Think about the gas spring taking the place of the crank and rods in a normal engine. In a normal engine, a piece of the "excess energy" from a power stroke gets recycled into the compression stroke of the next cycle, and the rotating inertia of the flywheel handles this very efficiently over a reasonably wide range of operating speeds, as long as the flywheel is heavy enough. If the engine is driving an alternator to generate power, that alternator doesn't have to handle any of the internal "recycling" of mechanical energy from one power stroke into the next compression stroke. And that is a good thing, because the instantaneous amount of power going in and out between the flywheel and the piston can be large - potentially a few times larger than the engine's normal power output, during the final stages of compression and the first stages of expansion. The alternator doesn't have to handle that, the flywheel takes care of it very efficiently.
Now, if you use a gas spring, if you want to use that concept of recycling mechanical energy from one stroke to the next, you are relying on the power stroke turning its energy completely into kinetic energy of the piston, then into compressed gas at the opposite end, then back into kinetic energy in the opposite direction, then into the compression stroke. As long as this is happening at the resonant frequency, it is all well and good, and the alternator need only bleed off the "excess" energy just as it does with the normal engine. But if you try to force this system at a speed other than the resonant frequency, now you are going to have to deal with instantaneous power going in and out of the magnetic side of this several times larger than the engine's power output.
That alternator might be 97% efficient but when you start compounding a 97% generating efficiency with a 97% motor efficiency for several times the engine's normal output because you are trying to force operation at an abnormal speed, that 3% loss repeated over and over again is killer.
These engines have to operate at their resonant frequency. Period.
Changing the gas pressure will change the resonant frequency, but it might also change the stroke of the piston. One way or another, the approach speed of the piston on the compression stroke will have to be in a pretty narrow range in order to achieve the next compression stroke without going too far.
Turbocharging, to increase the overall system pressure under load, would also increase the resonant frequency when running under boost, which is not a terrible relationship.
I suspect that controlling an engine of this sort is a nightmare, which could be largely why we haven't seen them in real world use.
