S/PDIF Toslink or Coax? Computer => DAC

[Lupin]

Thanks for all the feedback.

After reading all the replies I think I will go for optical Toslink.
Difference in SQ will be minimal, if any at all, and it provides a much cleaner cable manageable solution.

Apparently the re-clocking in the Singxer is not a big deal

 

[mansr]

In general Toslink generates more input jitter than coax.

Let's do an experiment. If we connect a reasonably low-jitter S/PDIF source to a receiver, we can measure the jitter of the recovered clock. For this test, I used an iFi Nano as the source, connected via a cheap dual-output converter to an even cheaper dual-input DAC. Triggering a scope on the source clock while displaying an edge of the recovered clock, the amount of jitter is readily observed.

This is the result when using a 0.5 m coax cable:

The histogram shows the frequency with which the waveform crosses midpoint at various times. As we can see, the distribution is more or less Gaussian with about 5 ns peak-to-peak.

Switching to a 1 m optical connection (cheap plastic), we get this scope image:

There is no appreciable difference to be seen. Clearly, the claim that Toslink gives more jitter than coax is not universally true. Maybe there exists a DAC that does behave that way. Before making a general statement, we'd need to measure a large number of DACs. Until someone does this, all we have is a claim without substance.

 

[Michael1977]

Let's do an experiment. If we connect a reasonably low-jitter S/PDIF source to a receiver, we can measure the jitter of the recovered clock. For this test, I used an iFi Nano as the source, connected via a cheap dual-output converter to an even cheaper dual-input DAC. Triggering a scope on the source clock while displaying an edge of the recovered clock, the amount of jitter is readily observed.

This is the result when using a 0.5 m coax cable:
View attachment 129393

The histogram shows the frequency with which the waveform crosses midpoint at various times. As we can see, the distribution is more or less Gaussian with about 5 ns peak-to-peak.

Switching to a 1 m optical connection (cheap plastic), we get this scope image:
View attachment 129397

There is no appreciable difference to be seen. Clearly, the claim that Toslink gives more jitter than coax is not universally true. Maybe there exists a DAC that does behave that way. Before making a general statement, we'd need to measure a large number of DACs. Until someone does this, all we have is a claim without substance.

The Wikipedia claim of increased jitter over S/PDIF Toslink seems to be a bit out of date as it was a more suitable explanation of the phenomenon in legacy D/A converters which weren't as good at clock recovery as modern circuits.
The case to be made for Toslink lies in its electrical isolation which virtually eliminates noise-causing interference.

 

[Vincent Kars]

Clearly, the claim that Toslink gives more jitter than coax is not universally true.

Fair enough.
To the best of my knowledge modern DACs use asynchronous sample rate conversion. This is very effective in eliminating input jitter. So my question is does your measurements show the difference between coax and optical or does it show that modern DACs are very effective in reducing input jitter?

Anyway, in 2010 you could find results like these: https://www.thewelltemperedcomputer.com/Intro/SQ/Toslink_Coax.htm showing coax to perform better in the jitter department.
Today you can find results like these: https://www.audiosciencereview.com/forum/index.php?threads/wiim-mini-review-streamer.33099/ showing todays Toslink is an excellent performer.

 

[maty]

SOUND DEMO (very good sound recording from loudspeakers). Try to watch it in 4K quality (AAC 192 kbps).

Digital Cables Compared: Usb Vs Coaxial Vs Optical - Choose Wisely!

 

[Rednaxela]

Not planning to, but suppose I gave it a listen and I would hear differences, should I then not be very, very worried?