Do you understand what a baffle diffraction step is and why it's the source of your problem?
Have someone cup their hands around their mouth (like a bullhorn) and talk to you. Got louder didn't it?
That's because the same amount of sonic energy is now radiating in a smaller cone. If you were standing off to their side, it would have gotten quieter, not louder.
If they could talk in really low frequencies, this trick would not work. Why? Because the baffle (their hands) wouldn't be big enough for the sound to interact with it. There would be no change in volume from the front or the side from them holding their hands up... unless they grew much bigger hands.
Same thing happens on a speaker. High frequencies project almost entirely forward and so are 6db higher than low frequencies which project in all directions.
But when you embed a speaker in a bookshelf, or wall, or the ground: you make a bigger baffle. Suddenly all of the sound is being projected forward instead of only high frequencies. All frequencies become +6db relative to what the bottom frequencies were doing before.
There is no speaker that, without adjustment, will perform properly both embedded in a wall/bookshelf/ground and also sitting on a stand in a room. The only possible exception would be one that, by virtue of a really big baffle or really limited LF response was already functionally an infinite baffle. Physics doesn't allow it.
On the upside: DSP room correction may adjust this for you.
(and yes, you don't want to use a rear-port that's blocked. That would make the mass in the tube functionally heavier (and spl dependant) and ruin your sound.