Digital Transport Quality (and why it *may* matter).

[Davesrose]

My player converts to PCM and hands it to the receiver over HDMI. Too much of a pain to hand the bitstream across. The player is better at decoding all the various formats than the amp. I can't remember exactly what the problem was, but I tried to set it up bitstream at first and bailed because it wasn't as flexible that way.

There's probably less issues with an optical disc player too with lossless surround vs it just converting to PCM. Way back when, my first HD disc player was a HD-DVD player, and I think it did convert everything to multi-channel PCM. Since then, I've bitstreamed all my discs and other devices except the Apple TV (since it doesn't support it). One of the main issues I've been trying to point out with streaming or mobile devices is that there can be other audio settings you're not aware of (I mentioned a night mode...there's also "voice enhancement" DSPs with TV systems). I'm just bringing this up for those of us that are also watching content networked (or also perhaps certain portable devices).

 

[Davesrose]

CDs sound better if you put them in the freezer!
Even better if you give them the cryo treatment! https://forum.audiogon.com/discussions/cryo-treating-cds
Of course, you need to demagnetize them from time to time(very important!!!!!!!): https://6moons.com/audioreviews/furutech5/demag.html
And also you need this piece of tech: https://6moons.com/audioreviews/nanotech2/pro.html
Clearly, you failed to push your CD library to its utmost potential with your primitive improvements of a CD's periphery. You need to have a holistic approach in audio, consider the entire CD!

All hail @castleofargh you've shown me I was only tipping the iceberg when it comes to the marker trick or shaving the edges. I have lots of work to do with my library of CDs!

 

[Ghoostknight]

i really like the videos from PS-Audio and they also covered CD Transports... one thing they apparently do is using a buffer + galvanic isolation between the analog and digital section in their transports, which sounds somewhat logical beneficial
And i kinda trust these videos (and therefore this company, even if its way too expensive for me maybe overpriced but they seem to have a point atleast), i wasnt able to confirm all of them but most of the ones i was able to cover seem to be true

tho... im also wondering how much the cd transport effects playback "overall"... if we can connect our dac to a PC, we can also use the CD Drive of the PC without it being that much worse right?... it probably isnt isolated (but you can isolate the pc in different ways from the dac) but buffered by the OS

tho with PC`s i noticed they can sound quite a bit different... Windows sets a "kinda" good standard but their is a gap between windows and "better/tweaked" solutions (like with linux) tho this seems mostly software related
personally i rather tweak the PC connection instead of searching for a better transports (they seem kinda expensive..) because a PC is much more versatile (atleast in my setup), for living room audio i would probably go with a music server/streamer

In Fact... with PC you can also neglate all the CD stuff by using Digital Files thats also the reason why i never went for CD/Vinyl... a (good) PC setup just seems superior but it has its own little quirks like software

 

[Davesrose]

i really like the videos from PS-Audio and they also covered CD Transports... one thing they apparently do is using a buffer + galvanic isolation between the analog and digital section in their transports, which sounds somewhat logical beneficial
And i kinda trust these videos (and therefore this company, even if its way too expensive for me maybe overpriced but they seem to have a point atleast), i wasnt able to confirm all of them but most of the ones i was able to cover seem to be true

That's what Paul banks on. He gives enough of an argument that someone might believe without wanting to research further. And they're always for something when it happens to include their products. I've thought what demonstrates this the most is his video on audiophile network devices. Since they don't sell those products, he went ahead, and said they don't make a difference. Wonder if he'll come out with another video if they do decide to get into that game. Like seriously, I'm now floored by the amount of snake oil you can sell in this category ($1,000 network switches, $250 LAN silencers, $500 LAN cables, etc).

 

[Ghoostknight]

I've thought what demonstrates this the most is his video on audiophile network devices. Since they don't sell those products, he went ahead, and said they don't make a difference.

this also somewhat surprised me tho

I'm now floored by the amount of snake oil you can sell in this category ($1,000 network switches, $250 LAN silencers, $500 LAN cables, etc).

well true, im also not a huge fan of this
while i dont think they do nothing they are often 1. way overpriced anyway 2. while they might sell a better solution for a specific thing their might be way less expensive and better solution (easy non-snakeoil example... Analog EQ`s or DSP`s vs a PC)

 

[Davesrose]

well true, im also not a huge fan of this
while i dont think they do nothing they are often 1. way overpriced anyway 2. while they might sell a better solution for a specific thing their might be way less expensive and better solution (easy non-snakeoil example... Analog EQ`s or DSP`s vs a PC)

They really don't do anything. They claim they reduce noise, but they don't. The noise they claim that they reduce doesn't effect audio. These folks that claim they make a difference aren't in any kind of environment that has a huge amount of noise that begins with. If noise was so important, we couldn't have streaming music (because it's served from servers that's from thousands of server farms and miles of cable).

