Measuring jitter of audio interface

[LionIT]

I would like to measure the jitter in the analog domain of my audio interface when using the Toslink signal as master clock (therefore the jitter induced by the source and propagation).
To do this I have to loopback on the AD input of the interface itself, which works with the same master clock.
The doubt that I cannot dissolve is whether the fact of sharing the same clock, and therefore the same jitter, false the measure in some way, canceling or doubling the level of jitter in the result.
I cannot rationally understand if there is any wrong mechanism in this chain of measurement therefore.
Could anyone clarify?
Thanks

 

[amirm]

You will be fine as jitter manifests itself far more IN DAC and at any rate, it is often not the clock itself that is the problem.

 

[AnalogSteph]

The doubt that I cannot dissolve is whether the fact of sharing the same clock, and therefore the same jitter, false the measure in some way, canceling or doubling the level of jitter in the result.

Your concerns are definitely warranted. In a loopback test, jitter will tend to cancel, especially the lower-frequency components. If you want to do this accurately, a clean external signal source or ADC will be needed, depending on direction.

 

[LionIT]

Well, two slightly contrasting opinion.

I took two meas to see the results.

Test file is the usual J-Test at 48kHz created with REW and played with my streamer, connected with Toslink to my audio interface (clock resolved to optical).
REW RTA is set to 64k Blackman-Harris 7.

This is the spectrum of the Toslink digital input.

And here the Toslink input routed to analog out (21dBu lev) and acquired with the analog input of the same interface (0dBFS are about 22dBu).

Honestly, I dont know how to judge these results in terms of jitter representativeness with respect to the measurement chain, for the doubt mentioned initially.