[460414]

From a technical point of view (not particularly interested in potentially biased human hearing test unless its ABX/blind) which is superior out of: coaxial, USB, or optical?
AFAIK: USB is the only one which can do 32bit, but until we have true 32bit or even full 24bit recordings and DACs which will actually utilize the 24/32bits; I don't see that as relevant.
 
Essentially it comes down to which one has the lowest jitter? And the implementation?
 
 
 
For my own information:
I have desktop PC which probably instead ideal for noise as its geared towards high performance (Ex TOTL CPU, GPU, Mobo)
I also have an ASUS STX II which can output SPDIF coaxial.
Which one should I use?
 
Also which bit-rate and sampling rate should I use in windows? Running 24 bit with 16 bit "recordings" (sine waves for tests) shows the noise floor is lowered by over 10db (Cant see due to accuracy of measuring device: Rohde and Schwarz Upv). 

 

[spruce music]

Lowest jitter should be USB if it is an asynchronous USB.  This places timing of the bit stream at the DAC using its own crystal clock.  Optical and coax must get timing from the external stream they are fed to some extent and this is less accurate.  Also some affordable DACs do now have 32 bit DAC chips and accept 32 bit inputs.  Results are still limited by thermal noise to around the 20 bit level.
 
As for sample rate use whatever rate is native to the recording, for bits nothing wrong with running 24 bit for everything.

 

[Roseval]

From a technical point of view one can only say that there is no free lunch.
 
SPDIF over optical provides excellent galvanic isolation but is said to be down in the jitter department.
SPDIF over coax doesn’t provide galvanic isolation but in general has a improved jitter performance compared with optical.
http://www.thewelltemperedcomputer.com/Intro/SQ/Toslink_Coax.htm
 
 
USB if used with asynchronous synchronization has zero input jitter.
The lack of galvanic isolation and the complexity of the protocol are the downsides.
 http://www.thewelltemperedcomputer.com/Intro/SQ/USB_SPDIF.htm

 

[watchnerd]

  From a technical point of view (not particularly interested in potentially biased human hearing test unless its ABX/blind) which is superior out of: coaxial, USB, or optical?
AFAIK: USB is the only one which can do 32bit, but until we have true 32bit or even full 24bit recordings and DACs which will actually utilize the 24/32bits; I don't see that as relevant.
 
Essentially it comes down to which one has the lowest jitter? And the implementation?
 
 
 
For my own information:
I have desktop PC which probably instead ideal for noise as its geared towards high performance (Ex TOTL CPU, GPU, Mobo)
I also have an ASUS STX II which can output SPDIF coaxial.
Which one should I use?
 
Also which bit-rate and sampling rate should I use in windows? Running 24 bit with 16 bit "recordings" (sine waves for tests) shows the noise floor is lowered by over 10db (Cant see due to accuracy of measuring device: Rohde and Schwarz Upv). 

 
This is all legacy / old tech.
 
We should be asking about USB-C and Thunderbolt 3.

 

[460414]

   
This is all legacy / old tech.
 
We should be asking about USB-C and Thunderbolt 3.

Any reasonable DACs which support that yet?

 

[watchnerd]

  Any reasonable DACs which support that yet?

 
I'm not sure what the criteria for "reasonable" is, but Focusrite and UAC have models now.  I would expect most interface makers to come out with updated models over the next 6 months.

 

[spruce music]

USB 2.0 isn't going to disappear overnight.  USB3 is compatible with it for one thing.  As USB 2.0 has plenty of capability for digital audio and is already widely available on DACs the OP's question seems just fine.  I fail to see any potential advantage that USB 3 or Thunderbolt will offer in regards to audio sound quality over USB 2.0. 

 

[460414]

  USB 2.0 isn't going to disappear overnight.  USB3 is compatible with it for one thing.  As USB 2.0 has plenty of capability for digital audio and is already widely available on DACs the OP's question seems just fine.  I fail to see any potential advantage that USB 3 or Thunderbolt will offer in regards to audio sound quality over USB 2.0. 

 
^
@watchnerd
 
Bandwidth isn't an issue? 32/384 is more than enough?

 

[watchnerd]

   
^
@watchnerd
 
Bandwidth isn't an issue? 32/384 is more than enough?

 
Bandwidth isn't an issue for playback.
 
But....
 
If you're doing production work, USB-C has much much less latency.  That matters a lot for people using DAWs.
 
Gear advancements often start on the pro side, which trickles over to the audiophile consumer side a few years later, often at much higher cost.