> It's an HDMI cable! The video signal has CRC in it and is packetized, it's either going to make it or it isn't.

I don't disagree with your main point, but this actually isn't quite true. The HDMI signal is split into 3 distinct interleaved periods: video data, data island and control. Video data is not packetized and the only possible error detection it has is from TMDS signaling, but no such error handling is required by the TMDS spec. You can absolutely get imperfect transmission of an HDMI video signal due to cable or other electrical problems. Auxiliary packets in the data island, including audio data, do have an error correction scheme (BCH + TERC4).

Feel free to check out the spec: https://glenwing.github.io/docs/HDMI-1.4b.pdf

Yes, but when when there's signal degradation in an HDMI cable it shows obviously, not as a subtle video quality change. If you can see an HDMI signal with white blinking dots everywhere, you can be confident that the cable is not altering the signal in any really detectable way.

"If you can see an HDMI signal with white blinking dots everywhere, you can be confident that the cable is not altering the signal in any really detectable way."

I have seen an HDMI transmission with white blinking pixels at content edges that were not from the original signal. I ran test signal through that source and verified the errors on an HDMI waveform monitor, after which I tested the cable and threw it away.

HDMI video is not magically impervious to error because it is digital. It's actually a pretty straightforward transmission scheme.

This was while working on an HDMI-output device, working with team members who had written the original HDMI specifications (so, yes, I've read them).

It is possible to have subtle degradation in HDMI-delivered content. It's rare, but possible. As one might expect, comprehensive failures (e.g. black screen, periodic HDCP key mismatch) are far more common, as are negotiation/configuration failures (e.g. wrong resolution, wrong framerate, wrong color space).

HDMI cables used to be a great example, but now there are something like 8 different levels of certification and only the highest have the bandwidth to carry a 2160p, 60fps, 10 bit, 4:4:4 signal. And as far as I can tell, they are all passive so your source or sink can't warn you that your cable isn't compatible. You'll just get artifacts.

The AmazonBasics cable will do 18Gbps, so it can do 2160p, 60fps, 10 bit, but only with chroma subsampling (which is fine because most video sources have chroma subsampling). The Amazon cable may work with a 4:4:4 signal, but it's not certified for it. There's a level above that for 48Gbps cables.

I totally agree with the general idea, though. As long as a cable for a digital signal meets the specifications set by the standards body, you're good.

Okay, so the last time I bought hdmi cables I got 2.0 rated. Just looked and HDMI 2.1 rated 48Gbps cables are $25-35 each for 6 foot. That's not totally unreasonable. Now how is a $350 cable or a $800 cable any better?

One thing I can say about copper is that it is noisy.

We are running FPGAs (with NICs) and were getting lots of packet errors over the copper cable SFP cables. Turns out, we either needed to write out noise cleanup logic into the FPGA, or add in a noise cleanup module which would add an extra 50ns to packet processing to the FPGA. Simple solution was to convert to fiber and the noise and errors all went away.

I'd say if audiophiles truly want unmolested electricity, they ought to use optical fiber for as much of the run as possible.

Digital transmission/storage ensures the signal is not affected no matter how much noise is on the line (well, as long as it can be transmitted at all...)

Now, what I think you mean is that the copper in the cable picks up noise either from the device on the other end or from the surroundings and that has the possibility of affecting the DAC.

Even moderately well designed receiver should be able to galvanically isolate itself from outside sources of noise without having to resort to fiberoptics or wireless. It's just couple of dollars max for extra few components on the board.

For an example possibly familiar in computing circles, Infiniband has easily-accessible counters to record transmission errors. In a properly-built system with hundreds of communication-busy nodes connected with copper, you see remarkably low error rates, even in comparison with the spec (maybe a few a week in my experience).

Infiniband is using extremely isolated and well balanced cables for passive/copper transmission. Also the cable lengths are very short for EDR and beyond (3m-5m is max usable length). For anything longer you need to use fiber cables with embedded converters.

