I always thought the moaning about the lack of sample accurate editing in Logic was pretty ridiculous for 99% of all the people who moaned. (I'm only leaving that 1% out there in case I ever find someone who I actually agree NEEDS sample accurate editing. I expect I'll meet Sasquatch first.)...I think the range is actually kinda big (128 samples or something?)...
Again with the arithmetic: assuming one is recording at 44.1k, and assuming that the edit range is 128 samples, that means that your edit could be off by as much as 1/344th of a second, or just under 3ms. Assuming the 128 sample rate stays constant, (does it?), then at 96k the edit could be off by 1/750th of a second, or about 1.33 ms.
Mind you, that's the most the edit would be off. Plain old dumb luck says that half the time it would be 1.5ms or less at 44.1k, .65ms or less at 96k.
Maybe the moaners all work exclusively with musicians whose internal timing is somehow synced to to the US Naval Observatory's Atomic Clock, (or Joel Hamilton's), but I certainly don't. If I'm moving tracks around and concerned about phase, (like drums), I move them all together. Worrying about whether the bass player is as much as 3ms early or late coming into the bridge? I don't have the energy for that. And prior to the last 15 years or so, that concern wasn't even an option. I'm pretty sure I like one or two recordings that were made before then.
To me, it just seems like people who moan about it are way more wrapped up in the computer science of it than the reality of audio recording. No orchestra would meet the same accuracy standard for tempo that somebody recording a $250 drum kit with 6 month old heads thinks they need - nay, DEMAND - in their software. For their myspace page. I ain't havin' it.
/rant