Why does SPDIF basically suck?

[Summit]

4 hours ago, Audiophile Neuroscience said:

 

I do not claim anything apart from telling you what I heard on what gear.As I said YMMV.

 

You can favor whatever digital interface you want. All am asking for is for you and everyone else that says that S/PDIF is inferior or crap to tell me with which audio gear you made this apple to apple comparison. It’s good practice that one declare the reference if ask for.   

[Ralf11]

5 hours ago, mansr said:

Did you find a video of me or what?

 

you are in that remake of a Bergmann film

[Ralf11]

Thread Summary:

 

1.  S/PDIF solves the galv. isolation problem of USB noise transmittal

2.  but S/PDIF may have jitter problems

 

what else?  (in simple terms for the technically ignorant consumer who may stumble across this thread)

[mansr]

21 minutes ago, Ralf11 said:

S/PDIF solves the galv. isolation problem of USB noise transmittal

Only if it is transformer coupled or optical. Many implementations are not.

[Ralf11]

oops - I meant optical

[Audiophile Neuroscience]

9 hours ago, adamdea said:

11 hours ago, Audiophile Neuroscience said:

Mans, I looked you up. It's official, I am old

...and I feel like a dumb monoglot

I have to admit to being slightly disappointed (and impressed as hell) that Måns' English does not have the heavy Swedish (?) accent I had imagined in my head when I read his posts.

 

Yeh, in my mind's eye, @mansr was a German middle aged ( or older) man. All I can say now is this very bright young man needs to learn to agree with me more often purely out of respect to his elders .

[mansr]

8 minutes ago, Audiophile Neuroscience said:

Yeh, in my mind's eye, @mansr was a German middle aged ( or older) man. All I can say now is this very bright young man needs to learn to agree with me more often purely out of respect to his elders .

Young? I don't even get carded in the States any more.

[Audiophile Neuroscience]

8 hours ago, Summit said:

 

You can favor whatever digital interface you want. All am asking for is for you and everyone else that says that S/PDIF is inferior or crap to tell me with which audio gear you made this apple to apple comparison. It’s good practice that one declare the reference if ask for.   

 

I listed MSB, Bricasti and Gryphon,,,,,Levinson but not direct comparison - and various others I cannot recall all the details well enough to list except I know I compared the dig outputs as well as dig filters and upsampling  (if offered). I wasn't doing it for a scientific comparison, just my own interest, general impression.

[Audiophile Neuroscience]

6 minutes ago, mansr said:

Young? I don't even get carded in the States any more.

 

Haha, I saw (what I thought was ) you in a 2012 video about optimizing Linux on ARM. You looked about 12 (kidding). when I was 30 I still looked about 17 which annoyed me no end.Then one day I looked in the mirror and wondered why I saw my father looking back! That annoyed me even more!

[mansr]

4 minutes ago, Audiophile Neuroscience said:

Haha, I saw (what I thought was) you in a 2012 video about optimizing Linux on ARM.

Such a video does exist.

[Audiophile Neuroscience]

4 hours ago, Ralf11 said:

oops - I meant optical

 

Seems those Coriolis forces are just as active in the Northern hemisphere at pulling out the wrong words

 

Hey and watch those transmittal vs transmission different meaning nuances - noise transmittal sounds like the noise arrived in a memo

[Ron Scubadiver]

It sucks, who says so that actually knows?

[miguelito]

12 hours ago, Summit said:

How does a word clock connection like DCS  change how the S/PDIF clock sync work?

If you read the beginning of the thread, one of the issues with SPDIF is clocking jitter. In the USB case you can have a buffer that you can reclock on the DAC side because you can tell the transmitter "ok stop sending data". Cannot do that with SPDIF. The way dCS and others have resolved this issue is by having a high precision master clock clocking the whole system - no need to tell anyone to stop since the whole thing is in sync. 

[Fitzcaraldo215]

12 hours ago, miguelito said:

If you read the beginning of the thread, one of the issues with SPDIF is clocking jitter. In the USB case you can have a buffer that you can reclock on the DAC side because you can tell the transmitter "ok stop sending data". Cannot do that with SPDIF. The way dCS and others have resolved this issue is by having a high precision master clock clocking the whole system - no need to tell anyone to stop since the whole thing is in sync. 

There are technical architectures and there are implementations.  With the right implementation, spdif does not suck, nor does USB or any other method.  But, the architecture of the scheme affects what must be done in the implementation. Also, anecdotal reports about this or that "sounding better"  are likely much more about implementation quality differences rather than revealing any generalizable truths about the respective architectures.

 

For PC interconnection, a USB port for audio output is almost universally available, but a port suitable for spdif, AES, Toslink, etc. is not necessarily always present. That is +1 for USB.

 

I have no no doubt that the DCS scheme is a good one, and I am reasonably sure that they have gone to great lengths to minimize jitter in the clock leads to which everything is synced.  But, look at the expensive hoops this architecture has made them jump, including their costly external clock, no big deal, though, if you can already afford a DCS DAC.  

 

Other well implemented schemes using spdif similarly have to go to great lengths in the implementation to wring out the jitter inherent in that architecture, with the receiving unit having to accurately synchronize with the clock in the sending unit.  It is doable in a way to render jitter inaudible, but at a cost in development and implementation.