We have a cluster with ~1000 nodes and all generations of Infiniband (DDR/QDR/EDR) equipment, and routing all these copper cables without damaging them is not easy.

The quality of digital tranmsission over copper cables is dependent on the size and clarity of "eye", which is the pattern which two sinusodial signals when run through and oscilloscope. Lower quality cables (HDMI/DP/Coax/etc.) has blurry and small eyes, which increases chances of artifacts and drives the error correction modules harder, while better cables are easier on the both sender and receiver, since they produce lower number of errors to correct.

Trivia: Amphenol / Gore has high speed digital interconnect cables called Eye Opener Plus [0], which also used some of the Infiniband cables that we use.

[0]: http://www.spectra-strip.com/Eye_Opener_Plus_Cable.cfm?Nav=E...

> Digital transmission/storage ensures the signal is not affected no matter how much noise is on the line (well, as long as it can be transmitted at all...)

There is no such thing as a digital transmission. All signals are analog.

Digital transmission refers to the whole transceiver system, not just the transmission medium. If I can put a "1" in and reliably get a "1" out, it doesn't matter that it was only a 0.9 part way down the wire.

But that's the question, isn't it? If 0.0001% of the time you get 0.45v instead of 0.9v on the far end, now what?

I suppose you need to read up on digital protocols, it is all well handled. If it wasn't there could be no Internet or digital media or even processors or motherboards.

That is a higher level, though.

I'm just so tired of all the bullshit where it "Must be a 1 or a 0". That is patently not true. HDMI video, to name one example very relevant to this thread, has no error correction, and bit errors will indeed crop up as the "sparkling" artifacts GGGGGP reported.

Sorry, copper is not ‘noisy’! Any medium has limits and methods to properly use it. E.g. shielding / grounding. Have to look at whole system. Agree fibre might be more practical in some situations.

Obviously don’t spend $1000 on a normal passive HDMI cable, but I’ve found that increasingly the all-or-nothing argument doesn’t work for some reason. I recently had to upgrade all my HDMI cables to get 4k HDR video working reliably. Before upgrading, the video would work, but I would occasionally see white “sparkles” on the video output. Googling this problem led me to many people recommending upgrading HDMI cables if you have very old cables.

Certainly there are bad cables, and sparkles are clear signal dropouts.

The description of that cable though, says Perfect-Surface Technology applied to extreme-purity silver provides unprecedented clarity and dynamic contrast.

An HDMI cable can not affect the "clarity" or contrast of the image or audio it carries. It can just succeed or fail (and failure results in obvious pixel noise, not some overall lowering of contrast or clarity like in analog)

All cables have some probabilistic failure of corruption, it isn't binary. It's the same as what going from 5 9s in your application to 7 9s, which involves a huge cost with distributed databases multiple data centers etc. etc... For most people it's a waste, but if you really value not having the occasional white speck, it could be worth a better cable, which may cost a lot more to recover design costs of making a cable very few will want to buy.

I don't know this world. But could it be because the HDMI protocol doesn't require correcting all errors? And newer cables produced fewer corrupt packets which might have resulted in "sparkle"?

Different HDMI cables have different bandwidth which is a result of their physical construction, same way as you have different bandwidth ethernet cables. Sending signal on the edge of cable's capability may be causing intermittent problems.

I've always wondered whether I could sell audiophile SATA cables for those who listen to music from their hard drives...

Edit: a commenter below has a link to an actual audiophile SATA cable.

Yes, you can absolutely sell "HD SATA" cables. To paraphrase, "There's a new audiophile born every minute".

https://jcat.eu/audiophile-sata-cable/

I once saw a shop that sold gold plated fiber optic cables...

Many years ago I saw gold plated TOSLINK cables. They were $500. TOSLINK is a pure fiber audio connection standard that uses multimode fat plastic fiber.

This made me snigger perhaps more than I should do. I wonder if the people selling this shit actually know or not. They have to I suppose.

huh, besides the BS, isn't the file loaded into RAM first anyway?

