I don't look at it like a higher air-to-fuel ratio per se so much as an improved mixing of the air and the fuel that is available. Many stove designs would pass the excess air too far away from the hottest part of the fire, providing no increased burn efficiency, but definitely creating a cooling effect on the fire and flue. With enough turbulence inside the firebox, however, the entire cavity becomes filled with a more evenly mixed mixture of gases and air. With a full load, the effect of turbulence isn't as pronounced, so there will be fuel-rich areas between the splits that just don't get enough oxygen in the primary burn area. Therefore, I believe that burning smaller amounts of wood promote more complete combustion, but it certainly is difficult to heat that way unless the smaller amounts of wood are being dropped into a burn pot in pellet form.
I thing the OP was right on when he mentioned the presence of the wood in the full load shielding the walls and glass from the direct effects of IR radiation. Still, if IR through the glass was largely responsible for this burst of felt heat, it fails to explains why the actual air temperature he is experiencing rises so quickly. IR radiation passing through air heats it up very little, so I think that convection may be a larger part of the picture in the case he is mentioning - both coming from the stove itself and coming from the other radiating surfaces in the room.
Which brings up and even more curious phenomenon I have experienced but can't quite explain. I can get my stove up to a certain temperature, but when I close the bypass I get an instantaneous wave of heat that really just hits me in the face. When I play the IR gun over all of the stove and pipe surfaces, I see no instantaneous rise in surface temperature (which, of course, is what I would expect). My feeling is that I must be feeling heat from the stove via convection vs IR radiation, but I still don't understand how this would occur without immediately increased surface temps.