Point being, when running in conventional, the Railsounds card is still a microprocessor controlling the sounds. If you ran in conventional without the "optional" 9V battery, then each time power is dropped for forward, neutral, or reverse operation, or when startup after a stop, then that processor is booting up from power for a second before it begins running. So it's possible the engine could be moving at speed and then because the processor booted up after already moving, it's unsure of the speed or state of inputs= sounds of prime mover speed might not match actual speed.
I'm not disagreeing with your descriptions, but in reality, the prime mover runs at a fixed RPM as it is a genset providing juice to the traction motors. The increase in sound is really the effect of "loading". Also keep in mind the 8 Notches of power level on the typical loco. I will defer to guys that know how the railsounds sound related to speed and load was designed, but in reality with a diesel loco, it's not really as simple as the loco speed strictly defining the level of sound, or at least I don't think it should be. A loco on it's own will be able to run up to a higher speed in notch one with no cars than it could with say ten cars. It might require moving up to notch 2 or 3 to get to the same speed, hense the increase in "loaded" engine sound. So from that sense, I'm not sure it is reasonable to even expect prime mover sound leel to match speed. All that said though again, I do not know what Lionel chose to associate the sound level change with, was it speed or was it sensed load. Especially for the early TMCC locos like this one (and knowing this GP9 uses a Pullmor AC motor) I suspect they did just associate sound level output with loco speed. It would keep things simple but probably not represent reality.