 

Asynch USB, properly implemented, has become a much more standardized and commoditized mechanism, an adaptation of an inexpensive, non-proprietary, universal USB data transmission standard.  Properly implemented, it positively and absolutely solves the clock synchronization issue in data transmission.  But, yes, proper software in the sending unit is necessary to control the data output rate and simple, straightforward logic is required in sending/receiving units. That seems no big deal these days, and it is much more economically implemented than something like a jitter free, external clock box with connections and interfaces.  That mechanism also requires programmed control logic, by the way.

 

There may be downsides to USB, including higher transmitted electrical noise.  But, there are also solutions to that, such as galvanic isolation in the DAC, again an implementation issue, but not exactly rocket science.

 

I see no horrors in these different architectural schemes.  I only see better or worse implementations.  But, I prefer a well implemented asynch USB architecture because I think it more economical and it easily supports higher audio sampling rates and channel counts.

 

 

 

[miguelito]

2 hours ago, Fitzcaraldo215 said:

There are technical architectures and there are implementations.  With the right implementation, spdif does not suck, nor does USB or any other method.  But, the architecture of the scheme affects what must be done in the implementation. Also, anecdotal reports about this or that "sounding better"  are likely much more about implementation quality differences rather than revealing any generalizable truths about the respective architectures.

 

For PC interconnection, a USB port for audio output is almost universally available, but a port suitable for spdif, AES, Toslink, etc. is not necessarily always present. That is +1 for USB.

 

I have no no doubt that the DCS scheme is a good one, and I am reasonably sure that they have gone to great lengths to minimize jitter in the clock leads to which everything is synced.  But, look at the expensive hoops this architecture has made them jump, including their costly external clock, no big deal, though, if you can already afford a DCS DAC.  

 

Other well implemented schemes using spdif similarly have to go to great lengths in the implementation to wring out the jitter inherent in that architecture, with the receiving unit having to accurately synchronize with the clock in the sending unit.  It is doable in a way to render jitter inaudible, but at a cost in development and implementation.

Completely agree on the diffs between design and implementation, and btw the vagueness of "this sounds better", especially comparing single ended vs balanced spdif.

 

Yes, the dCS clocks ain't cheap! I know first hand...  But even with a great clocked system, all in sync, if you feed it with an SPDIF source from a source that is not on that clock network you're back to square one. So there...

[Summit]

22 hours ago, mansr said:

Only if it is transformer coupled or optical. Many implementations are not.

 

What do you mean? All good DDCs have some form of transformer coupling to isolate from the (often noisy) USB source. The $179 Schiit EITR for example has it, so it’s not only implemented in very expansive DDCs. The GI is and has always been one of the pros of a DDC or external renderers.

[mansr]

1 minute ago, Summit said:

What do you mean? All good DDCs have some form of transformer coupling to isolate from the (often noisy) USB source.

Maybe all the good ones do. I was also talking about S/PDIF sources in general, not just USB interfaces.

[Summit]

15 minutes ago, mansr said:

Maybe all the good ones do. I was also talking about S/PDIF sources in general, not just USB interfaces.

 

To use the S/PDIF or any other digital interface direct out of a standard computer is never good for SQ and IMO no good example when we are discussing pros and cons of different types of digital interface, on a site named Computer Audiophile.

[Summit]

29 minutes ago, miguelito said:

Completely agree on the diffs between design and implementation, and btw the vagueness of "this sounds better", especially comparing single ended vs balanced spdif.

 

Yes, the dCS clocks ain't cheap! I know first hand...  But even with a great clocked system, all in sync, if you feed it with an SPDIF source from a source that is not on that clock network you're back to square one. So there...

 

I wonder who is comparing the vagueness of "this sounds better", single ended vs balanced S/PDIF??

 

 

Ohhh it’s you!!

[mansr]

14 minutes ago, Summit said:

To use the S/PDIF or any other digital interface direct out of a standard computer is never good for SQ and IMO no good example when we are discussing pros and cons of different types of digital interface, on a site named Computer Audiophile.

What about the iFi Nano?

[Summit]

6 minutes ago, mansr said:

What about the iFi Nano?

 

It’s not easy to implant GI in a portable battery-powered device. It’s one of the limitations of battery-power and is well known.

[mansr]

7 minutes ago, Summit said:

It’s not easy to implant GI in a portable battery-powered device. It’s one of the limitations of battery-power and is well known.

Whatever your excuses, it's an "audiophile" USB to S/PDIF bridge without a transformer.

[Summit]

5 minutes ago, mansr said:

Whatever your excuses, it's an "audiophile" USB to S/PDIF bridge without a transformer.

 

Is it? Okay use it then I don’t care. It probably go well with the rest of your "hifi" gear.

[miguelito]

34 minutes ago, Summit said:

 

I wonder who is comparing the vagueness of "this sounds better", single ended vs balanced S/PDIF??

 

 

Ohhh it’s you!!

Dude: You need to learn to read! I made no statement whatsoever of which are better. I just said that for clock dissemination, single ended is just fine.

[Ralf11]

46 minutes ago, Summit said:

 

To use the S/PDIF or any other digital interface direct out of a standard computer is never good for SQ ...

 

why?