, thereby sounding better.  Other techniques included playing different program material with different products.  Nothing new here.  These techniques are still being used today.

 

[Tony Lauck]

On 10/20/2017 at 6:11 PM, barrows said:

Of course, there will always be room for improvement: this is like the tangent line (is that what it is called, forget my math), which gets closer and closer to the axis, but never actually meets: same thing as diminishing returns...

 

For each individual, where they might stop, on this pursuit, will be different.  If they stop at all...

One of the principal aspects of this life is the potentially infinite gap between "need" and "want" for those who are exceptionally skilled at making themselves miserable.