> but utilizing a microphone instead that listened for the subtle tonal differences in each key strike. How accurate is that? What if paper crinkles slighly,

> but utilizing a microphone instead that listened for the subtle tonal differences in each key strike. How accurate is that? What if paper crinkles slighly, or a deformation in the roller doesn't keep the paper tight?