Is Audiophiledom a confidence game?

[jabbr]

32 minutes ago, mansr said:

You may only analyse the analogue output of the DAC or downstream components.

Right so pick a very dense miniature but cheap DAC. USB. Power supply droop and ground bounce affect the output. Electrically measurable. 

 

Result of cost cuts and design considerations, less bypass caps, smaller, higher ESL caps, etc. etc. 

 

Yes these issues can and are designed away. Motherboards are advertising better SQ for gaming etc. Everthing is market driven. If good enough sound sells then you get good enough sound.

[Samuel T Cogley]

48 minutes ago, jabbr said:

I would take a cheap, but high speed, motherboard, get a cheap HDD and a really crappy PSU.

 

One copy of the file would be heavily fragmented, each block max distance from the next. The other copy would be unfragmented. The trial would be arranged such that the file was not cached.

 

The scope would look at the pattern of HDD seeks as evidenced by power draw. 

 

What's the goal here?  To construct a underperforming PC with the intention of creating an environment that might explain why people of high audiophile esteem claimed to hear differences in bit-identical files in the 2000s?

 

Your proposed test environment is an environment that a typical "computer audiophile" would avoid like the plague.

 

I concede that there is a set of "computer audiophile best practices" that has evolved over time and this very web site is a big reason why.  But getting back to the thread topic, those best practices do not require $5,000 USB cables or $10,000 Ethernet cables and the results of your hypothetical (and IHMO, rather dubious) tests do not in any way validate the "need" for those cables.

[jabbr]

15 minutes ago, Samuel T Cogley said:

 

What's the goal here?  To construct a underperforming PC with the intention of creating an environment that might explain why people of high audiophile esteem claimed to hear differences in bit-identical files in the 2000s?

 

Your proposed test environment is an environment that a typical "computer audiophile" would avoid like the plague.

 

I concede that there is a set of "computer audiophile best practices" that has evolved over time and this very web site is a big reason why.  But getting back to the thread topic, those best practices do not require $5,000 USB cables or $10,000 Ethernet cables and the results of your hypothetical (and IHMO, rather dubious) tests do not in any way validate the "need" for those cables.

That’s essentially the point. The core of computer audiophile should be a set of best practices. Those aren’t a con game. I can explain why, in electrical terms, those best practices are best. I say this because among other things I’ve read books that are available to anyone.

 

The voodoo cables are to be avoided and we can say exactly why. 

[Samuel T Cogley]

2 minutes ago, jabbr said:

That’s essentially the point. The core of computer audiophile should be a set of best practices. Those aren’t a con game. I can explain why, in electrical terms, those best practices are best. I say this because among other things I’ve read books that are available to anyone.

 

The voodoo cables are to be avoided and we can say exactly why. 

 

Ok, so we agree.  I'm flummoxed. 

 

[lucretius]

1 hour ago, jabbr said:

 

That at was an ode to plisskens offer.

[lucretius]

1 hour ago, jabbr said:

The scope would look at the pattern of HDD seeks as evidenced by power draw. 

 

I'm beginning to see that this is related to dirty power.

[crenca]

3 hours ago, jabbr said:

That’s essentially the point. The core of computer audiophile should be a set of best practices. Those aren’t a con game. I can explain why, in electrical terms, those best practices are best. I say this because among other things I’ve read books that are available to anyone.

 

The voodoo cables are to be avoided and we can say exactly why. 

 

3 hours ago, Samuel T Cogley said:

 

Ok, so we agree.  I'm flummoxed. 

 

 

As far as I can tell, jabbr's point (as well as Tony's, etc.) is not that two bit identical files are inherently different - it is that in real world electronics lots of things can and do happen that can lead folks (particular "subjectivists) to hear differences in actual playback chains (i.e. they are all "suboptimal").  Upstream he says this is a "crucial" point, and that he "holds objectivists to a higher standard".  In other words, he is on a bit of a crusade (though he won't put it that way).  He also seems to believe all this is important to the root cause(s) of the subjectivist vs objectivist debate and whether Audiophiledom is a confidence game.

