LePou
Member
Tolerances, I would say...
Creating an amp model from a schematic is going to make it sounds like an "ideal" amp, where all the components have that exact value and behaviour, when in the real world there could be up to a 20% of difference, for instance...
I've been talking to Thomas Serafini (the guy from Simulanalog and now Overloud) about this some days ago, and he told me that, for example, a real pot could not be perfectly logaritmic or linear, so it could react way different from an ideal one... this could explain the differences on the tonestack settings or the tone pot on an overdrive pedal...
Something I don't understand is the huge sound difference between different models of the same amp (Revalver 6505 vs 7170 or Revalver Mesa vs TH1 Mesa, etc.), but it could be dued to the schematics used and approximations...
You (Alu ?) are totaly right about tolerances, they are part of the assumptions that I was talking about. Apart from the assumptions related to the models, there are also approximations due to the numerical methods (or technics) used. As an exemple, take the Spice tube model presented by Koren. Real time evaluation of this model is very heavy on CPU. Numerical methods (or technics like lookup table, polynomials approx.) are most like needed to speed up the calculations of this model and this is another source of differences.
LePou