[johnjen]

snip  
As for Yggy burn-in, I strongly recommend new owners listen and enjoy the journey rather than just run it continuously and check in after 150 hours.  The changes it undergoes are quite fascinating and made for a really interesting couple of weeks worth of listening.  I'd liken it to an impulse response wave dying out, as time goes on the swings in sound get milder until one day the line goes flat and it just sounds right.

I heartily agree with this.
 
I always enjoy 'tracking' the changes as the equipment 'settles in'.
 
JJ

 

[Currawong]

I'm wondering if the USB input is improving with use, as I don't notice as much difference now with other transports. It could be me though.

 

[landroni]

 
It depends on the USB source, I use an iFi USBPower and it cleaned up some of the sibilance when use with the Raspberry Pi

 
On my Mac Mini it made no difference, went over to the RPi as the SQ was better.

 
Was this vanilla RPi, or RPi with Digi+ boards? 
 

 

[b0bb]

   
Was this vanilla RPi, or RPi with Digi+ boards? 
 

No RPi->Yggy via usb not, using the Digi+ board
 
I used the RPI as the mediaplayer instead of the MacMini + Audirvana

 

[MattTCG]

  I'm wondering if the USB input is improving with use, as I don't notice as much difference now with other transports. It could be me though.

What kind of transports are you using? Just curious. 

 

[Jamiee]

  I'm wondering if the USB input is improving with use, as I don't notice as much difference now with other transports. It could be me though.

 
Nope, The USB definitely improves with age - like the rest of the Yggy.
 
That said, I still prefer the AES/EBU input (with COAX - S/PDIF coming in a very close 2nd) when paired with a Stello U3 USB converter.
 
My digital chain (so far this is the best my system has sounded):
Win10 PRO Desktop PC - Jplay v6.2 - Fidelizer v7.1 - ProcessLasso v8.9 > J-CAT Battery PSU (feeds OS SSD) > J-CAT USB CARD > 2.0m Curious USB Cable > IFI iPurifier > USB Regen (amber) with Teradak DC-30W > Curious Regen Link cable > April Music Stello U3 > MIT Digital Proline AES Cable > Schiit Yggdrasil.

 

[DACLadder]

  I'm wondering if the USB input is improving with use, as I don't notice as much difference now with other transports. It could be me though.

Perhaps.  You got me to try AES/EBU input and I like it very much.  Seems to have a little more lower range energy while retaining the clear, open midrange and smooth highs.  Slight differences to the direct USB input though. 
 
I also turned off my Yggy for a week to try another DAC.  It had been previously powered on continually for 6 months.  Powering back up cold sounded OK but does sound better after a few days of warming back up.  As I suspected the initial long period of sound changing seems to be just a break-in component.          

 

[Currawong]

 

  I'm wondering if the USB input is improving with use, as I don't notice as much difference now with other transports. It could be me though.

What kind of transports are you using? Just curious. 

 
At the moment I have an iUSB 3.0 feeding both the Gen. 3 USB and an Audiophilleo 1/Pure Power which I can select between. I am also still using the Soundaware server via AES.  I guess if things continue as they have been, it may end up being a similar situation as I had with the Hugo, where I used just a Schiit Wyrd instead of bothering with the whole Audiophilleo set-up.

 

[SeanS]

  I'm wondering because these things seem like easy to test with access to the right equipment, checking input data with packages received at each point in the chain at each point in time... I mean, if the "bits are bits" hypothesis doesn't hold in real-time streaming, there has to be a mechanism for differing perceived differences from something that nominally should unmistakenly be 0s and 1s.

 
Bits are bits.  But we're not talking about the bits, error correction takes care of bits that get lost, etc.  What we're talking about is noise that rides into the "analog" output circuitry of the DAC.  That's why we talk about "galvanic isolation" with USB.  The noise can come from many places in the circuitry from the source to the DAC, but end result is that it ends up getting into the "analog" output signal and we hear the effects.