 

He is wrong.  Not in the particular, but in its relative importance to the Audiophile Confidence Game. "the con" does not rest on this or that particular "misunderstanding" around digital, real world non-ideal implementation, and the all too real "grey areas" in any techno endeavour.  These are proximate and even contributing realities, but they are in no way sufficient or and while I would not call the reality that he points to "un-important", it is at the end of the day what the con game rests on.

 

Now, this is not to say that digital audio is not something that is not resistant to the con  - it's just that the truth of all this is not going to move the debate in any significant direction or in any way upset the status quo.  There is too much inertia built in.

 

What will is changing demographics, major market disruptions, a consumer awareness, etc. etc.

[esldude]

2 hours ago, crenca said:

 

 

As far as I can tell, jabbr's point (as well as Tony's, etc.) is not that two bit identical files are inherently different - it is that in real world electronics lots of things can and do happen that can lead folks (particular "subjectivists) to hear differences in actual playback chains (i.e. they are all "suboptimal").  Upstream he says this is a "crucial" point, and that he "holds objectivists to a higher standard".  In other words, he is on a bit of a crusade (though he won't put it that way).  He also seems to believe all this is important to the root cause(s) of the subjectivist vs objectivist debate and whether Audiophiledom is a confidence game.

 

He is wrong.  Not in the particular, but in its relative importance to the Audiophile Confidence Game. "the con" does not rest on this or that particular "misunderstanding" around digital, real world non-ideal implementation, and the all too real "grey areas" in any techno endeavour.  These are proximate and even contributing realities, but they are in no way sufficient or and while I would not call the reality that he points to "un-important", it is at the end of the day what the con game rests on.

 

Now, this is not to say that digital audio is not something that is not resistant to the con  - it's just that the truth of all this is not going to move the debate in any significant direction or in any way upset the status quo.  There is too much inertia built in.

 

What will is changing demographics, major market disruptions, a consumer awareness, etc. etc.

Excellent summary post.  

[marce]

7 hours ago, jabbr said:

If you talk to people working in the field, in order to achieve increases in bit density require real work, physics. Moore’s Law doesn’t just happen but requires real behind the scenes work. What they do to make advances in these days is amazing (to me).

If you look at it that way, there is some amazing stuff, I thought you were on about signal propagation specifically.

[crenca]

2 hours ago, crenca said:

 

 

As far as I can tell, jabbr's point (as well as Tony's, etc.) is not that two bit identical files are inherently different - it is that in real world electronics lots of things can and do happen that can lead folks (particular "subjectivists) to hear differences in actual playback chains (i.e. they are all "suboptimal").  Upstream he says this is a "crucial" point, and that he "holds objectivists to a higher standard".  In other words, he is on a bit of a crusade (though he won't put it that way).  He also seems to believe all this is important to the root cause(s) of the subjectivist vs objectivist debate and whether Audiophiledom is a confidence game.

 

He is wrong.  Not in the particular, but in its relative importance to the Audiophile Confidence Game. "the con" does not rest on this or that particular "misunderstanding" around digital, real world non-ideal implementation, and the all too real "grey areas" in any techno endeavour.  These are proximate and even contributing realities, but they are in no way sufficient or and while I would not call the reality that he points to "un-important", it is at the end of the day what the con game rests on.

 

Now, this is not to say that digital audio is not something that is not resistant to the con  - it's just that the truth of all this is not going to move the debate in any significant direction or in any way upset the status quo.  There is too much inertia built in.

 

What will is changing demographics, major market disruptions, a consumer awareness, etc. etc.

 

should have read over it:

 

"while I would not call the reality that he points to "un-important", it is at the end of the day what the con game rests on."

 

should read

 

"while I would not call the reality that he points to "un-important", it is at the end of the day not what the con game rests on.

 

&

 

"Now, this is not to say that digital audio is not something that is not resistant to the con"

 

whhaaattt??    

 

I am trying to say that while digital audio is in fact more resistant to the con game culture than just about all the rest of audiophiledom, it is in not here where the crux of the matter lies...

