Does anyone hear actually buy an entire rig preassembled? With a single instrument, you can get a feel for timbral accuracy, so its a good place to start, but clearly you have to move on to music that uses multiple instruments and voices to complete the evaluation.
But to answer your question. Piano first, followed in order by cello, violin, french horn, then soprano and tenor voices.
I have heard speakers that do all of these well but fail on piano. That is a deal breaker. On the other hand, I have never heard a speaker that gets piano right and fails on the rest. So the lynchpin is piano.