[landroni]

Nobody needs real-time for simple audio playback.

"Simple audio playback" involving an external DAC requires by definition a real-time streaming of audio bits to the DAC. Every single time you listen to an audio file. The DAC receives each sample one at a time, (hopefully) at the expected moment, which is where timing errors (i.e. jitter) originate from. And that's why good clocking is so important and so much talked about. If external DACs were receiving all of the bits samples in batches (with buffering), say like when you load something in RAM, then jitter wouldn't exist as an issue in audio, and nobody would be talking about things like femto clocks.

 

[landroni]

  I have 7x the experience (35ish vs. 5ish years) design experience with DACs as opposed to USB interfaces.  Therefore, I am at a lesser point on the USB learning curve.  Since all of our DACs from Bimby up are upgradable, earlier adopters of our gear have been able to take advantage of USB upgrades as I have progressed along the learning curve.

Does this subtly imply that one USB implementation in a real-life product can perform better or worse than another USB implementation in a different product? And if this were true, then what are the chances that differing digital interfaces with very different underlying schemes would perform exactly the same in the context of D/A conversions? Mike may be keeping his cards close when it comes to the effects on sonic performance of a Wyrd, but in the past he hasn't shied away from suggesting that e.g. Coax SPDIF (RCA or BNC) would perform sonically better than USB...

 

[johnjen]

  mate this made me laugh so hard. there is so much irony for this to fall on me. 
snip

I'm glad you derived some humor from all of this.
 
JJ

 

[landroni]

 
 
So, why not put Wyrd-like USB devices inside DACs?   Because some of the advantages of these devices acrue to it not being electrically inside the PC and also not being electrically inside the DAC.

Exactly. This is also the reason why DACs got externalized in the first place (and got isolated in tank-like enclosures), to take them as far as possible from the noisy guts of a general purpose computer. Hell, Jason wouldn't even countenance putting a DAC and amp in the same box because of isolation concerns.
 
To quote @Torq who discussed some of these issues in this post:

Now, suppose you eliminate these interfaces entirely and play from a source directly housed in your conversion device. That might be a memory card, a RAM buffer (those are pretty much ALWAYS the final source) or a spinning disc ...

Guess what?

While you've eliminated the VISIBLE digital interface, there is still one present, and the eventual analog output is still potentially affected by any noise it generates. Which is one major reason why DACs got externalized from the transports/players in the first place!

 

[madwolfa]

  "Simple audio playback" involving an external DAC requires by definition a real-time streaming of audio bits to the DAC. Every single time you listen to an audio file. The DAC receives each sample one at a time, (hopefully) at the expected moment, which is where timing errors (i.e. jitter) originate from. And that's why good clocking is so important and so much talked about. If external DACs were receiving all of the bits samples in batches (with buffering), say like when you load something in RAM, then jitter wouldn't exist as an issue in audio, and nobody would be talking about things like femto clocks.

 
But DACs have buffers (albeit relatively small). And in USB asynchronous mode the clock (and incoming data rate) is controlled by DAC, not source. So the jitter is all but eliminated.
 
Please correct me if I'm wrong, Mike.

 

[madwolfa]

  I'm glad you derived some humor from all of this.
 
JJ

 
But he had a point - where is the line when it just gets silly?

 

[landroni]

  But DACs have buffers (albeit relatively small).

Care to back that up? (More than an one-liner please...)

 

[madwolfa]

Care to back that up? (More than an one-liner please...)

 
I can't find the source now, that's why I wanted Mike to chime in...
 
EDIT: my point about asynchronous USB still stands. any comments on that?

 

[jacal01]

Thanks -- but they're actually sure they can!

And...if you can't, then there's something wrong with (insert here - your ears, your equipment, all of the above, yada, yada)

 
And they actually can.  Cue the particular ringing in your ears...
 
Which is the cute endearing way that your inner ear hair cells ring the bell as they say "I'm outta here.", their time in service being generally determined of how much you abused them in the past.  My own little inner ear cacophony is induced no doubt in the main by my hanging around warbirds on the flight line in my youth, ± more than just a few rock concerts.
 
But I assume everyone's hearing experience is different, and those that sheltered their ears in their lifetime more from loud excessive noise probably currently enjoy the benefits of husbanding their stereocilia hair cells, still having most likely low audible sensitivity as well as extensions at both ends of the usual hearing spectrum.       

 

[castleofargh]

 

Nobody needs real-time for simple audio playback.

