Having dug through the original PNAS article this is about, there's nothing particularly interesting here. Different makers produced distinct instruments. They

Having dug through the original PNAS article this is about, there's nothing particularly interesting here. Different makers produced distinct instruments. They neglect the other major schools of violin making, such as that based around Steiner in Germany or the French makers, and show that they can find some spectrograms of singers that look like the spectrograms of the different instruments. Historically this is uninteresting, as we have no reason to think that the makers were consciously trying to imitate a particular kind of voice. It's not interesting physics (and the physics of violin acoustics have been done about to death). It's not interesting mathematics or statistics. And it has no implications for violin making.