Topic: Best Syncing: MADI+ADAT or all MADI or Wordclock?

System:

UFX+ (master)
UFX (adat synced)
M16 AD Pro (MADI synced)
M16 DA (MADI synced)

I've noticed some weird lags between units and wonder if the ADAT and MADI sync (or jitter) at rates different enough to cause issues? Would using wordclock in this set up instead of ADAT/MADI be better? What if it was all MADI? (About to do an upgrade so all units will have MADI)

2 (edited by ramses 2022-05-17 19:16:46)

Re: Best Syncing: MADI+ADAT or all MADI or Wordclock?

I think its important to understand 1st what problems you mean.

> I've noticed some weird lags between units

What lag between do you mean between which units? What is the signal flow / use case?

You can use the RTL utility for measuring so that it's possible to talk about real numbers:
https://oblique-audio.com/rtl-utility.php

Besides this my understanding is, that (full) latency compensation by the ASIO driver is only possible for analog ports, where the converter latency is known to the driver.

Digital I/O through ADAT and MADI has only a very small latency of only a few samples if I remember right, so should not be an issue. It could be the case that you have a little offset between e.g.:
- AD from UFX+
- AD from UFX through ADAT, from M16 AD through MADI, as the driver doesn't know what you connect and thus doesn't know converter latency

Regarding clock quality/jitter: UFX, M16 AD and M16 D/A support Steadyclock, thus they should have no problem with any clock source (in this case UFX+), refresh the clock and use internal clock for final conversion.

I personally would have chosen the same setup: UFX+ as master (the directly connected recording interface), so that the clock rate can be learned from the application (DAW) and be distributed further to all clock slaves.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub14

3 (edited by explodingPSYCH 2022-05-17 19:41:54)

Re: Best Syncing: MADI+ADAT or all MADI or Wordclock?

My tests doing various loopback combinations yielded variance in latency that I wasn't expecting. For example sending audio from my DAW through a UFX+ output and back in through either of the slave devices gives 13 samples of latency (adat) but the MADI is 2 samples early. However, if I send audio from the DAW through one of the slave devices outputs and back through the UFX+ my both are around 40 samples late.

Nothing to cry about, but I've had some strange issues creep up when doing more complex parallel routing in Totalmix. The 32 samples of latency seems to stack up depending on routing. I also have quick measurements I did when I first noticed the issue that were more like 76 samples.

I'll use the tool you linked to get better results. In the meantime I'm trying to understand how to set up the most uniform response I can out of the devices. Could AD/DA models be causing issues?

(I'm upgrading the UFX to a UFX+ and adding a 1610, so that should leave the DA M16 [non-pro] as the only last gen model [I think]. I'd still like to keep that in uniform sync if possible though.)

EDIT: Here are the results of loopback tests using RTL Utility:
Buffer 512
(values in order: Measured RTL, Return, Correlation)
UFX+ > UFX+: 1106, -6.2, 24.4
UFX+ > UFX  : 1143, -6.2, 24.5
UFX+ > M32  : 1148, -6.2, 24.5
UFX > UFX+  : 1118, -6.5, 23.5
M32 > UFX+  : 1102, -5.8, 25.6