[Ralf11]

Moore’s "Law" is really just an Empirical Generalization

 

... just sayin'

 

...keepin' it real

[fas42]

4 hours ago, crenca said:

I am trying to say that while digital audio is in fact more resistant to the con game culture than just about all the rest of audiophiledom, it is in not here where the crux of the matter lies...

 

I would say the crux of the matter is that using conventional, "objectivist" thinking creates audio systems which are always "sub-optimal" - currently! Which leaves a gaping hole that one can drive a train through - perfect for a "con", if one wishes to call it that... so, plug that hole - and then all the madness goes away ...

[jabbr]

6 hours ago, marce said:

If you look at it that way, there is some amazing stuff, I thought you were on about signal propagation specifically.

Right, for example suppose there was someone who works on something like this: https://www.anandtech.com/show/11925/western-digital-stuns-storage-industry-with-mamr-breakthrough-for-nextgen-hdds -- not exactly this because of confidentiality but close enough to make the point here, and suppose this engineer/scientist had something to say about "noise" or variable physical encoding of bits in a file on a hard drive, and their effects on voltages occuring during readout -- would you believe this individual regarding "bits are bits" on a file?

 

Hint: I would take this information seriously. (I did).

[Ralf11]

why doesn't a buffer cure any of those problems?

[wgscott]

The comments are delightful:  https://www.anandtech.com/comments/11925/western-digital-stuns-storage-industry-with-mamr-breakthrough-for-nextgen-hdds/578055

[jabbr]

6 hours ago, Ralf11 said:

Moore’s "Law" is really just an Empirical Generalization

 

... just sayin'

 

...keepin' it real

Keeping it real, Bogotin makes the take home point, that because of Moore's Law, digital IC chips are designed to pack more and more gates on a smaller and smaller die which affects the rise time of the logic. Shortening the rise time causes ringing in receiver circuits, consequently a newer batch of a chip can cause signal integrity issues. In particular he notes that cheaper logic chips have shorter rise times ...

 

15 minutes ago, Ralf11 said:

why doesn't a buffer cure any of those problems?

 

It may but things may be more complicated behind the scenes than are realized. A circuit that is working barely ok because of ringing may get tipped over the edge because of ground bounce or power supply droop. The circuit as designed may not be exactly like the circuit as shipped.

 

But yes good buffers help a great deal. Cheaper devices tend to skimp on buffer sizes.

[marce]

This is why we use software such as in my case:

https://www.zuken.com/en/products/pcb-design/cadstar/products/pcb-analysis-and-verification

Cadence:

http://www.flowcad.ch/cms/upload/ds_pcb/Allegro_PCB_SI_Datasheet.pdf

etc.

So power requirements can be simulated at the frequencies of concern, so you get an impedance map of your power delivery system at a selection of frequencies. That is why with digital just playing with your main poser supply is pretty much a waste of time, you have to look at the whole system and understand how digital devices derive their power. So everything is important and correct decoupling is one of the most important factors. If you look at the current pulse of a digital wave, it is a sharp peak, corresponding to the rise time, with a sharp fall of when the required voltage has been reached. Its getting current to the device for that initial switching that is critical, it is that initial supply consisting of on die capacitance, board plane capacitance, local decoupling caps (in that order) that supplies the initial current, these are then topped up by local reservoir caps (larger 1uf-10uf). After that comes either larger reservoir caps or more often these days a local LDO or SMPS point of load supply. Finally the main supply that provides the main digital supplies, usually 3 or 5V depending on the logic, with separate 12V's for analogue. Though in the quest for reduced power, smaller sizes, some designs now work of low analogue supplies some times +/-5V.

So power requirements are worked out (should be, cheaper designs are often done by rule of thumb or an engineers beliefs. at all operating frequencies and a well engineered power deliver system will handle the load.

Fast rise times, signal integrity basic, add a series resistor next to the transmitting pin, and I do mean next to the pin, this line has to be as electrically short as possible, this will slow the rise time and limit the energy pit into the line. Or an a.c. parallel termination at the far end to absorb the energy and reduce the reflection.

