Is Digital Audio Transmission Analog? [video]

[Ken Tajalli]

The *transmission* (physical layer) may be anything you want -the whole point it so make it irrelevant. The actual information/data ("application layer") is digital, and hence unaltered

Yes, absolutely.
As I said in a connected thread, digital transmission carries no audio, so the transmission side of it is irrelevant.
It is akin sending a recipe to another chef to recreate a dish from scratch. you can send it by Morse code or email, that is irrelevent.
The recipe carries no food! can not be contaminated. As long as the chef at destination (the DAC) can read the recipe perfectly, he can cook an identical dish (the music).
That's what digital data is. Nothing but a set of instructions for the DAC to recreate the audio voltages from a Powersupply.

 

[KSTR]

I think it is clear and universally undisputed that bit-perfect symbol transmission is required to have identical analog output, and it is readily achieved.

Nothing but a set of instructions for the DAC to recreate the audio voltages from a Powersupply.

That's where the hard part is in DAC design. Bit-perfectness alone is not sufficient to render the exact same final analog signal for a number or reasons:

• EMI issues (those often manifest themselves further down the chain, the "ground loop" type of scenario comes to mind where the signal at the DAC may still be perfect but falls apart at the end of an unbalanced cable to the amp)
• Clock jitter which may come from the transmission itself (SPDIF etc), local clock jitter (from power supply disturbances), combined with hidden resampling (like ESS chips do in many of their operating modes)
• Voltage reference disturbances which may be random but can also be data-depedent.

So in a complete system setup it is actually possible that the final analog signal as we use it can be different even with one and the same DAC device, depending on circumstances, even though bit-perfectness was never violated.

 

[rs172]

Yes, absolutely.
As I said in a connected thread, digital transmission carries no audio, so the transmission side of it is irrelevant.
It is akin sending a recipe to another chef to recreate a dish from scratch. you can send it by Morse code or email, that is irrelevent.
The recipe carries no food! can not be contaminated. As long as the chef at destination (the DAC) can read the recipe perfectly, he can cook an identical dish (the music).
That's what digital data is. Nothing but a set of instructions for the DAC to recreate the audio voltages from a Powersupply.

It all seems logical to me, I think you gave a great example here with food.
Coming from Computer Science, according to what see here, all streamers are identical when used as digital transport (sound wise) and companies like hifi-rose etc are scamming potential customers and all these youtube reviewers who say they hear "dramatic" differences between streamers are either lyres being paid by the manufacturers or imagining differences (their brain tells them they are hearing better sound quality).

But I have to say that there is a youtube channel "Virtual Hifi" who puts out demos through a high performance microphone in his room and lets you hear the differences between different streamers using the same dac. The last test where I definitly heard a difference was between the Eversolo dmp a6, dmp a6 Master Edition and dmp a8. I could easily hear the soundstage differences using headphones (a clear difference between the A6 and the A6 MasterEdition, and barely a difference between the A6 Master Edition and the A8)
The question is, is it a hoax? You are welcome to check his review yourself and comment afterwards

 

[Purité Audio]

Amazing just how powerful the imagination is.
Keith

 

[Ken Tajalli]

EMI issues (those often manifest themselves further down the chain, the "ground loop"

A streamer can not add noise of any kind. Decent DACs have isolation built-in.
Even a noisy USB from a PC is no match for a galvanic isolator.

• Clock jitter which may come from the tr

A few years ago, perhaps. These days the jitter issue is no isuue.

• Voltage reference disturbances which

Even a decent $100 DAC has almost no noise or PSU issues.

But aren't we side tracking?

 

[voodooless]

We all know there are possible mechanisms that could cause audible issues, but:

a. The likelihood with current well-designed equipment is negligible, except for maybe ground loops, which is a hit-and-miss anyway, and have little correlation with product quality
b. The audibility claims are never of the kind that can be ascribed to the possible issues described above

 

[pablolie]

...But I have to say that there is a youtube channel "Virtual Hifi" who puts out demos through a high performance microphone in his room and lets you hear the differences ...

You do realize Youtube is not a great example when it comes to supposedly listening to minute differences?

 

[rs172]

You do realize Youtube is not a great example when it comes to supposedly listening to minute differences?

