If you picture 1/3 cord, 4' high x 8' long x 16", filled with eight round 2' dia. rounds. Lots of air space.
Now picture 1/3 cord, filled with 32 round 12" dia. rounds. Lots of air space.
They are equal percentages of air to wood volume.
I did a real world test a few years ago, since buried in another thread of a forgotten name.
I filled two 1/3 cord racks as a cradle for 8' logs. 8' long x 8' wide x 4' high. Two cord of logs.
Cut into rounds and stacked in racks I got 1 1/2 cord or something close to that, a bit more if I remember right.
Split into stove wood size I gained volume going from rounds to splits, but not all that much. 1 2/3 cord approx.
Again, this is all from memory, but the gain in minimal from rounds to splits, maybe a sixth of a cord gain in two cord volume test. The loss from logs to rounds is far more, and logs to splits was a loss of approx. 1/3 cord.
This also check consistently with the twenty cord loads of logs I've been buying for six or more years now, approx. twenty loads.
I get about fifteen to sixteen cord of good wood, and a cord plus of junk wood (punky, hollow, feather weight airy crap, and good burning short cut-offs that don't stack)
If anyone knows the old thread I'd love to see it again, as I spend a bit of time playing with it. Curly Cherry did the dowels on the work bench thread years before that, and I always questioned if it translated to real world results.
Some university project broke stacked firewood volume down into three categories with percentages of wood, bark and air.
Again, going from memory, bark was approaching ten percent volume in a cord of stacked wood. That of course would be heavily dependent on species and size of rounds, and seems it would also factor in figuring charting the weights of a cord. Maybe the difference in bark for limb wood and trunk wood cancels out.
In the end, who cares.