$300 too much for a 40cm digital interconnect?

[louballoo]

Definitely good enough, it is. But what of noise, for instance when I ran a squeezebox classic connected to a dac, the dac maker told me to run the optical connection, because of what I understood to be Heligoland noise. Wouldn't there be similar mechanisms operating when connecting a computer via USB to a dac?

The optical connection provides complete isolation between the two electrical systems, the squeeze box and the DAC. This shouldn't be necessary with any quality gear and shouldn't be a concern for audiophiles. I use an optical connection because it can go farther without issues.

"Heligoland" ? I'll have to look that up, but I suspect it's just another obscure term that ******* like to use without knowing its meaning.

Louballoo

 

[Peter Stockwell]

"Heligoland" ? I'll have to look that up, but I suspect it's just another obscure term that ******* like to use without knowing its meaning.

Louballoo

Sorry about that, didn't proof read. Should have been "ground plane". And it's true I know next to nowt about 'lectric circuits, let alone the innards of a computer.

 

[ItemAudio]

The optical connection provides complete isolation between the two electrical systems, the squeeze box and the DAC. This shouldn't be necessary with any quality gear and shouldn't be a concern for audiophiles. I use an optical connection because it can go farther without issues.

The question Peter asked was about USB grounding - your reply re: optical seeks to change the subject.

With USB, we have to ask: what does the ground plane look like?
With optical - different question: we have to ask: what does the electrical environment of the driving clock look like?

In theory, all manner of things look fine that just don't work in practice. Of course, yes, the DAC - potentially, theoretically - can eliminate transport aberrations. DAC makers all implement similar measure to isolate and reclock, and speak about their 'proprietary' system as if it's unique - and that sounds jolly persuasive - in theory - until you put it to test. From those without practical experience in this area, there tends to be a lot of regurgitated hearsay, and naive trust in theories advanced by certain manufacturers. All I'm saying is: it's not that simple: try stuff.

 

[Still]

This picture answers the question: what do I think item's listening room look like? ;-)

 

[Robert]

With USB, we have to ask: what does the ground plane look like?
With optical - different question: we have to ask: what does the electrical environment of the driving clock look like?

In theory, all manner of things look fine that just don't work in practice. Of course, yes, the DAC - potentially, theoretically - can eliminate transport aberrations. DAC makers all implement similar measure to isolate and reclock, and speak about their 'proprietary' system as if it's unique - and that sounds jolly persuasive - in theory - until you put it to test.

You have this the wrong way round.
The theory is demonstrated to be correct when put to the test.

 

[PhilCTTE]

I'm not sure if SPDIF has some synchronization mechanism (like USB does), but if not (likely), then "yes", assuming the software is done correctly, ie. real-time system, that guarantees latencies below a certain point (say, 50-100μs for a userspace application - music player), as opposed to "normal" operating systems, which achieve about 1-5ms during idle, much more during load.

That is - on common PC hardware.

Again - I'm no hardware designer (John, Dominik or someone like that might have a better answer), but the only "always perfect" solution here is USB, .. ehrm .. "asynchronous USB".
If SPDIF has no means of source<->destination synchronization and the clocks are not synchronized as well, then there's a real possibility of a clock speed mismatch. I believe that's called jitter. Even if there was destination-side buffering, faster source clock would fill the buffer completely and the destination would have to start dropping data. Similarly, if the source clock was slower, the buffer would soon be empty, with no data to play.

That's not the issue with USB, where the destination can temporarily stop the source from sending more data (through URB), so the buffer can be kept under control. Similarly, when the data get corrupted during the transfer (USB checksums data), it can be re-transmitted again and added to the buffer, without you noticing it. It (in theory) always results in a "bit-perfect" 1:1 reproduction, with playback correctly clocked by the destination's exact clock. The source here is merely a data storage, not an active sender, like it's in the case of SPDIF.

Note that I'm not saying "asynchronous USB" is easy to implement in hardware, AFAIK some manufacturers do it wrong (incorrect crystal sync with buffer ?).

While SPDIF works perfectly only when the source and destination have clocks synchronized, USB can work perfectly even on a common PC hardware, running a common OS, because of the buffer.


PS: If SPDIF can actually be destination-controlled and thus buffered and I got that part wrong, then please disregard it.