 

[gregorio]

Even though the premise is that PCs are noisy, and these devices magically eliminate noise (even though measured, they don't)....

That is the fundamental fallacy upon which so much audiophile snake oil is based. From various audiophile “de-crapifiers”, audiophile digital cables (USB, Ethernet, optical, etc.), DDCs, transports, audiophile network switches, etc. It’s a fallacy based on falsely applying analogue signal principles to digital signals, it’s a fundamental misunderstanding/misrepresentation of how and why digital audio was invented and of what it is!

Noise is the whole reason why digital audio/communications was developed. Forget poxy little rip-off audiophile companies, forget even the pain of telecoms giants getting a never ending stream of complaints and demands for refunds because noise/interference made many telephone calls indecipherable. No, the big impetus was because noise caused the deaths of tens of thousands of military personnel and tens of thousands more of civilian deaths during WWII. Noise caused errors in military communications; squadrons of bombers obliterated residential areas instead of military targets, other squadrons got decimated by overflying areas of high anti-aircraft defences because noise stopped them hearing the instruction to “avoid”, similar things happened with warships, submarines and battalions of artillery and soldiers, and they kept happening! Huge resources were put into combatting noise and numerous experimental systems were developed and trialled, including operational PCM systems but by the time the war ended, engineers were still largely working in the dark. The work continued but more in the civilian world of the telecom giants.

We often quote Shannon in this subforum but tend to focus on just one part of his theory, the part pertaining to the theorem of capturing all the information in a time varying function (EG. An analogue waveform) but in fact that’s a relatively small part of his famous “A Mathematical Theory of Communication” 1948 paper and we don’t tend to mention the huge achievement in the rest of it. The paper is effectively all about noise and the application of advanced stochastic math to combat it. Firstly by introducing the concept of “information entropy”, dealing with the issue as a “probability” issue by encoding the information in binary “bits” and formulating theorems to combat ANY effect of noise on those “bits”. The result was that engineers were no longer working in the dark, they had proven mathematical equations to encode the information and to calculate the channel bandwidth and data rates required to transmit and receive that information without any noise/interference induced error. Again, this is information in general, it doesn’t only apply to digital audio. Your phone can transfer trillions of bits of data a second from long term storage to RAM, to it’s processing cores and back again without crashing every second because Shannon gave engineers a mathematical blueprint to achieve the task without noise/interference causing any change or loss of information. And of course this applies to all digital devices and is why Shannon is called “the father of the digital age”. Incidentally, Kotelnikov, a Soviet scientist, arrived at the sampling theorem several years before Shannon and achieved much of what Shannon’s paper achieved in his doctoral thesis “The Theory of Optimum Noise Immunity”, a year before Shannon’s paper was published but this was unknown in the west until many years later.

So the audiophile claims are just nonsense, there is noise/interference when digital information is transferred but digital is immune to it (so there can be NO benefit from trying to remove it). That’s why digital was invented, it’s what digital is, and if that were not the case then the digital age would not exist.

Sorry for the long post but hopefully some will find it interesting or illuminating.

G

 

[Elegiac]

CDs sound better if you put them in the freezer!
Even better if you give them the cryo treatment! https://forum.audiogon.com/discussions/cryo-treating-cds
Of course, you need to demagnetize them from time to time(very important!!!!!!!): https://6moons.com/audioreviews/furutech5/demag.html
And also you need this piece of tech: https://6moons.com/audioreviews/nanotech2/pro.html
Clearly, you failed to push your CD library to its utmost potential with your primitive improvements of a CD's periphery. You need to have a holistic approach in audio, consider the entire CD!

Oh man. Lovely. This'll be a good read.

 

[gregorio]

Wasn't there something about Apple integrating a Fletcher-Munson compensation curve into their DAC software?

Not into their DAC. Loudness Normalisation (the “Sound Check” feature of IOS/MacOS) partially uses a Fletcher-Munson loudness contour, so it’s maybe referring to that but Apple are very secretive about what they’re doing so it’s near impossible to know for sure. Although it’s proven that their DACs are ruler flat. Incidentally, guess where the Fletcher-Munson curves were first applied in practice.

i really like the videos from PS-Audio and they also covered CD Transports... one thing they apparently do is using a buffer + galvanic isolation between the analog and digital section in their transports …

A real classic bit of audiophile marketing there. A tried, tested and fruitful marketing ploy that’s over 40 years old, the old “lie of omission”. Let me add the missing part to this: “one thing they apparently do is using a buffer + galvanic isolation between the analogue and digital section in their transports” - The missing part at the end is: Yes, just like every other CD player, even cheap $40 OEM CD drives from the 1990’s! To make matters worse, this really isn’t anything to do with “digital transport” anyway, it’s an analogue issue.

while they might sell a better solution for a specific thing their might be way less expensive and better solution

Yes there is a much better and much cheaper solution. Doing nothing at all costs $0 and is a better solution because as already explained in my previous post, digital is already immune to noise.

