In line 3073 there was an unproper scope of cast to double.
// do the actual interpolation in double-precision arithmetics
// because it is a bit sensitive to rounding errors.
- const double factor = timeB == timeA ? 0. : static_cast<double>( ( time - timeA ) / ( timeB - timeA ) );
+ const double factor = timeB == timeA ? 0. : static_cast<double>( ( time - timeA ) ) / ( timeB - timeA );
const ai_real interpValue = static_cast<ai_real>( valueA + ( valueB - valueA ) * factor );
result[ std::get<2>(kfl) ] = interpValue;