You know what's next - sell audiophile-grade RAM. Gold-plate everything.

full stack audio grade computer, 40000USD

Aim higher. That's cheap compared to the $90,000 DAC you need to buy https://darko.audio/2015/11/the-select-dac-ii-msb-technology...

It's called ECC... And I happily pay for it.

Yes, which is why this is one of the more absurd pieces of audiophile gear.

Anything "audiophile" in the digital domain is all equally absurd. The only difference is that some require a half-second of consideration and others require literally no thought at all to identify as stupid.

I wouldn't go as far to say "anything", but certainly most. As long as a DAC with sufficient resolution and stable clock is being fed with enough bits, then yes.

I would go that far.

If it’s all-digital and marketed to audiophiles, it’s a scam and/or absurd. You can’t improve upon bit-perfect. (Room calibration being a notable exception.)

You mean like those people, who argue, that: "Everything is a sequence of Zeroes and Ones, therefore the signal either gets transmitted or not!", while totally leaving out the fact, that this digital signal materializes in the real world via an analogous signal, which is electricity, and therefore each digital data transmission (as long as not optical) is submitted to physics that happen totally outside of the "digital domain"?

I don't understand the point you're trying to make. Everything is ultimately subject to physics, including optical signals. Nobody claims otherwise.

But it is incorrect to say that digital transmissions must be at the mercy of physics. Digital transmissions can defy physics. There are numerous techniques which ensure that a signal reaches the destination with precisely zero flipped bits and exact timing. And even when the signal has no integrity mechanisms, in practice the error rate will be low enough to never matter in practice.

And you know what, even if bits were getting flipped and jitter was extreme, you still wouldn't have signal degradation in the ways described by audiophiles; you would get a raised noise floor. Random error is noise. Noise is random error.

Of course this is never an issue. We can send digital signals that are millions of times more complex than digital audio, with zero problems, using equipment that is insanely cheap. The assertion that the extremely low data rate PCM audio signals have some special risk associated with them is delusional.

> I don't understand the point you're trying to make. Everything is ultimately subject to physics, including optical signals. Nobody claims otherwise.

Sorry, I misformulated my sentence.

It should have read (emphasizing what I left out from the sentence in my previous post):

> You mean like those people, who argue, that: “Everything is a sequence of Zeroes and Ones, therefore the signal either gets transmitted or not!”, while totally leaving out the fact, that this digital signal materializes in the real world via an analogous signal, which is electricity, and therefore each digital data transmission (as long as not optical) is submitted to the physics of electricity, that happen totally outside of the "digital domain"?

Though, I think you are nitpicking, because, from the rest, it should be totally obvious, what I wanted to say. Use the error correction, man ;-)

Therefore:

If an analogous signal's "success" can suffer from the material, through which it travels, then a digital signal will suffer for the same reasons, simply, because it is not a digital signal, but, materialistically, both signals are of the same sort.

Even if we could create the perfect mathematical concept in our brain and have a solution for anything, as soon as we step into the real world, that is, the material side of affairs, many unexpected things can happen, that have to be accounted for at the next time.

In engineering we call this a "race condition". I like the definition on FOLDOC: https://foldoc.org/race%20condition

> ”Anomalous behavior due to unexpected critical dependence on the relative timing of events.“

Or, in other words: chaos!

I still don’t understand the point you’re trying to make. Certainly doesn’t have anything to do with how actual systems in the real world work.

You don't understand the concept of "race condition"?

For what it's worth I deal with race conditions on a daily basis with multi-threaded programming. What I don’t understand is the point you’re trying to make and how it relates to real digital audio systems. Can you please explain it.

Sure.

tl;dr: As soon as the "digital domain" steps into the "physical domain", you can not guarantee a "side effects" free system.

Let's start with a real-life example:

I have a Macbook from 2006, which has a Firewire port. I have an external audio-interface (Onyx Satelite Pro), that has a (Logitech Z4) active speakers system connected via line-out. When I got the Onyx, I had to realize, that the delivered Firewire cable was too short for my setup, so I had to purchase a longer one. I took great care to buy the cheapest one in the store (which was a Radio Shack type), because, hey, digital is digital, or isn't?!.