G

 

[Davesrose]

Noise is the whole reason why digital audio/communications was developed. Forget poxy little rip-off audiophile companies, forget even the pain of telecoms giants getting a never ending stream of complaints and demands for refunds because noise/interference made many telephone calls indecipherable. No, the big impetus was because noise caused the deaths of tens of thousands of military personnel and tens of thousands more of civilian deaths during WWII. Noise caused errors in military communications; squadrons of bombers obliterated residential areas instead of military targets, other squadrons got decimated by overflying areas of high anti-aircraft defences because noise stopped them hearing the instruction to “avoid”, similar things happened with warships, submarines and battalions of artillery and soldiers, and they kept happening! Huge resources were put into combatting noise and numerous experimental systems were developed and trialled, including operational PCM systems but by the time the war ended, engineers were still largely working in the dark. The work continued but more in the civilian world of the telecom giants.

Not to disagree with the fundamental that digital systems greatly mitigate noise (as the SNR only needs to represent an on or off during a set time), but as nuclear technology was advanced for military use, so was technology for digital computers. The first digital computer, ENIAC, was created to calculate ballistics, and was funded by the US Army (it wasn't created to save lives).

Digital systems are normally irrelevant with noise...as the bandwidth needed is ones and zeros with a given speed. If there is an issue with the cable (such as poor interface termination, or not within spec), then there can be a drop in bandwidth and the effects would be noticeable with audio/visual. It would be evident as blocking or hangups with video, or popping or static in audio. This gets to be further mitigated with LAN cables, in which TCP/IP sends data in packets. That means that if there is an error, there's less of a chance to be picked up as there's an inherent buffer. It's also a fundamental reason why a $20 LAN switch measures the same as a $1000 LAN switch, or a $10 cat6 cable measures the same as a $500 one. And also more so that specifications for LAN now have data bandwidths that well exceed data bandwidth that audio formats needs.

 

[bigshot]

There's really no reason to split the transport and DAC into two separate components if you want to play a CD. Just get a CD player. They're inexpensive and they do just as good of a job of playing a CD as doing it the hard way.

 

[gregorio]

The first digital computer, ENIAC, was created to calculate ballistics, and was funded by the US Army (it wasn't created to save lives).

Actually, as far as I’m aware, Colossus was the first digital computer built for code breaking in England but it was dismantled after WWII and was classified until the 1970’s. Shannon didn’t invent digital, he invented the sampling theorem and how to communicate digital information without error/loss. Shannon’s paper was the first to mention digital “bits” and indeed the ISO proposed that the official international name for “bits” be the “Shannon” but he refused the honour stating that he didn’t invent the concept or the term “bits”. The earliest digital computers were effectively relatively simple programmable calculators.

Digital systems are normally irrelevant with noise...as the bandwidth needed is ones and zeros with a given speed.

No, or rather not exactly. Digital systems are somewhat different to the early digital “calculator” type devices, they needed to deal with data that represents information. So rather than just entering a few relatively simple numbers and spitting out a relatively simple (though difficult to compute) number, a great deal more data needs to be input and output, and at these higher data rates, noise/interference becomes an issue and error rates increase proportionally. Therefore, the “bandwidth needed” is the ones and zeros representing the information itself plus the ones and zeros necessary for redundancy, to ensure communication unaffected by noise/interference.

This is what Shannon (and Kotelnikov) invented, it’s effectively a two pronged attack on noise/interference. Firstly you encode information as binary data (for example using the sampling theorem) which is inherently immune to low and moderate levels of noise/interference then you incorporate data to check for (and correct) errors in the case of high levels of noise/interference. That’s why Kotelnikov called his dissertation “The Theory of Optimum Noise Immunity”. Shannon’s paper provides the math for calculating (and mitigating) error rates down to any arbitrary level. Without this, “the digital age” could not exist, it would be limited to nothing more than just relatively simple digital calculators.

