Tone Deft wrote:you're talking at the sample base level, so 1/48kHz = ~21uS, 96kHz is half that, 192 is 1/4 that. midi is 33kBaud IIRC, a period of 30uS. with midi it's theoretically impossible to get perfect sync at 48kHz without doing a lot of averaging and operations that would slow down the machine's reaction time. yet people have been using midi sync for decades.
with true sync I could invert one device's metronome sounds and have them cancel. or not cancel them but have the combined metronome sounds double in volume.
there's a huge difference between the 2-3bpm wiggles that the OP is seeing, the uSecond sync I'm talking about. that would be an incredible improvement, very very noticeable. that you're referring to is much more subtle and overkill for most users. it's a sexy idea but I'm not worried about. a Pro like you goes nuts about it. word clock is a more suitable route to take.
forget it toni.. such a beat wobble we have in live is doing several ms of unacuarcy..
I studi clock sync problems since they started to get problematic from the mid 90´s and had discussions with many developers also some that do the real sync stuff for tv stations incl.
I even influenced some products that try to deal with the issue..
On an apple you should get a stable tempo reading..reaktor can do it and fire samples precisly.. you can patch an ableton alike mini app in reaktor that is just running tight on an external midi clock..
so when i can do it within reaktor it must be possible for a machine code written app aswell... YOu just need to have that as a goal on your agenda.. Roger linn was able to do it perfectly on the old mpc´s..the later akai developers wasnt anymore.. Either because they was unableton or precission was just not on theire agenda anymore.