Now - that makes NO sense to me...you have clock sync - what does "shrinking" accomplish...
Its all about the relative position of the files...
I think there may be some confusion here about the difference between two different things -
delay and
drift. Jim is referring to drift and you are referring to delay.
Delay refers to the the difference in the amount of time a distinct audio event/snapshot (for instance a snare hit) is captured by two distinct sources. A direct feed/SBD source will be 'immediate' realtively-speaking by comparison to a source captured by ambient mics, whether on stage or out in the room. It is the difference in time between the sound waves reaching the direct mics feeding the mixing console and the sound waves reaching your ambient pair - a difference that should approximate ~ 1 milisecond per foot of distance that is between the ambient mics and the sound source. When distinct sources like a SBD and an AUD feed are clock-sync'ed in a 4 track recorder or multitrack set-up, you will still have to correct for delay but will have no drift.
Drift, in contrast to delay, refers to the difference between two distinct recorders' internal clocks' rendering of time. Unless two distinct recorders are pro units with the ability to link by wordclock or by locking/syncing to the time information provided by a digital feed from some other 'master clock' device like an outboard A>D converter, etc., you will over the course of a recording have a slight difference between one recorder's rendering of, for instance, 1:25:36.97 and the other recorder's rendering of 1:25:36.97
Drift between the clocks of two separate digital recorders - even of the exact model/type - will be constant if they are not clock-synced in some way, and drift thus increases in a direct relationship with an increase in the length of recording time. So, when you find and line the source up at a sync point (like a snare hit) at the beginning of two distinct sources that were not clock-synced, and then you find a sync point at the end of the recording, you will find that one source is off from the other by some small amount that will make a mix of two disparate recordings lined up at the beginning fall out of sync by the end. You have to fix this
drift by time-scaling (essentially a time-shrink or expand process) one of the two sources to match it to the other. It may also be done, as some do in Wavelab, by chopping up one source and continually re-aligning it at multiple time intervals through the length of the recording, but that is both much more cumbersome AND much less accurate because there will still be some drift apparent (and possibly audible) at the end of each time interval before the following new sync point.