xlights 3.x had three types of papgayo output
1) Autofaces. These create a face on a megatree or matrix. No further setup needed
2) Corofaces. You had to define the 10 phonemes and list the channels to be turned on.
3) PictureFaces. You provided 10 image files (Like a snowman) each with a different mouth shape.
The recent work on Integrating Papagayo saw
Dan Kulp implement (2), Corofaces
Gil implemented (3), Picture faces.
No one has re-implemented Faces.
If you still need this, you would need to use the Papagayo tab, import in a Papagayo file *.pgo, and create an xml file. This file si xlights 3.x. Open that file and it is converted.
So we have 2 out of the 3 singing faces implemented. The most used are done. I bet 75% are corofaces (2), 20% were picture faces, maybe 5% were the faces effect.