"Simple audio playback" involving an external DAC requires by definition a real-time streaming of audio bits to the DAC. Every single time you listen to an audio file. The DAC receives each sample one at a time, (hopefully) at the expected moment, which is where timing errors (i.e. jitter) originate from. And that's why good clocking is so important and so much talked about. If external DACs were receiving all of the bits samples in batches (with buffering), say like when you load something in RAM, then jitter wouldn't exist as an issue in audio, and nobody would be talking about things like femto clocks.

wouldn't that make streaming digital audio an analog thing?

I'm really a noob when it comes to all the usb protocols, but in my shameful past as a gamer, I seem to remember the default refresh rate for usb mouse to be 125hz and special drivers to try and reach 1khz that not all usb cards could really hold constantly.

 

[madwolfa]

wouldn't that make streaming digital audio an analog thing?

 
I think what he described was a legacy synchronous mode, where the clock was enslaved by the USB source and DAC couldn't control anything (=> jitter). Which is basically a non-issue now with async USB.

 

[KoshNaranek]

One would think that newer USB protocols would not have issues related to the data stream, but it appears to not be the case.

My own experience is that the OS makes a difference on the same hardware.

Mike has spoken earlier in this post regarding differences between hardware with the same OS. (Mac OS)

 

[Torq]

One would think that newer USB protocols would not have issues related to the data stream, but it appears to not be the case.

My own experience is that the OS makes a difference on the same hardware.

Mike has spoken earlier in this post regarding differences between hardware with the same OS. (Mac OS)

 
The last USB audio protocol was the 2.0 spec, which is about 15 years old now; it wasn't updated for USB 3.0 or 3.1.
 
In principal, as a pure streaming protocol it works well enough - the clock is where it should be, it has more than enough bandwidth, etc.  The problem is more to do with how the actual protocols are implemented on the computer (e.g. there are some timing issues with the latest version of OS X and on some Windows boards) and, more importantly, the electrical characteristics of the interface, and its signalling, at a physical level.
 
Power and data run together.  Power is often very noisy since the vast majority of source devices aren't audio-focused and don't have to worry about noise on the power lines upsetting analog outputs as they work purely in the digital domain.  Additionally, all of the data-centric (storage, peripheral interconnect, etc.) protocols that exists for USB are error-detected/corrected, where audio isn't.
 
A better solution, for USB 3.0, would have simply been to shoot the entire file down the wire, with ECC data, up front, and let the receiver buffer and play it.  That wouldn't fix all issues (you still have digital interfaces in the DAC then, but at least they can be optimized for audio purposes), but it would eliminate or ameliorate the issues of noise, isolation, consistency, power levels and so on from the interface.

 

[bigro]

Out of Curiosity. Has any one of the Audio Streamers /Music Servers Paid any attention to Improving the USB or SPDIF hardware possibly with a homegrown OS that specifically addresses the issues with Digital Audio That a do all, master of none OS like Windows or crappy USB Implementations can introduce? I have not done much digging as I just starting to consider getting one, but the reviews of the few I have read it seems Airplay, Bluetooth, Pandora, and whatever else they cram into this tiny box along with some random DAC and a bunch of things that are shiny is the norm. I have not seen any Mention of Any R/D dedicated to improving the method of delivering the audio data to your DAC.
 
We can now get some great external DACs for very reasonable prices and clearly a lot of time and effort has been put into making these products . It would be Nice if the point of origination would see the same type effort put into it. Instead we are running around the Windows, Linux, USB sucks VS. SPDIF Circle, Chasing our tails. Sure the DAC could have the USB Optimization H/W inside instead of an external box, which would be nice as there are less things to buy and connect, however to me it seems the Point of Origination of this possibly sub par Data Stream should be looked at with far more scrutiny. If there is a quality Data Stream to begin with then the need for said box would possibly be low to none, and the circle of OS variations and which flavors are better or fighting USB drivers would probably be minimized.
 
Back to my original question. Has any one run across a Music Server out there that is not just a bunch of stuff in a box, one that has seen some effort in addressing some of the issues seen with USB audio or OS quirks etc?

 

[johnjen]

   
But he had a point - where is the line when it just gets silly?

True enough.
But that is where each person makes that determination for themselves.
 
Some are willing to pursue their curiosity to amazing levels of minutia, while others not so much.
 
What I find 'interesting' is any attempt to denigrate anyone or any other line of inquiry in their pursue where their curiosity takes them.
Just because it makes no sense to one, does that mean it has no value or worth for another?
 
I personally want to encourage the spirit of curiosity and inquiry whatever it might be, and where ever it might lead, as these efforts can certainly lead to increased understanding, one way or another.
 
JJ