Hi Lewm,
Again, I don't build phonostage or preamps so I don't claim to be an expert on RIAA. There are many others who have posted on this issue on this thread and other threads, like the one I linked before, that are far more qualified to speak on the subject than I am.
As I understand it, the issue is not if RIAA correction should be done because it should. The question is during reproduction how faithful to the curve does this correction need to be to sound good? Some phonostage and preamp designers argue for strict adherence. Many more argue that the extra components need to conform to strict adherence add more bad things than they help. I think anyone who has experienced changes do to cap and resistor swapping can appreciate this concept.
IMO, Raul has always argued for strict adherence to the mathematical models that define the RIAA curve. Ok, that's his opinion and he's entitled to it. But I have never heard any other phonostage/preamp designer insist on this rigorous interpretation. Raul and Jose have built a very fine product. However, there are other products out there that do compete very well against their box. So why doesn't one sound clearly superior to another? I don't know, but this .1 dB adherence does/did not translate into any sonic revelations when I heard it. Unless something else in the implementation is holding it back.
As to the Soundsmith SG, it seems to me that Peter has posted that the SG does do RIAA correction. Perhaps, and this is my interpretation of what I'm reading here, not in the more conventional way because of the natural response of the cartridge used which is also not your conventional cartridge.
In the end it all comes down to how does one wish to select a product. By arguing numbers? Or by listening?