[sdiy] "sample accurate light swing"

[Richie Burnett]

> I think it more likely it's just a consequence of the buffer size of the
> "ACB" sound generation.
> If it was done on purpose, they should have gone with 2ms.

It is exactly 2ms of variation on the TR-8.  But, as you said, it drops out 
naturally in the modelling if you choose a 2ms block size for the digital 
audio synthesis.  For the TR-8 with a 120BPM four-to-the-floor kick pattern 
the onset times are like: 0ms, 499 ms, 1001ms, 1501ms, 1999ms, 2499ms, 
3001ms, 3499ms, 4001ms, 4499ms, etc...

> The human auditory system uses inter-aural delay as a source of spatial...

Agreed.  +/- 1ms in absolute timing for the downbeat of all instruments that 
are active on that step is probably not perceptible at normal tempos.  Maybe 
with the exception of "machine-gun" drum fills.  But +/- 1ms variations in 
the onset of one instrument *relative to another* would definitely be 
audible.

-Richie,