Soooo,
i think i figured at least the hornresp horn modeling part correct.
The following figures resulted:
Tuba30, external width 30", dual 12" Driver (applies to all)
Here the Stack of 4, V-Plated, Driver is the 3012LF, max SPL:

- T30 30 dual 3012LF 4Stack maxSPL VPlate.png (10.67 KiB) Viewed 197 times
Here the Stack of 4, V-Plated, Driver the LaVoce, max SPL:

- T30 30 dual LaVoce 4Stack maxSPL VPlate.png (9.97 KiB) Viewed 197 times
Here The T30 single vs stack of 4 without V-Plate, max SPL:
Titan 48, external width 30", dual12" loaded (applies to all)
Here the Stack of 4, V-Plated, LaVoce Driver, max SPL:

- T48 30 dual LaVoce 4Stack maxSPL VPlate.png (10.64 KiB) Viewed 197 times
Here the T48 single vs. Stack of 4 without V-Plate, max SPL:

- T48 30 dual LaVoce 4Stack vs single maxSPL.png (11.69 KiB) Viewed 197 times
So, first of all: if anyone feels there has to be an error, give it to me please.
Assuming this (more or less) is correct, it seems as the T30 really goes a bit lower, even in stacks of 4.
The difference between T30 4Stack and T48 4Stack is just around that 5db (Comparing 4 cabs to 4 cabs, so half the Volume of space from the Cabs)
Doing 8 T30 stacked V-Plated:

- T30 30 dual LaVoce 8Stack maxSPL VPlate.png (10.79 KiB) Viewed 197 times
To compare the same Volume of cabs - we even surpassed the T48 limits. T48 brings a dent at around 80 Hz, from there on the T30 (double cab count, same volume) tops it out.
And the roll-off down below is smoother within the T30 stacks, no matter which variant.
The limiting point seems to be the Driver, not the cab itself.
So, to get to my questions from the beginning of ths trip:
- I don't think there is something the T48 would achieve i could not reproduce by just building more of the T30. The better upper range extincts within the same volume of cabs (there is doulbe the amount of drivers included)
- i need double the amount of drivers and amp power to achieve that with the T30, on the same power there is around 5 db difference but still the T30 plays lower
BUT the storage space needed still the half. The result of 8 T30 will blow me away anyway, even if the simulation is typically of by 2-3 db.
- The Hornextension doesn't seem to be in any sensible relation comparing the size of the horn to the result. I don't think i will follow that, maybe when i run out of ideas what to build next.
As the driver used limits the output situation, i will keep an eye open for other options that may fit (and will come back with the questions about them when they simulate better)
If anyone sees errors in there, please hint me on them i will try to fix that. Or any ideas which alternatives to simulate.
Edit: And in addition, the approximation with doubling the amount of cabs gives you x db, plating them gives you y db: Somehow not that far away, but it's just not that linear and effects every band in another way. And the plain calculating does not include limits of the driver you achieve, they come in first.