When I came home and exchanged the two cables the first thing I realized, much to my surprise, was, that the audio sounded different! The old cable sounded more "relaxed", "warm", "analogous". The new sounded "faster", "analytical", a little "sterile". How could that be? All I changed was the cable...?!

Conventional audio wisdom, especially of those, who call audiophiles "audiofools" dictates, that a digital signal is just zeroes and ones. The sequence either get transmitted or not. Add in the timing component, which is either to the point or off, and things should be pretty clear: What goes in is what comes out. No loss of information. Should there be such loss or distortion, we can error correct it, at least up to a certain degree. Therefore, one always get a perfect, or near perfect, signal.

That, at least, is the argument.

If that argument (alone) would hold true, and be sufficient to describe all, that happens in such a system, I would not have been able to hear a difference, by exchanging the digital interconnect, right?

But I did!

Therefore there must be something happening outside of the digital domain, that influences the signal. In my case, clearly the cable!

And now I come to the conclusion, which should answer your question:

> What I don’t understand is the point you’re trying to make and how it relates to real digital audio systems.

Former example covers exactly your request for a real digital audio system. What happens here is, that we step out of the digital domain into the real world domain, which means the physical and chemical domain. Namely, the domain of audio components. These components are made out of matter. In the case of electrical signals (Firewire, USB, etc.), the medium, which transmits our signal, is the same, which would transmit an analogous audio signal. Signals get transmitted as electricity with ever changing voltages (leaving out optical transmissions now). Therefore all, that applies to an analogous signal transmission, applies to digital signal transmission, as well, as long as the transmitting medium is electrical.

Now, you may, say, that that is not of importance, since the information carried is still a sequence of binary data, and it gets transmitted or not, given the theory behind it, it should be perfectly reproducable. But then, I ask you, why the difference in sound?

Given a perfect program (our digital concept), when applying it to the real world (physical/chemical), race conditions may occur, simply, because of unpredicted side effects (you know the importance of this), that come with this domain. Our perfect program, here, would be the theory of lossless, bit-perfect data transmission: the digital domain. Or - Theory. Our real world would be the cables, connectors, building blocks of electronics, which are: the chemical and physical domain. Or - Matter. And there is something happening in the material domain, that makes an impact on the sound. I don't know what that is. But I can clearly hear it on sub 1000€ system. Now imagine, how much more this could be heard on a stereo system, that cost 25.000€ or more...

Closing, I would like to requote the definition of "Race Condition" from http://FOLDOC.org/Race%20Condition:

> Anomalous behavior due to unexpected critical dependence on the relative timing of events.

The anomalous is the key word here. There are side effects, that happen outside of the "digital domain". And it is these side effects, that change the sound. And audiophiles know this. If a cable can change the sound, then capacitors, resistors, whatever, may well change the sound, too.

> ... given the theory behind it, it should be perfectly reproducable. But then, I ask you, why the difference in sound?

Placebo effect. 100% guaranteed. It is literally impossible for digital audio signals to degrade in any way that even vaguely aligns with your descriptions. It is analogous to using a different brand of SD card to get richer, more vibrant colours out of your digital camera. It is analogous to changing the ethernet cable on your computer to make web pages more insightful.

Bits don't work that way.

Whereas the placebo effect is very real and very powerful.

Firewire is also a two-way packetized, error corrected protocol. If a cable was not delivering the signal reliably, the consequence would have been a lower maximum transmission rate, not error that could affect sound. If there was any change in the signal due to the cable, your Onyx interface could not maintain a stable connection with your computer and you'd get no sound at all.

Your attempt to describe how digital signals are transmitted is actually just plain wrong.

Furthermore, race conditions have absolutely no relevance to this discussion.

I wrote:

> [Digital] Signals [when not transmitted optically] get transmitted as electricity with ever changing voltages [...]

