yep, he records stuff track by track, then when you listen to everything together, the two groups of tracks are "split".
in cubase you would move the audio tracks into a "folder track", add 12 ms predelay to them and you´re done. in max i am delaying the midi vs the audio if required.
your story about clock problems because of temperature changes is somehow interesting, but seems a bit odd.^^
I may not yet be totally understanding everything, but his description of "layering" voices that can be degraded by undesirable… call it "interphasing" where the overall results can be altered by millisecond variables is a different method than just stacking tracks. He's intermixing and combining waveforms to yield a desired sum. The precision needed to keep that under control
consistently seems like a different issue than everyday delay compensation and making sure everybody "starts together on the one" so to speak.
Internal computer clocks drift in teeny tiny amounts from all of the causes I mentioned. That's mainly because computer hardware - especially 20-year-old hardware was built to adequate standards for word processing, calc, etc. The build standards required for 90% of computer uses is not all that high. Then we came along and impressed music production on them. Then we forced them to do digital recording. Of course, worst of all, we started using them to look at kitty cats on the internet.
The less-than-perfect-accuracy produced by our hardware is not an issue 99% of the time. BUT, i can easily see a scenario where the system runs at 120.0001 BPM on Monday but at 120.0100 on Tuesday and maybe drifts a little to and fro on Wednesday. Now, synth voice you laid Monday and added another to on Tuesday are running ever-so-slightly out of sync on Friday – not enough to hear as out of sync, but maybe enough to slightly alter the combined sound of them played together so now you gotta futz with them.
If our stuff was all made by say, a co-venture of Raptor and Focusrite, the crystals that determine clock speed would have constant temp monitoring and control,
really high-resolution displays, light-speed CPU, busses and lots of other Swiss-type precision goodies. The computers would also cost as much as a small house and so they don't have that precision that 99.9% of users don't need anyway.
gert79's only problem is HE is that remaining .1%……If you trigger external gear and record all tracks in the same way (same buffer, same sample rate, same depth, same interface), a simple offset should be the only thing needed to apply to keep everything super tight. That being said, a "track stack" or layered sound of 3 or 4 really cool synth sounds should be layered as 1 stereo mix BEFORE the DAW and then recorded... use a mixer or if possible a multi-timbrel "Combo/performance" patch if you have that feature available (this avoids "phasing" on sounds that have a lot of movement and would be one solution); thus not dealing with trying to align the related tracks to sample accuracy in the DAW.
GAWD! Sample accuracy? THAT would be, what? 44,100 times worse??
Anyway, I'm impressed that gert79 clearly
knows what he wants and has a plan to get it by extracting every last bit of performance out of his "obsolete" Mac, DAW, interfaces and I was going to say synths, except
those, just like old guitars, are
never labeled as being "obsolete"… just "vintage".
Just like us…… 
