Odd MTC Sync Problem (Urgent!)

I wonder if anyone has any idea what is going on, here…

For my own arcane reasons I need to sync Logic and Nuendo. MTC via an IAC bus is the obvious choice.

When I press play in Logic, Nuendo joins in but with an offset. If I press stop in Logic, then Nuendo stops. Press play again, and Nuendo joins in again but picking up where it left off, plus an additional delay. Basically, Nuendo is syncing to MTC + [unidentified running clock time]. Very, very odd.

If I use a third-party timecode app (Horae) it reads Logic’s output MTC fine, and Nuendo syncs perfectly to Horae with no offsets (and no offsets are set in the apps). If I set up two IAC busses and sync via Horae then, also, it works. But direct IAC bus between the two apps - weird offset.

Any thoughts?



What is the sample clock source for both applications ?

The difference between a timecode reading application and a DAW is that the TC app will follow the incoming TC independently from a sample clock.

A DAW can jump to a TC position (or MTC or SPP or Serial P2) but after this jump it must release the incoming TC and follow the hardware sample clock delivered by the audio hardware. If for some reasons this clock does not run exactly at the same rate compared to the incoming TC clock then an offset will start to grow.

The idea behind this is : if your master application generate a MTC that is clocked to a different sample clock compared to the sample clock of the second application, then this second application will not be able to synchronize to the incoming MTC because of out of sync clock rates (as a side note for most DAWs and audio hardware cards clock rates cannot be adjusted in realtime, eventually some DAWs can resample to adapt incoming and output sample rates, but this is consuming CPU power and deteriorate sound quality).

To my knowledge, Nuendo cannot resample to follow an external TC, but he can follow it using an external synchronizer that will recreate a word clock slaved from the incoming TC. This mean as well that you’ll need in this case some audio hardware with a word clock input. Anyway this is a bad solution because MTC is far from being a jitter free solution and its delay is not exactly known. So there is no guaranty that your slave machine or software will be frame and even less sample aligned even if both sample clock frequencies are well matched.

A better solution would be to give the same sample clock to both audio hardware through a word clock common reference. And the best for a more precise synchronization would be to use something like Rewire (for internal connections) or Video ref + TC or VST system link (for external connections) where timings are well known, jitter free and compensated. The advantage of VST system link is that it does use an audio digital output for synchronizing, so that there is no need to make manual delay measurements and adjustments.

Both on internal sync. But the thing is, Nuendo is not jumping to the address that Logic is sending. It’s jumping to the address, plus an offset which seems to be related to system up-time. It’s very odd.

In other words:

(1) The computer has been running for 1 hour. Logic starts playing from address 01:00:00:00. Nuendo jumps to 02:00:00:00 and starts playing.

(2) The computer has been running for 1 hour, 5 minutes and 27 seconds. Logic starts playing from address 01:00:00:00. Nuendo jumps to 02:05:27:00 and starts playing.

I am, needless to say, baffled.


Yes this is strange Nuendo should jump to the right location each time and eventually drift if the “internal” clocks do not have the same source.

Seems like there is some kind of Jam sync or something that is silently adding this running offset between Logic and Nuendo. Perhaps you could try to use another virtual midi port utility to link the two softwares, actually you are using the default system one if i’m right (i’m not used to the MAC world).

Or you could try to use 9pin sync using a virtual serial port, if Logic can output in this format.