Yes I know that Youtube is 256 kbps lossy audio format and even so, I can't dismiss what I hear. The sound seems to be a bit more open. Please watch the attached link using headphones and tell me what you think. Maybe I am imagining and its all in my head but I listened to the test with my eyes closed so that I won't be able to see which one of the streamers is playing at that specific moment.
Please share your comments

 

[pablolie]

Yes I know that Youtube is 256 kbps lossy audio format and even so, I can't dismiss what I hear. The sound seems to be a bit more open. Please watch the attached link using headphones and tell me what you think. Maybe I am imagining and its all in my head but I listened to the test with my eyes closed so that I won't be able to see which one of the streamers is playing at that specific moment.
Please share your comments

There is also the slight possibility someone had an agenda as they put the video and its sound together...

 

[voodooless]

All of the audio is horribly distorted. It’s unclear how good the level matching is, or what other settings may have been different between devices. Plus you do not have the same parts of the audio to compare with. All this makes it highly unreliable.

 

[Purité Audio]

Yes I know that Youtube is 256 kbps lossy audio format and even so, I can't dismiss what I hear. The sound seems to be a bit more open. Please watch the attached link using headphones and tell me what you think. Maybe I am imagining and its all in my head but I listened to the test with my eyes closed so that I won't be able to see which one of the streamers is playing at that specific moment.
Please share your comments

I actually have those two units in front of me now, there is no difference in SQ or in ‘soundstage’ through the same speakers of course.
Keith

 

[antcollinet]

The question is, is it a hoax?

Well two points.

Even tiny differences in the mic position in the room, or changes in the position of other objects in the room (such as the person carrying out the tests) can result in differences in the sound as the sound field is disturbed.

Even if this is not the case - if you are not comparing blind, level matched, and statistically relevant, then your comparisons are likely impacted by cognitive biases - which can cause you to perceive differences even when none exist.

 

[KSTR]

A streamer can not add noise of any kind.

Of course it can (as can any other source), like lots of common-mode EMI/noise on the cable.

Decent DACs have isolation built-in.

I'm not aware of many DACs or other devices that have built-in isolation for the standard interface which is USB. TOSLINK yes, by design, SPDIF/AES3 input should have isolation but that's not always the case. AES3 cable shields are often connected at both ends to the chassis which completely defeats the purpose of the signal isolation transfomer. Same for shielded Ethernet connections.

A few years ago, perhaps. These days the jitter issue is no issue.

How do you know when almost nobody, including Amir, tests for jitter of sources or jitter susceptibility of devices? Note that the J-Test that Amir does only tests for data-dependent jitter of SPDIF/TOSLINK-fed DACs (for USB it's totally irrelevant) with a source -- the AP -- that is jitter free.

Even a decent $100 DAC has almost no noise or PSU issues.

Yeah, maybe some modern $100 device have actually decent reference voltages and general good supply, but that's not always the case.

----------:---------

We simply do not know what's happening at the consumer's home, we don't know the devices and don't know the setups. So we cannot simply make a statement that this stuff is all non-issue in practice.

Obviously, "decent" DACs etc show no issues when tested in isolation under lab conditions... which is a somewhat circular thing as "decent" by definition here means "measures well" under just that circumstances, within the restrictions of the actual tests performed (see above).

But aren't we side tracking?

No. Quoting initial post "And for this reason, everything digital can be subject to audible difference from digital audio cables to digital output of streamers."

 

[pablolie]

Of course it can (as can any other source), like lots of common-mode EMI/noise on the cable.

It can happen in very extreme circumstances, but highly unlikely to find any emissions that distort even the analog sound these days. For digital signals, it does not have an effect at all.

I'm not aware of many DACs or other devices that have built-in isolation for the standard interface which is USB. TOSLINK yes, by design, SPDIF/AES3 input should have isolation but that's not always the case. AES3 cable shields are often connected at both ends to the chassis which completely defeats the purpose of the signal isolation transfomer. Same for shielded Ethernet connections.

These are *digital* interfaces. Unless the spec demands it, shielding is a waste of money. Digital signals don't care about physical media as long as it's up to standards.

We simply do not know what's happening at the consumer's home, we don't know the devices and don't know the setups. So we cannot simply make a statement that this stuff is all non-issue in practice.