You wrote:

> Your attempt to describe how digital signals are transmitted is actually just plain wrong.

Wikipedia writes:

> A waveform [...] as a digital signal [...]. The two states are usually represented by some measurement of an electrical property: Voltage is the most common [...]

Therefore I will leave it here.

Are you referring to noise being transferred to the analog stage, or digital data transmissions being corrupted?

The latter is very unlikely with asynchronous USB DACs since USB will retransmit corrupted packets, and jitter isn't a concern since the DAC buffers then uses it's own clock.

The former, well yeah, it's impossible to completely eliminate RFI no matter what you do, short of optically isolating the source, putting the analog equipment in a faraday cage, and running it all off of batteries.

See my reply to sjwright.

(But only if you want to bash your head against the wall with an epic misunderstanding of how digital signals work.)

Maybe you just need some HD ram to bring out the sparkling highs.

That still leave the dirty old storage drive. Fortunately Mojo audio have got that one covered. “Integrated Sorbothane resonance control under the chassis, main board, and SSDs isolate resonance.”

https://www.mojo-audio.com/music-servers/

There is now audiophile ethernet cable so why not.

> it's either going to make it or it isn't

This isn't _strictly_ true - I've seen basically the equivalent of noise in a HDMI picture (white dots all over the picture), and eventually tracked it back to a faulty HDMI cable.

I did replace it with the cheapest (<£5) HDMI cable I could find though, so there is that.

That's not 'noise', that is packets that didn't make it. It's not having a perfect picture vs. no picture, but about data packets either going from A to B or not going from A to B.

Say the HDMI cable is crap and it can only do 100 packets per second, and HDR needs 1000 packets per second, the HDMI chips are going to negotiate a HDMI transfer settings without HDR. Going back further, say the HDMI data path is negotiated to do only basic HDMI, say, DVI-like video. If even that is too much for the cable to handle, say, it constantly drops a few packets, you would get either perfect pixels or missing pixels.

HDMI doesn't have packets. It's largely digital VGA. It's very easy to flick individual bits, and the effect is basically the same as analog noise.

The only difference is the threshold.

You are correct. I suppose serial data is more like a clocked stream than packets. Perhaps I was thinking of TDMA instead of TDMS. Flicking bits is easy indeed, which should be mostly mitigated with the requirements on the conductor and shielding.

> very easy to flick individual bits

Is there any error correcting code/protocol in HDMI spec to fix individual bit flips?

No. There's an 8b/10b encoding to balance current but that doesn't help you fix much of anything.

Displayport has packets with a checksum but I don't think there's any resending, only discarding.

hdcp

hdcp uses the DDC AFAIK. The data is scrambled but still RGB.

but wont decode if randomly modified

It's a stream based cipher, so you'll get the same amount of blips as you'd get with normal bit damage. Difference is probably that you can get a whole missing frame as a blip instead of just missing pixels, because the stream key operates on frames instead of single pixels.

You are right. I went back and read hdcp spec. Now I understand my confusion. What I experienced multiple times with hdcp and bad/low quality cable/connection was screen going blank for a short moment. I always assumed it was cryptography itself breaking apart, but the spec requires resetting connection in case of logged TMDS layer errors

" While the receiver may reach a determination sooner, the receiver must determine loss of synchronization at least by the time it has detected 50 consecutive data island packets with ECC errors. The HDCP Receiver must either assert the REAUTH_REQ bit of the RxStatus register or de-assert HDCP_HPD to the upstream transmitter once it determines loss of synchronization. "

The examples you linked are crazy. It is possible, especially with the Ethernet, to just check the packet error rate, and see if there is a problem.

There's something to be said for a bit of extra effort in filtering, but most of what audiophiles pursue has no basis in electrical engineering reality.

And then there's even weirder audiophile accessories like cable elevators, amplifier stands, CD "demagnetizers", and "quantum chip" stickers -- all complete nonsense which doesn't even have any plausible way of affecting the audio.

