Hello!
I'd recommend to have a look at the approach taken in this thread instead:
Mouth animation using Rhubarb Lip Sync
The swaps are not a long fading transition but several image attachments alternating thanks to the audio track that you give to Rhubarb.
Another approach if you wish to have smooth transitions is to deform your mouth using meshes weighted to bones. In each animation, you arrange the bones to form the shape you need. This way, when you switch from one shape to the other, the interpolation should look smooth and natural.
For this to work well it means that all or most images are already in place, a good approach is to have a separate jaw image or a hole in your face mesh, and the open mouth and teeth below all already visible but hidden by the face when the lips are in a closed position.