 

[Torq]

   
Bits are bits.  But we're not talking about the bits, error correction takes care of bits that get lost, etc.  What we're talking about is noise that rides into the "analog" output circuitry of the DAC.  That's why we talk about "galvanic isolation" with USB.  The noise can come from many places in the circuitry from the source to the DAC, but end result is that it ends up getting into the "analog" output signal and we hear the effects.

There is no error correction on the commonly implemented audio interfaces.  S/PDIF (whether optical or electrical) and USB Audio do not provide for error correction at all, though the HDMI audio specification does include ECC support.
 
If a bit gets flipped, which given the general 0/5v signaling standards in use is rather unlikely (that'd take a lot of cable attenuation or some serious interference ... like wrapping the interconnect around an alternating/pusling power source), the DAC will either output the corrupted data as-is, which may well be inaudible since it will be, at most, 1/22,000th of a second of audio, cause a glitch or get smoothed out (successive approximation schemes are affected by the previous data item, for example).
 
However, such occurrences should be very infrequent unless dealing with a marginal connection in the first place and your point that it is noise, carried or induced, that gets into the analog stages of the DAC or its buffering/output, is the real issue.

 

[earnmyturns]

 
There is no error correction on the commonly implemented audio interfaces.  S/PDIF (whether optical or electrical) and USB Audio do not provide for error correction at all, though the HDMI audio specification does include ECC support.

Bits are encoded as waveforms. Depending on source clock accuracy and waveform distortion, there will some variability on when each bit is registered at the destination -- jitter. That will affect how the analog signal is reconstructed, so good DACs try to reclock the bits.

 

[Torq]

Bits are encoded as waveforms. Depending on source clock accuracy and waveform distortion, there will some variability on when each bit is registered at the destination -- jitter. That will affect how the analog signal is reconstructed, so good DACs try to reclock the bits.

That deals with jitter. Which is a time-based error.

ECC is generally concerned with errors in the data, not its clocking.

Correcting a flipped bit requires a means to detect it (e.g. a checksum) and redundant data (or a retransmission-capable protocol). S/PDIF and USB Audio have neither, so there no possibility of deterministic error correction. HDMI Audio includes both - so as long as there are enough parity bits to correct the number of bit errors in your stream, you're golden.

Flipped bits aren't very common since the signaling voltage differential between 0 and 1 is large enough to require some pretty gross interference or attenuation to induce.

Most DACs do some level of jitter correction. No non-HDMI consumer DACs do error correction on the data stream as they simply don't have a way to detect it in the first place.

 

[earnmyturns]

That deals with jitter. Which is a time-based error.

ECC is generally concerned with errors in the data, not its clocking.

My point is, why worry about bit errors and ECC, which as you said are very unlikely, while jitter as well as electrical analog noise are much more prevalent?

 

[Torq]

My point is, why worry about bit errors and ECC, which as you said are very unlikely, while jitter as well as electrical analog noise are much more prevalent?

I'm not worrying about bit errors. I'm just interested in seeing the discussion properly framed.

I'm also an engineer ... which means that I'd default to using a transport mechanism that didn't suffer from two sets of issues that have to be addressed (e.g. noise and jitter with USB) when a cheap and simple alternative standard exists that only has to deal with one of them (e.g. jitter with an optical connection).

And using a USB output just to have to first fix USB's issues and then deal with any issue endemic to what is ultimately being converted to seems like an exercise in creating problems for the sake of it. It only makes sense if you don't have other choices.

Hopefully the next USB Audio standard won't be limited to just a streaming protocol. The interface is already fast enough to just transfer the whole source file to the DAC in the same fashion as a file to a hard drive. No timing or data integrity concerns ... the DAC just buffers the entire source file (or large sequenced chunks of it) and then plays from its internal buffer.

 

[landroni]

 

simple alternative standard exists that only has to deal with one of them (e.g. jitter with an optical connection).
 

Isn't Coax with galvanic isolation also free of noise issues? Or will it depend on specific implementations, i.e. how much noise is being picked up and passed downstream the audio chain?