Steve...where does jitter occur?
Jitter starts at the master clock in the transport or async USB interface. It then just gets worse as it passes through buffers, translators and cables. The DAC receiver may add more or reduce it a bit. The D/A chip generally adds more.
Is it more likely to come from the transport/music server or from the DAC?
Both have contributions. The worst will be the transport or device that contains the master clock.
Can it be eliminated completely and how would that be done?
Cannot be eliminated completely, but it can be reduced to the point where other system noise and distortion is more prevalent. This is not easy though. One must use very low noise fast-responding power supplies for the master clock and the associated circuitry. This results in tight voltage regulation. Likewise, the D/A chip must have very low noise fast-responding regulation on the supply voltages. DACs without upsampling and with good voltage regulation can delivery a very low jitter result. You generally get what you pay for here.
Alfe - the jitter solution is actually fixed with async USB, but it is highly dependent on the implementation. Most are poorly implemented. There are however other pitfalls that are more difficult to overcome, such as the effect of the playback software and audio stack. Most of these disappear when one uses a network-driven solution such as Sonos or SB, but hi-res is the limitation of these right now.
Cerrot - you continuously complain about the SQ of cheap USB converters and poor designs. This is like buying a cheap CD transport and concluding that all transports are garbage.
There are some high-end reviewers such as JA and Steven Stone that use USB converters for their digital references, and for good reason. They outperform everything else.
Steve N.
Empirical Audio