The ECM makes linear blending calculations between cells, which can be done in excell as well. It's the bases all timing calculators work off of, EFI or UDC. Without data logging we cannot confirm it, but it has been confirmed on other platforms.
The duration table is nothing more than a calibration, or conversion, table. It's where the ECM takes desired volume and turns it into time, since the ECM cannot measure volume but it can measure time.
If you look at the labels I put on you can see that there are lots of calibration points in the low load region, and fewer as load increases. This is because at low loads the ECM needs more accurate data to keep the motor smooth, surge free, and stable at a constant load. As you increase load you can decrease the calibration points and rely more on blending because the increase load smooths out the operation... but that is also why an improperly tuned duration table can mess with the way a truck runs.
In very simple terms (as there are many limiting and adjusting tables we don't see in UDC) when the ECM does it's calculation for fuel it looks at TPS, then rpm. Using those two inputs it determines how much fuel needs to be injected (we can't see those tables in UDC). Now that it knows how much fuel it goes to the pressure table to see the commanded pressure for that fuel amount and rpm, then using the commanded pressure and desired fuel the ECM then looks to the duration table to determine how long to open the injector, and lastly it looks to timing as to when to start the injection event.
So as an example, lets say we are towing a trailer up a grade. The TPS (or cruise) is saying we need more fuel and the ECM decides to run 79% load, rpms are fixed at 2200. So from there the ECM say's okay at 79% load (110mm3 is what the ECM uses) and 2200 rpms the pressure table says 16682 psi (115 MPa is the unit the ECM actually uses). So now the ECM knows we want 110mm3 at 115 MPa, so we go to the duration table. There aren't row's/columns for 110mm3 and 115 MPa so we extrapolate like the ECM will. The 4 cells we need to blend (OEM duration table) are 1740 2102 in the 68.9% load row and 3300/3400 columns, and 1654 2001 in the 75.1% load row and 3300/3400 columns (based on MADS labels, see the above photo for other labels). It doesn't matter which way you blend first, but the desired value is halfway between all 4 vales. So I'll blend the row's first. (1740+2102)/2 = 1921, which is the value for 110mm3 at 110MPa. Then (1654+2001)/2 = 1827.5, which is the value for 110mm3 at 120MPa. TO find 110mm3 at 115 MPa we take (1827.5+1921)/2 = 1874.25, I rounded to 1874. So the commanded injector open time for this equation is 1874us.
This is also how we add fuel with the duration table, by telling the ECM it takes longer than it actually does to get the desired fuel. It's also why the lie-o-meter gets worse with duration table modifications.
Does that help?