In HX, we have DSP elements that can change the sound; filters, gain, tube simulators, power supply sag, etc. All of these start at a neutral point, like the DSP is zeroed out. When making HX models, we measure the change of sound in a small bit of analog circuit, and then we set the DSP elements to make those same changes. All of the sonic changes have to be matched manually or semi-manually. Anything that is "feel" related comes along with matching the measurements. If it's not quite there, we can tune it manually. It's a long process, but everyone liked the results. In addition, this type of modeling easily allows original models to be made.
Agoura modeling is different. Like I said, I can't get into it, but it doesn't start with DSP set at a zero point that we then manually configure.