Bits is bits?

[Soothsayerman]

On 8/6/2019 at 12:57 PM, Indydan said:

I found this article through Darko. The author makes a case that it isn't a case of Bits are bits. I wonder what others think?

 

 

https://www.upscaleaudio.com/pages/bits-is-bits?fbclid=IwAR20r8A_bUYGlDjzsbFfEKzNAB06vDsYOgTf_TUf7xXwpps6-kdggIay4bU

 

 

The gist of this article is absolutely correct and I know this from working in the telecommunication and enterprise computing industries for decades.  What is baffling to me is why it seems to be such an issue in the audio world. 

 

Part of it may stem from poor power supply design and implementation within audio equipment and the fact that manufacturers skimp on parts and power supplies and end up creating a very electrically noisy environment or an environment that does not manage heat well.   The other is that the application itself does not lend itself well to tolerating data buffers to deal with  error correction and then there is the whole jitter issue which everyone deals with in a variety of ways.  Clocking and frequency stability are issues.

 

In enterprise computing, power regulation and power cleanliness is a very important factor because it has a large impact in computing errors, downtime, equipment longevity etc and of course heat is an issue on the IC level and the environmental level.  Cables and connections do go bad and you would not really think that would be the case because they are seemingly inert but they're not.  Not all transmission mediums are the same (multimode vs singlemode fiber even though they're both glass) but once you get to a certain level, if everything else is accounted for, they should have no impact. 

 

It's a mystery wrapped in an enigma lying at the bottom of a beer glass that I'm sure has something to do with money.

 

 

[davide256]

Mostly agree with the article... lesser digital  cable design affect on sound is  very much like being the "3rd listener in a whisper chain". But find it hard to justify cable prices over $500... there just isn't that much cost in the materials/design. Not buying any Ethernet cable discussion as thats pure digital domain

[mansr]

38 minutes ago, marce said:

Nope its just a quick simulation with two LS00 gates and a few mm of 50R track between them, it was more to show the current pulses as we look at square waves as voltages but rarely as current pulses... The data is derived from the IBIS models of the driving circuitry, this sim is as basic as it gets two gates and a wire.... 

Well, a resistive load wouldn't result in i = dv/dt as seen here. It has to be something more complex. Yes, real circuits are rarely resistive, so this is of course what actually happens. Not disagreeing there.

[marce]

My main aim was to show digital (square wave signals) are pulses, current pulses, which is why the simulation software uses S parameters among other stuff. (to get the simulation to match the scope shots, you have to model the scope probe loading on the signal, most scope manufacturers have modelling info for their probes, its quite cool when the sims match the scope screen, don't get to play with it much these days,but if ever get time I've been meaning to download some USB driver IBIS data and model some USB cables, did some SPDIF years ago on DIYAudio several years ago to show just adding a 75R resistor cleans up the wave nicely.)

My own view on the article is a set of relatively true statements mixed together in a bit of a mishmash to promote the belief in cables...

Quote

digital processors must recognize and decode many different voltages (16 or more for Ethernet)

This is my favourite bit.... ALL baseband Ethernet I have worked on has been digital, either on or off, not 16 different voltages... Maybe they are mixed up with broadband, which is frequency-division multiplexing, so I am unsure where the 16 different voltages comes from.

[esldude]

1 minute ago, marce said:

My main aim was to show digital (square wave signals) are pulses, current pulses, which is why the simulation software uses S parameters among other stuff. (to get the simulation to match the scope shots, you have to model the scope probe loading on the signal, most scope manufacturers have modelling info for their probes, its quite cool when the sims match the scope screen, don't get to play with it much these days,but if ever get time I've been meaning to download some USB driver IBIS data and model some USB cables, did some SPDIF years ago on DIYAudio several years ago to show just adding a 75R resistor cleans up the wave nicely.)

My own view on the article is a set of relatively true statements mixed together in a bit of a mishmash to promote the belief in cables...

This is my favourite bit.... ALL baseband Ethernet I have worked on has been digital, either on or off, not 16 different voltages... Maybe they are mixed up with broadband, which is frequency-division multiplexing, so I am unsure where the 16 different voltages comes from.

Probably from the uninformed belief that 16 different voltages are transmitting 16 different bits on ethernet. 

[marce]

Probably, it is disinformation though and confuses the issue...

 

[Soothsayerman]

22 minutes ago, marce said:

My main aim was to show digital (square wave signals) are pulses, current pulses, which is why the simulation software uses S parameters among other stuff. (to get the simulation to match the scope shots, you have to model the scope probe loading on the signal, most scope manufacturers have modelling info for their probes, its quite cool when the sims match the scope screen, don't get to play with it much these days,but if ever get time I've been meaning to download some USB driver IBIS data and model some USB cables, did some SPDIF years ago on DIYAudio several years ago to show just adding a 75R resistor cleans up the wave nicely.)

My own view on the article is a set of relatively true statements mixed together in a bit of a mishmash to promote the belief in cables...

This is my favourite bit.... ALL baseband Ethernet I have worked on has been digital, either on or off, not 16 different voltages... Maybe they are mixed up with broadband, which is frequency-division multiplexing, so I am unsure where the 16 different voltages comes from.