The high rise time of some devices is a big problem for signal integrity, and often we have to work on basic interfaces with simulation because a logic device has a 1ns or less rise time, to drive a signal a few mm, requiring some form of termination.

[marce]

Reliability, all the companies I have done work for consider reliability a major concern. Look at the case of Amstrad computers and disk drive problems to see what lack of reliability does to a company... 

[jabbr]

Most audiophile companies come from a low speed analog expertise. Digital and high speed digital is relatively new. The digital interfaces eg USB typically come from small third party companies eg DIYINHK, Amanero etc. Do you have any idea what percentage of these small firms (producing $50 boards) use the design methodology you’ve outlined?

 

I think  if there is a widespread belief that “bits are bits” then folks won’t see the imperative of proper high speed design (audio is only 20-20k). Yes it’s possible to allow bits to be bits when properly designed (of course) but in consumer boutique audio how often is it done?

[mansr]

21 minutes ago, jabbr said:

Most audiophile companies come from a low speed analog expertise. Digital and high speed digital is relatively new. The digital interfaces eg USB typically come from small third party companies eg DIYINHK, Amanero etc. Do you have any idea what percentage of these small firms (producing $50 boards) use the design methodology you’ve outlined?

 

I think  if there is a widespread belief that “bits are bits” then folks won’t see the imperative of proper high speed design (audio is only 20-20k). Yes it’s possible to allow bits to be bits when properly designed (of course) but in consumer boutique audio how often is it done?

So let's say they've botched the USB input on the DAC. Why would you then trust these very same companies to produce a magical de-eviliser to fix it?

[jabbr]

12 minutes ago, mansr said:

So let's say they've botched the USB input on the DAC. Why would you then trust these very same companies to produce a magical de-eviliser to fix it?

 

Each company is different. I’m not trusting that everything has been done perfectly vs good enough. We aren’t given details nor shown SIV models. Heck it would be great if all cos published SPICE models but that’s not the case.

 

 

[pkane2001]

1 hour ago, jabbr said:

I think  if there is a widespread belief that “bits are bits” then folks won’t see the imperative of proper high speed design (audio is only 20-20k). Yes it’s possible to allow bits to be bits when properly designed (of course) but in consumer boutique audio how often is it done?

 

The problem as I see it is that companies (and audiophiles) are latching on to exactly the argument you're making that bits are not just bits in all parts of digital transmission. While this happens to be true, you and I know that there's little evidence that the effect of the RLC characteristics of a USB or ethernet cable have much to do with sound quality. And yet, that exact argument is reinforcing the belief of an average audiophile that they need to spend $500+ on an ethernet cable or that they absolutely must have that latest USB reclocker, along with an expensive USB cable, along with an expensive power supply, along with.....

 

Theory and speculation about audibility of noise sources in the digital domain is fine and healthy. What's not fine is the way audiophiledom is currently viewing USB audio: that it's a complete disaster, that it can only be fixed with 5 different chained cleanup/reclocking devices, expensive cables, and everything, including NAS and network switches, being powered by expensive LPS.

[jabbr]

Fair point. When I hear about some of these way over complicated setups my head spins about what could be going on. 

 

I do do believe that drawing attention to the importance of proper design of the digital inputs does place impetus on the manuf to do more work in this area (eg Schiit and Phasure and I’m sure others). 

 

On one hand we can talk about whether a DAC is ESS or AKM or AD or TI chip based but these other factors may be more important.

 

The DAC manuf should supply the cable and the AC cord.

[crenca]

2 hours ago, pkane2001 said:

 

Theory and speculation about audibility of noise sources in the digital domain is fine and healthy. What's not fine is the way audiophiledom is currently viewing USB audio: that it's a complete disaster, that it can only be fixed with 5 different chained cleanup/reclocking devices, expensive cables, and everything, including NAS and network switches, being powered by expensive LPS.

 

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...

[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.