If you believe in the placebo effect, then perhaps the nonsense can have a verifiable effect!!!

I think this is a straw man, since a majority of audiophiles will tell you that a 1000 dollar HDMI cable is stupid. nobody actually buys that shit.

What does matter is getting a decent HDMI cable. I'm going to get a braided cable which I know supports the bandwidth of HDMI 2.1 (I think that's the current version).

1k? No. Probably more like 25.

Technically there is jitter in the digital domain, and it's possible to see problems at high bandwidth signals, but in low bandwidth transmissions like Flac, you're absolutely right.

If you have nanosecond timing requirements, then those might be useful, but for audio they are overkill.

https://en.wikipedia.org/wiki/Jitter

The difference is that with digital signal there is limited number of components that actually make a difference and as soon as a certain level of quality is reached (basically reliable transmission) nothing can be improved further.

For listening purposes the transmitted data can be buffered on the receiver DAC. This means that as long as the buffer on DAC is not emptied everything else may be failing badly and it will have zero effect on your listening experience.

FLAC????? Ew... Every audiophile knows that WAV sounds the best. FLAC quality is mediocre compared to WAV.

Update: geez, it was satire...

...or was it?

https://serato.com/forum/discussion/1541194

https://www.head-fi.org/threads/flac-vs-wav-format-surprisin...

You were not downvoted because people took your comment literally, you were downvoted because being snarky runs against the Hacker News guidelines, specifically:

> Be civil. Don't say things you wouldn't say face-to-face. Don't be snarky. Comments should get more civil and substantive, not less, as a topic gets more divisive.

Sometimes snarky comments will be forgiven if they are substantive. Yours contributes very little beyond what the parent/grandparent comments had already mentioned.

Thank you for the quick heads up! I’ll make sure to familiarize myself with said guidelines. It’s very useful!

I’ll pay a bit extra for a nicely wrapped cable with really good anti-tangle and strain relief properties, because that’s convenient.

And that's exactly what real high quality cable is.

Professional cables are usually flexible, with robust insulation and connectors. They may also be fire rated. Quality control ensures that cables will always be up to specs because people doing the installation have other things to do than dealing with defective cables.

Like most pro stuff, pro cables are not "better", they are cables that won't be a problem.

Sometimes “professional” just means “designed to handle a lot of abuse”.

I often counter with the idea of transmitting an ebook over the same cable. "Does it become more detailed, in depth with more character development?"

The only thing I can agree is better on a digital cable is increased shielding to avoid ground noise.

Admittedly I chuckled. The amazon cable you referenced is full of satirical comments.

One theory I read (for good quality USB cables to a DAC) is that reduced packet loss reduces jitter, which means that the clock rate of the DAC doesn't have to adjust in order to keep the buffer populated, but not overflowing. Sounds plausible, but I'm not sure that packet loss is high enough to actually cause this.

A working USB cable does not create packet loss and definitely does not fix packet loss; if packet loss occurs because of the cable, the cable is defective, nothing else. There is no good quality and bad quality, there are cables that work as expected and cables that are defective.

But there is capacity, and USB1.1 may need less than USB2. That said, we can assume that we are talking about USB cables that are tested/verified/certified, which means they have to follow a few rules about how much Mhz bandwidth it supports, how much EMI it can deal with and how much drop in power it's going to cause. But of course, those are measurable facts and still have nothing to do with the data packets getting 'warm' or 'fuller' or some other subjective term.

More or less what I thought about the network cable in GP's post - if the network cable is changing how your FLACs sound, it's _broken_.

Yeah but still, the biggest loss is going to happen within your plug/jack connection. So if you really care, solder your cables in. Better than every gold plated connector your audiophile purse is ever going to buy.

Also: Acoustics are what matters the most (and they are the hardest to get right, just by throwing money at the problem, that is why it is not usually beeing discussed that much).

Jitter is itself an audiophile scam. The audible consequences of extreme jitter is an increased noise floor, which itself is being drowned out by the sound in a typical "quiet" room.