Drift correction confusion

[dotnet]

I have a question for the Mac audio buffs about drift correction with aggregate audio devices in macOS (Audio Midi Setup). The Apple documentation I found on this gives conflicting information.

https://support.apple.com/en-us/HT202000 states:

To set the clock source for the Aggregate Device, choose the device from the Clock Source menu. Choose the device with the most reliable clock.
For each device that is not the clock source, select Drift Correction.

In https://support.apple.com/en-au/guide/audio-midi-setup/ams094c7edb4/mac, it says:

If all of the devices in the aggregate device aren’t synchronised through hardware, click the Clock Source pop-up menu, then choose the device with the most reliable clock. Select the Drift Correction tick box for each device.

So, should drift correction be enabled for each device, or only devices other than the clock source?

If I’m using two USB DACs as output devices and combine them as an aggregate device, what should I listen for to determine whether drift correction is set properly? Is it obvious/audible?

 

[MajorFubar]

I too found the advice confusing and tend to just tick all boxes. If I don't then it always seems that one of my aggregated interfaces experiences more PB latency compared to the others

 

[dotnet]

I too found the advice confusing and tend to just tick all boxes. If I don't then it always seems that one of my aggregated interfaces experiences more PB latency compared to the others

How do you detect that? Is there something to listen for?

I only have the slave checkbox checked and I do get occasional (perhaps once every 20-30 minutes or so) drop-outs, short blips really, that at first I blamed on Qobuz. Then I realised they happen with local media playback, too. I wonder whether that could be a symptom, i.e. the time drift eventually exceeding the buffer size.

I’ll run for a while with all boxes ticked and see how that goes.

 

[MajorFubar]

How do you detect that? Is there something to listen for?

Yeah, just simply that the output from one is obviously out of sync with the other (ie fractionally later). I use multiple aggregated audio interfaces in my home studio, one of which sends a headphone mix to a vocalist in a different room. It's obvious when that particular interface is dragging behind because I can hear on my headphones that the vocalist's timing is late, which always needs to be corrected in post.

 

[dotnet]

Yeah, just simply that the output from one is obviously out of sync with the other (ie fractionally later).

I see. This would be rather more difficult to hear in my case, where one DAC does the L+R speakers and the other the subwoofers. The subs are crossed over at 80Hz.

I’ve checked all boxes now and will watch for drop-outs anyway.

 

[BrianBaughn]

I'd believe it if someone said clock drift could cause phase issues between the speakers from one interface and the subwoofers from a different interface.