G

 

[Davesrose]

Actually, as far as I’m aware, Colossus was the first digital computer built for code breaking in England but it was dismantled after WWII and was classified until the 1970’s. Shannon didn’t invent digital, he invented the sampling theorem and how to communicate digital information without error/loss. Shannon’s paper was the first to mention digital “bits” and indeed the ISO proposed that the official international name for “bits” be the “Shannon” but he refused the honour stating that he didn’t invent the concept or the term “bits”. The earliest digital computers were effectively relatively simple programmable calculators.

ENIAC is considered the first general purpose programmable electronic digital computer.

No, or rather not exactly. Digital systems are somewhat different to the early digital “calculator” type devices, they needed to deal with data that represents information. So rather than just entering a few relatively simple numbers and spitting out a relatively simple (though difficult to compute) number, a great deal more data needs to be input and output, and at these higher data rates, noise/interference becomes an issue and error rates increase proportionally. Therefore, the “bandwidth needed” is the ones and zeros representing the information itself plus the ones and zeros necessary for redundancy, to ensure communication unaffected by noise/interference.

No, what I out lined is the simplified version of what a digital signal is. It's a binary set of numbers being sent at a given speed. You're getting into the digital system's error correction algorithm for checking data loss. The main complication about a data cable is whether it is serial (one bit at a time), or parallel (having multiple lanes of data being transmitted).

 

[gregorio]

ENIAC is considered the first general purpose programmable electronic digital computer.

Looking it up, it seems there’s some discussion on this subject. Colossus was completed and in use 2 years before ENAIC and was the first programmable electronic digital computer, although it was not “general purpose” but then arguably neither was ENAIC. The main difference seems to be that ENAIC had a big press launch in 1946 and computers entered the public consciousness while the 10 already working Colossus computers were being smashed-up and all records burnt by order of the British government.

You're getting into the digital system's error correction algorithm for checking data loss.

Error correction and/or detection is integral in digital data communication, as Shannon proved. Along with binary encoding it rendered digital immune to noise/interference, thereby invalidating the main/typical “selling point” of audiophile digital cables, switches, etc.

G

 

[Davesrose]

Looking it up, it seems there’s some discussion on this subject. Colossus was completed and in use 2 years before ENAIC and was the first programmable electronic digital computer, although it was not “general purpose” but then arguably neither was ENAIC. The main difference seems to be that ENAIC had a big press launch in 1946 and computers entered the public consciousness while the 10 already working Colossus computers were being smashed-up and all records burnt by order of the British government.

ENAIC is considered "general-purpose" because of its "general-purpose" programming. Meaning program logic for different applications, execute, and debug. Colossus was a precursor (and wasn't widely known about until the 70s). There were other various precursors of being electronic, digital, and programmable. That you did bring it up, I do see your point about digital computers saving lives in the war. My view was that it's ironic that the investment in advanced technology is often carried out by the military for weapons (but then the military also saves lives: not just code breaking, but medical advancements).

Error correction and/or detection is integral in digital data communication, as Shannon proved. Along with binary encoding it rendered digital immune to noise/interference, thereby invalidating the main/typical “selling point” of audiophile digital cables, switches, etc.

It's another topic that's after the concept of digital transmission (which again, is either serial or parallel). I'm not disputing that error correction is integral in a digital system and can keep an audio file "bit perfect". But it's still separate concept than transmission, which doesn't necessarily have error correction (for example, certain components on a data bus). The need, type, and amount of error correction can also be different depending on application. Scratches on a CD being handled in the player, digital transmission of a digital audio cable being handled by receiving component (that has data sent in one stream), or digital transmission of a network cable (that has data sent in packets). In a serial connection, bandwidth is determined by the speed in which bits are sent (toslink capping at 15 Mbps, HDMI 2.1 capping at 48 Gbps, cat6 capping at 10 Gbps up to 180ft). Now that network transmissions are in the gigabits, and are sent in packets, the need for error correction for audio is greatly diminished. With streaming video over internet, I've noticed the program might start with a fuzzy or pixelated picture until there's enough of a buffer for the full HD or UHD image to get to optimal quality. The DD+ audio is clear immediately.

To get back to audiophile network devices, the logic fallacies audiophiles are prone to really amuses me. I've heard them say they don't use a computer, because it's so noisy. And also you need an expensive music streamer to go along with your expensive audio setup. If the selling point is that the high end devices some how eliminates noise, then why spend $500 on a network cable? And then if you've spent that kind of money, and noise is supposedly not present, why do you need a $1000 audiophile network switch and 2 $100 LAN silencers on either end? More amusing, when I did engage with an audiophile about a controlled test, he was against the idea of using a PC as a source. But isn't the claim that it's too noisy and transmits noise? Wouldn't a LAN silencer be more pronounced in such a situation?