Implemented properly, the Excels have distortion and spectral decay properties that are simply embarrassing to ever other driver I've ever seen measured. I saw a Joseph Audio waterfall plot that was just gorgeous. I've been living with said driver in the Xd for several months and the precision, resolution and naturalness is simply astonishing. It's no wonder it's all over the place in so many speakers, despire the fact that not everyone does a great job of implementation. I don't think there's a plastic, paper, carbon fiber and certainly not kevlar cone that can match the Excel in a good implementation.
Ribbons, on the other hand, often have their own little demons, such as huge resonance peaks that need notch filters of their own. In the end, as usual, it comes down to implementation, not one material vs another.