Clock speed mismatch and buffering issues is what I understood SPDIF interface battles with. Unless you have a separate word clock I/O to re-align it all.

What happens to the sound with buffer overflow and underflow, do they both result in dropout/clicks with the sound or is it more of a smear, veil > lack of clarity.

I keep hearing that most well designed modern DACs can sort out most of these timing issues with buffering. When I was testing 96/24 hires source material via Optical SPDIF on the SBT against Async USB from my laptop I could detect a slight lack of clarity on decay of a single guitar chord. This was into my MDAC. I repeated the test and had no reason to prefer one or the other , but the test was sighted.

There are other issues with USB in that it requires a conductive cable which results in ground loop issues and other unwanted transmissions which is why i use an isolator. And again I tried with and without this component and heard an increase in definition/clarity.

Back to the point raised on $300 on a screened SATA cable. If you can isolate the DAC optically and put some distance between computer and DAC does it not make everything else in the chain redundant?

Wouldn't a USB to Optical convertor be the best solution prior to running it into a DAC when using a computer as a source. I know John Westlake was in the process of designing one.

What I havent tested yet is a quality audio board with SPDIF and word clock I/O. Anyone care to comment who have compared them both ?

 

[earwaxxer]

It seems to me that current day DAC's are getting better and better at clocking and jitter.

Take the Schiit Gungnir as an example. I recently got one, and I am very pleased at how well this DAC handles various digital inputs. I could not find a difference in sound between any of the SPDIF inputs. I have an old MSB dac that clearly sounds better via toslink. Using my Transporter as a server, the Gungnir sounds great from either toslink, coax, or BNC. I didnt get the USB because I feel USB still has much growing to do before it is ready for prime time!

 

[ItemAudio]

Back to the point raised on $300 on a screened SATA cable. If you can isolate the DAC optically and put some distance between computer and DAC does it not make everything else in the chain redundant?

Wouldn't a USB to Optical convertor be the best solution prior to running it into a DAC when using a computer as a source. I know John Westlake was in the process of designing one.

Opticis has such a component off-the-shelf: great idea, lousy execution.

However, there are now high-sped PCIe/optical isolators that operate upstream of the USB chipset, enabling a bare minimum of support for USB3 to be powered by a battery of linear PSU (depending on whether the DAC prefers to see a ground at that point). That really changes the game . . .

However, even in the T1-USB, which firewalls the whole PC behind an Adnaco S3B fibre-optic PCIe connection, what is reluctantly and none-too-accurately being called 'software jitter' still seems to be in play.

 

[avole]

item, are you sure it's wise to add links to the products you're trying to flog through your posts actually in your posts? Shouldn't they be restricted to the trade section?

 

[ItemAudio]

It seems to me that current day DAC's are getting better and better at clocking and jitter.

Take the Schiit Gungnir as an example. I recently got one, and I am very pleased at how well this DAC handles various digital inputs. I could not find a difference in sound between any of the SPDIF inputs. I have an old MSB dac that clearly sounds better via toslink. Using my Transporter as a server, the Gungnir sounds great from either toslink, coax, or BNC. I didnt get the USB because I feel USB still has much growing to do before it is ready for prime time!

I've not heard the Gungnir, but in principle this is what we would expect: some DACs are better at leveling differences between transports. Interestingly, those that excel at this are not always the most expensive, or best-sounding, converters. There seems to be little correlation between 'properly designed' by that measure, and highest performing.

Not to be taken too literally, the situation reminds me of car design, where there is an eternal trade-off between razor sharp sports handling that shows you every bump in the road, and wallowy, magic carpet, grand tourer suspension. Some top end and mid-range cars combine both virtues reasonably well, but many supercars go all out for speed and to hell with the ride: because creating immunity to bumps would interfere with their design goal.

DACs tend to be either designed for USB, or have a USB input. Is a DAC that sounds identical through SPDIF / USB really, really well designed? Or just not very good at doing USB? And when tested via SPDIF, which clock is driving it? When auditioned via USB, which computer? It's difficult to assess.

