Fitting of measured material data to MDL parametrization

“Given measured BSDF (or BTF) data, an MDL material for the model, and possibly hints / conventions, infer input parameters (and textures).”

Anyone out there who has solved this problem (in whole or in parts) and can share an experience report?

Isotropic measured BSDF are directly supported by MDL, although Iray’s photoreal mode is the only renderer that can take it.
Xrites AXF nowadays switched from BTF to already fitted analytical models, i.e. their tool can do the fitting. We do have a MDL material that matches xrites fitting target (nvidia::axf_support::axf_support contained in
practically all AXF we se by now are already fitted.
Fitting of BTF we did not do ourselves, so we have no experience.

Thanks for the reply.

Yes, I am aware of the existence of Xrite’s AxF and the implementation of their SVBRDF and car paint models in the material exchange package. Those models only fit fully opaque materials, though. Xrite’s services even seem to include more customized solutions…

However, that would give me data according to some fixed-function model. It’s not quite what I meant in my OP: A fitting tool directly at MDL-level would retain editability of the model and, given the raw data is kept, allow for refit at will (after e.g. hard-coding a parameter to become assumed fact, or adding/removing a layer of finish).