Discussion on the performance of oculus lipsync plug-in

After using this plug-in, I feel that oculus lipsync's mouth animation has some shortcomings, especially its mouth and voice matching are not very accurate, lack of authenticity, and some stiffness.
How do you think its mouth shape compares with other existing technologies and products in the world?
If it is not good enough, what do you think is the reason for this gap?
In the theory of this product, what limits it?
Is it one of the reasons that it only maps audio to 15 standard mouth shapes?
Is there any other reason?