 

[Mike42]

DACs tend to be either designed for USB, or have a USB input. Is a DAC that sounds identical through SPDIF / USB really, really well designed? Or just not very good at doing USB? And when tested via SPDIF, which clock is driving it? When auditioned via USB, which computer? It's difficult to assess.

What does any of this mean in practice? What happened to 'just listen and decide let your ears decide...'?

 

[Cereal Killer]

I think what hes trying to say is; Its easy to make USB sound like SPDIF, but not so easy to make SPDIF sound as good as (well implemented) USB.

 

[PhilCTTE]

Opticis has such a component off-the-shelf: great idea, lousy execution.

However, there are now high-sped PCIe/optical isolators that operate upstream of the USB chipset, enabling a bare minimum of support for USB3 to be powered by a battery of linear PSU (depending on whether the DAC prefers to see a ground at that point). That really changes the game . . .

However, even in the T1-USB, which firewalls the whole PC behind an Adnaco S3B fibre-optic PCIe connection, what is reluctantly and none-too-accurately being called 'software jitter' still seems to be in play.

Yes spotted that Adnaco on your site, but you are relying on how well designed the USB port is and its associated power supply. It would isolate the PC but not from itself. Would you still need a USB isolator ?. Whats your experience of them. £375 to isolate the usb. I guess it would save on everything else. $300 sata cables and all.

I'm still in the "blissfully unaware" camp when it comes to anything upstream in the audio conversion process. I care upto the point of the incoming signal to the DAC. I can't see that treating and "tricking out" anything in the PC if its isolated from the DAC would matter in regards to an Async USB interface. Always open minded to be persuaded otherwise by audition.

 

[ItemAudio]

The business end of the Adnaco is a pair of USB 3.0 outputs driven by a TI chipset, fed by the 5Gt/s optical transcoder (ie, military/telecom spec, not Toslink). And not much else.

You bring the 5V from an external supply - squirt your preferred flavour of clean juice (linear or battery) into a very simple board and the DAC sees nothing but the bare essentials required to deliver USB. It sounds like just noughts and ones.

Happy to lend you one to play with, Phil . . .

 

[avole]

Like a shark circling...

 

[ItemAudio]

Like a shark circling...

avole: doesn't saying 'neeeargh' every time I post - whatever I post - get boring?

Given your absolute faith in the impossibility of transports differing, I'm surprised you don't want everyone to hear the 'upstream foo' to expose the scam - why is that?

By contrast, I'm confident enough to put my money where my mouth is, and let people judge for themselves - on a free, no pressure, tyre-kicking, extended home loan - so they can actually hear something to have an opinion about.

 

[avole]

How can you hear foo, item?

 

[Cereal Killer]

By contrast, I'm confident enough to put my money where my mouth is, and let people judge for themselves - on a free, no pressure, tyre-kicking, extended home loan - so they can actually hear something to have an opinion about.

Good plan that , Avole, are you up for a dem?

 

[sonddek]

Cutting through the foo, what all of this boils down to is...

- if you're worried (duly or not) about electrical isolation, use TOSLINK or ethernet, or an isolator gizmo on USB
- if you're worried (duly or not) about the question of whether anybody, even item, hears modest upstream jitter, use a DAC which has a sensible buffer
- if you're worried (duly or not) about the 1's and 0's reaching your DAC accurately, make sure that your source software isn't fiddling with the bits

Game over, and don't waste any money on cryomobos and other foo sourcery.

The reason why item's posts are so long and incomprehensible is that he persistently seeks ways to insinuate that digital systems behave like analogue ones, viz. the bogus noisy room analogy. Digital systems are deliberately, by design, impervious to that kind of noise. They are a foghorn in that room of 100, and they only have to measure whether the foghorn blew or not. If they fail due to noise, they do so catastrophically, i.e. no signal at all. The fact that your computer boots up most days shows that item's effort to apply analogue signal ideas to digital circuits is largely a fallacy. Don't swallow the noise/jitter gibberish - it is designed to establish an obsessive compulsive bit-cleanliness fixation. Only buy DACs which don't amplify upstream jitter. Computers work fine. Spend your money on music.

 

[avole]

I can't hear foo, sorry. Also he'd have to send the stuff to France. Besides, I already have an opinion, which is what you'd expect given the field in which I work.

Edit: couldn't have put it better myself, Sonddek, a big +1