 

Ethernet uses different voltages depending on whether it uses POE or Gigabit Ethernet and what the muxing requirements are.   So, a signal over a Gigabit Ethernet cable represents two bits, instead of a single bit. In other words, instead of just using two voltages on a signal representing merely “0” or “1”, it uses four different voltages, representing “00”, “01”, “10” and “11”.  Factor in other engineering considerations and it can add up.  The data requirements and architecture for real time computing (streaming) are different than latent computing. But I'm sure everyone already knows this.  The data bus in streaming applications and therefore the transmission requirements are different than say a desktop environment.

 

[mansr]

12 minutes ago, marce said:

This is my favourite bit.... ALL baseband Ethernet I have worked on has been digital, either on or off, not 16 different voltages... Maybe they are mixed up with broadband, which is frequency-division multiplexing, so I am unsure where the 16 different voltages comes from.

From IEEE 802.3:

Quote

40.1.3 Operation of 1000BASE-T

 

The 1000BASE-T PHY employs full duplex baseband transmission over four pairs of Category 5 balanced cabling. The aggregate data rate of 1000 Mb/s is achieved by transmission at a data rate of 250 Mb/s over each wire pair, as shown in Figure 40–2. The use of hybrids and cancellers enables full duplex transmission by allowing symbols to be transmitted and received on the same wire pairs at the same time. Baseband signaling with a modulation rate of 125 MBd is used on each of the wire pairs. The transmitted symbols are selected from a four-dimensional 5-level symbol constellation. Each four-dimensional symbol can be viewed as a 4-tuple (An, Bn, Cn, Dn) of one-dimensional quinary symbols taken from the set {2, 1, 0, –1, –2}. 1000BASE-T uses a continuous signaling system; in the absence of data, Idle symbols are transmitted. Idle mode is a subset of code-groups in that each symbol is restricted to the set {2, 0, –2} to improve synchronization. Five-level Pulse Amplitude Modulation (PAM5) is employed for transmission over each wire pair. The modulation rate of 125 MBd matches the GMII clock rate of 125 MHz and results in a symbol period of 8 ns.

Not just on/off, but far from 16 levels.

[marce]

4D-PAM5, so 5 total, not basic digital though.

[marce]

Interesting stuff... Thanks for pointing out a missing segment in my knowledge. Read some easy stuff, could do with more, this is quite good.

https://www.dell.com/content/topics/global.aspx/power/en/ps4q01_patward%3Fc%3Du

 

[Samuel T Cogley]

2 hours ago, Soothsayerman said:

 

 

The gist of this article is absolutely correct and I know this from working in the telecommunication and enterprise computing industries for decades.  What is baffling to me is why it seems to be such an issue in the audio world. 

 

Part of it may stem from poor power supply design and implementation within audio equipment and the fact that manufacturers skimp on parts and power supplies and end up creating a very electrically noisy environment or an environment that does not manage heat well.   The other is that the application itself does not lend itself well to tolerating data buffers to deal with  error correction and then there is the whole jitter issue which everyone deals with in a variety of ways.  Clocking and frequency stability are issues.

 

In enterprise computing, power regulation and power cleanliness is a very important factor because it has a large impact in computing errors, downtime, equipment longevity etc and of course heat is an issue on the IC level and the environmental level.  Cables and connections do go bad and you would not really think that would be the case because they are seemingly inert but they're not.  Not all transmission mediums are the same (multimode vs singlemode fiber even though they're both glass) but once you get to a certain level, if everything else is accounted for, they should have no impact. 

 

It's a mystery wrapped in an enigma lying at the bottom of a beer glass that I'm sure has something to do with money.

 

 

I've been in IT and "enterprise computing" since the 80s and I'm really struggling to think of one example of $1000+ Ethernet cables that are used there.  In fact, I can't think of a single boutique "upgrade" in the realm of telecommunication or "enterprise computing".

 

Cables "go bad" all by themselves in your environment?  I'd check for a saboteur if I were you.

[Ralf11]

no no - audio cables "go bad" 'cuz sensitive ears audio is special snake oil

[Soothsayerman]

26 minutes ago, Samuel T Cogley said:

 

 

I've been in IT and "enterprise computing" since the 80s and I'm really struggling to think of one example of $1000+ Ethernet cables that are used there.  In fact, I can't think of a single boutique "upgrade" in the realm of telecommunication or "enterprise computing".

 

Cables "go bad" all by themselves in your environment?  I'd check for a saboteur if I were you.

It's not a common thing, but I have been in many many data centers where a cable has failed. I would really suspect a manufacturing defect or, someone did not handle it properly or a datacenter where there are lots of adds moves changes where it gets abused.   I never said anything about boutique cables.  A data center is a perfect environment, your living room is not.

[crenca]

1 minute ago, Soothsayerman said:

It's not a common thing, but I have been in many many data centers where a cable has failed. I would really suspect a manufacturing defect or, someone did not handle it properly or a datacenter where there are lots of adds moves changes where it gets abuses....

 

I worked net admin in a factory where we had long ethernet cable runs through plenum above the factory floor.  These cables would occasionally fail (with zero moves on either end) and my best guess was winter cold/summer heat expansion of the roof/plenum space...

