Some conjecture on this gadget thing.
Consider four performances with Chris Thile playing mandoline.
1. Live, A=440hz
2. Recorded and reproduced, A=440hz
3. Live, C=256hz
4. Recorded and reproduced, C=256hz
They may all be very nice, but I'm going to say the two live performances are going to emotionally best the recorded, and the two recorded performances will be mostly indifferentiable to each other.
Let's do the same again but bring in Edgar Meyer on bass to lay down some foundational tones over which Thile's mandoline sings.
Now we've got honest to goodness skingasms during both live performances, but something is amiss on the recordings or live with any lessor bassist who does not precisely intonate. Or with instruments that aren't incredibly well crafted which we can reasonably assume these two possess.
Bear in mind the emotional weight of the minor third is all about sound waves being in the exact ratio of 6/5. A fifth only sounds powerful with frequency ratios of 3/2.
I wager tightening up the mathematical relations of all the frequencies reproduced is exactly what the gadget is intended to do, and what tuning it defaults to is arbitrary or entirely for convenience, and this would be done because time or phase alignment is lost during the recording process.
edit:
Although I could very well be wrong <-link.