Be aware that it can to be somewhat more of a challenge to split a single stereo pair across two recorders than it is to record two separate sources that will be mixed together to separate recorders. At least some stereo pairs where maintaining tight phase correlation between the two channels is more critical, such as a coincident pair. Should be less of an issue with a wide-spaced pair.
Yea I wondered about the lack of precision when lining up independent sources in post. Does it affect critical time arrival differences? I'm sure it does. [snip..]
Somewhat. I think the issue will be if the error in precision falls within our perceptual tolerance for error. When aligning the two channels of a coincident pair and you zoom in, the peaks of both channels will/should always align, whereas with a split pair it's easy to see the time offset of peaks arriving from sources that are off-center. Sometimes you just need to split the difference when aligning a split pair with SBD. Additionally, perfect alignment is somewhat less critical when aligning two separate stereo sources, than when aligning the two channels of a single stereo pair.
On that last point but even more OT, there is a technique proposed maybe 20 years ago for mixing in spot mics with a main mic pair, which might be applicable to mixing in SBD with AUD, since a SBD feed acts essentially as a "collective spot mics" stem. It seeks to better preserve the imaging and depth cues of the main/AUD pair by delaying the spot/SBD signal by a few different randomized milliseconds or factions of, and pans those slightly delayed signals around to random static locations so they act like early reflections. The main pair/AUD source then retains unmolested first arrival status preserving its image cues, with the pseudo early reflections provide clarity and "direct-soundiness". Never tried that, but its an interesting approach modeled on the behavior of real world direct/early/late arrival acoustics.