[mansr]

2 minutes ago, crenca said:

I worked net admin in a factory where we had long ethernet cable runs through plenum above the factory floor.  These cables would occasionally fail (with zero moves on either end) and my best guess was winter cold/summer heat expansion of the roof/plenum space...

That's why you run them with a little slack.

[crenca]

31 minutes ago, mansr said:

That's why you run them with a little slack.

 

Even then, failures.  I speculated about them being pinched, weak points in the wires separating enough, etc.

 

 

[Samuel T Cogley]

1 hour ago, Soothsayerman said:

It's not a common thing, but I have been in many many data centers where a cable has failed. I would really suspect a manufacturing defect or, someone did not handle it properly or a datacenter where there are lots of adds moves changes where it gets abused.   I never said anything about boutique cables.  A data center is a perfect environment, your living room is not.

 

Ok, but you certainly implied that your experience in "many many" data centers validates the "need" for boutique cables in audio.  That's what we're talking about, by the way:  Audioquest marketing literature posted by the OP.

[Soothsayerman]

32 minutes ago, Samuel T Cogley said:

 

Ok, but you certainly implied that your experience in "many many" data centers validates the "need" for boutique cables in audio.  That's what we're talking about, by the way:  Audioquest marketing literature posted by the OP.

 

Do you always have an arrogant dickhead attitude or is it only when you're talking about audio?

Cables can make a difference depending on the system and environment they're used in. My belief is that the point of rapidly diminishing returns happens a heck of a lot faster than most cable manufacturers would want you to believe.  But the whole notion that a bit is a bit in the grand scheme of essentially a very complex data bus is not true when it comes to streaming.   Now if you think that is snake oil, I personally do not care. 

 

[crenca]

"Do you always have an arrogant dickhead attitude or is it only when you're talking about audio? "

 

Easy there Biggin. Samuel has been around for a while and does not possess a bad attitude.  Assume the limitations of a comment box first....

 

 

[Samuel T Cogley]

2 minutes ago, Soothsayerman said:

 

Do you always have an arrogant dickhead attitude or is it only when you're talking about audio?

Cables can make a difference depending on the system and environment they're used in. My belief is that the point of rapidly diminishing returns happens a heck of a lot faster than most cable manufacturers would want you to believe.  But the whole notion that a bit is a bit in the grand scheme of essentially a very complex data bus is not true when it comes to streaming.   Now if you think that is snake oil, I personally do not care. 

 

People that try to treat the entire audio experience based upon solely on empirical evidence are a bit laughable to me because hearing does not work that way and I think "hearing" is the whole point of the hobby. At least to me it is.

 

Your knee-jerk ad hominem isn't helping your cause.  You started with an appeal to authority, hence:

 

Quote

The gist of this article is absolutely correct and I know this from working in the telecommunication and enterprise computing industries for decades.

 

So Audioquest marketing material is "absolutely correct"?  I hope you'll forgive my skepticism.  And I'm just not understanding how your background validates, in any way, said marketing material.

 

[Ralf11]

I want an an arrogant dickhead USB cable - I can use it for punk rock

[Samuel T Cogley]

11 minutes ago, Ralf11 said:

I want an an arrogant dickhead USB cable - I can use it for punk rock

 

Punks are more generically obnoxious.  I don't think arrogance is really a factor. 🙂

[Soothsayerman]

10 minutes ago, Samuel T Cogley said:

 

Your knee-jerk ad hominem isn't helping your cause.  You started with an appeal to authority, hence:

 

 

So Audioquest marketing material is "absolutely correct"?  I hope you'll forgive my skepticism.  And I'm just not understanding how your background validates, in any way, said marketing material.

 

I didn't call you an arrogant dickhead I asked about your attitude. It was actually an interrogative.  Like I said. I believe the point of rapidly diminishing returns occurs quickly in cables and that a bit isn't always a bit.   If that has too much gray area for you, there are a lot of things in life that are likely going to disappoint you.

[jabbr]

31 minutes ago, Soothsayerman said:

But the whole notion that a bit is a bit in the grand scheme of essentially a very complex data bus is not true when it comes to streaming.   Now if you think that is snake oil, I personally do not care. 

How exactly is an Ethernet frame different in streaming than any other TCP/IP application? Exactly how could any teenie little audiophile company like audioquest have the resources to even come close to the Ethernet cable technology of Belden — including the REVConnect? 1Gbe is child’s play for the big boys, even 10Gbe is old school — now check out that encoding and those levels and jitter / eye patterns that need to be hit there. Just tell me how you are going to beat Belden ... no show me.

[Samuel T Cogley]

4 minutes ago, Soothsayerman said:

I didn't call you an arrogant dickhead I asked about your attitude. It was actually an interrogative.  Like I said. I believe the point of rapidly diminishing returns occurs quickly in cables and that a bit isn't always a bit.   If that has too much gray area for you, there are a lot of things in life that are likely going to disappoint you.

 

I see you're dodging the question, but I'll try one last time: How does your career background validate that Audioquest marketing material?  I honestly want to know.