גם S/PDIF טוב יותר, גם אם זה לא נובע ממבחן זה ישירות
http://db.audioasylum.com/cgi/m.mpl?forum=digital&n=148562
Both the SB3 and touch have separate crystals for 44.1 and 48 families. The ones in the Touch are little lower in jitter but not by a very large amount.
Both the SB3 and Touch have a "clock mux", this chooses which crystal is used for generating the rate of the output stream.
In both the Touch and SB3 the signals go through a "connector board" which contains the external world jacks.
Now to the differences:
SB3: the clock muxing takes place in the FPGA, the S/PDIF generation takes place in the FPGA. It goes from the FPGA directly to the 75ohm line driver then on to the connector board. On the connector board the trace widths and spacing are not optimized for proper 75ohm impedance. The EMI suppression network does strange things to the impedance of the signal driving the line, the result is a major increase in reflections on the transmission line to the DAC.
Touch: the clock muxing is handled by a separate low jitter mux which drives a low jitter flop which is used to reclock the S/PDIF data stream. The 75ohm line driver is a better design than the one in the SB3 giving lower jitter performance. There is no EMI suppression network as in the SB3. The trace widths and spacings on the connector board are much closer to producing 75ohms than those on the SB3.
The net result is a significantly lower jitter S/PDIF stream with much lower reflections than on the SB3.
I have not done much testing with the TOSLINK out. You should still get the advantages of the lower jitter S/PDIF stream to begin with. The power supply regulation and decoupling in the Touch is better than the SB3 so that might help the TOSLINK. My guess is that the differences in TOSLINK between SB3 and Touch are not going to be as great as the differences in coax.
The designers of the Touch seem to have put a fair amount of effort into optimizing the S/PDIF output. The SB3 did not seem to get that much design effort on the S/PDIF.