The stuff you're talking about performs flawlessly in far more critical and demanding environments than any home. Military. Space navigation. Hospitals. etc etc. At 800Gbps speeds too. Your music streaming at a few Mbit/s is safe.

 

[KSTR]

You intentionally misunderstand what I'm talking about, don't you? Besides opening secondary theaters (space navigation) to distract from the topic.
Sorry, end of discussion from my side. Have a nice day.

 

[amirm]

How do you know when almost nobody, including Amir, tests for jitter of sources or jitter susceptibility of devices? Note that the J-Test that Amir does only tests for data-dependent jitter of SPDIF/TOSLINK-fed DACs (for USB it's totally irrelevant) with a source -- the AP -- that is jitter free.

AP is not jitter free. No device is.

Regardless, the J-test signal that I run comes over USB to the DAC, simulating exactly what happens when you play music. AP is not involved in that output pipeline. I have also played J-test with Roon with many streamers to same results.

Anyway, the points you are raising are already covered in my video. Problem for folks making these claims is that not one of them has come up with a single measurement showing the output of the DAC changes with different streamers. You would think they are highly motivated to find at least one scenario but none exists. So while the probability exists, there is no evidence of it happening in practice.

 

[pablolie]

You intentionally misunderstand what I'm talking about, don't you? Besides opening secondary theaters (space navigation) to distract from the topic.
Sorry, end of discussion from my side. Have a nice day.

I do not "intentionally misunderstand" anything you wrote. I replied to your point with the simple technical truth about physical interfaces, and the fact these have been proven in far more demanding and critical environments than playing a -believe it or not- undemanding music stream. The problem is there's too much charlatanerie in audio, making up esoteric arguments that for some odd reason don't hold true in applications that would be far more likely to uncover such issues.

 

[Blockader]

Of course it can (as can any other source), like lots of common-mode EMI/noise on the cable.


I'm not aware of many DACs or other devices that have built-in isolation for the standard interface which is USB. TOSLINK yes, by design, SPDIF/AES3 input should have isolation but that's not always the case. AES3 cable shields are often connected at both ends to the chassis which completely defeats the purpose of the signal isolation transfomer. Same for shielded Ethernet connections.


How do you know when almost nobody, including Amir, tests for jitter of sources or jitter susceptibility of devices? Note that the J-Test that Amir does only tests for data-dependent jitter of SPDIF/TOSLINK-fed DACs (for USB it's totally irrelevant) with a source -- the AP -- that is jitter free.


Yeah, maybe some modern $100 device have actually decent reference voltages and general good supply, but that's not always the case.

----------:---------

We simply do not know what's happening at the consumer's home, we don't know the devices and don't know the setups. So we cannot simply make a statement that this stuff is all non-issue in practice.

Obviously, "decent" DACs etc show no issues when tested in isolation under lab conditions... which is a somewhat circular thing as "decent" by definition here means "measures well" under just that circumstances, within the restrictions of the actual tests performed (see above).


No. Quoting initial post "And for this reason, everything digital can be subject to audible difference from digital audio cables to digital output of streamers."

I agree with your points I just want to add that fancy R-2R dacs usually have galvanic isolation. Soekris 2541, Holo May for example.

 

[voodooless]

I agree with your points I just want to add that fancy R-2R dacs usually have galvanic isolation. Soekris 2541, Holo May for example.

Well, it didn't help the Soekris one bit. At least the Holo is a decent performer.

 

[Blockader]

Well, it didn't help the Soekris one bit. At least the Holo is a decent performer.

The goal of the galvanic isolation is basically to cut off the noise coming from the source, PC for example. Those noise artifacts may not be visible if the source is clean enough. For example, I was able to hear the polling rate of my mouse or hear a high pitched noise when my gpu is working hard while MiniDSP 2x4 was connected to my PC. This is not the case with Motu M2 I am using right now.

Talking about Soekris dacs, IIRC Soekris 2541 has around 94db SINAD unlike Soekris 1421. This aside, ultimately -80db distortion is as inaudible as -120db distortion assuming we are listening at 85db average. Holo May, Motu M2, Soekris 2541 are equally transparent.

We always ask audiophiles to test their ears. Maybe we should put ASR members to test too. https://www.klippel.de/listeningtest/ Is there a point to worry about -80db distortion if I can't hear -80db in distortion listening tests? If you can hear -80db distortion, I would like to know.