It is not easy to integrate animated content in Sozi presentations. Usually, the embedded content is not aware of the "frame change" events emitted when moving from a frame to another.
For this reason, the Sozi media extension wraps the HTML video and audio elements so that we can react to frame changes as can be seen in the Media.js source file.
I don't know how Enve works, but I suspect it is designed to start playing when the document is loaded.
To be honest, I don't really understand how the animations are implemented, and how your second example fixes the issue.
If you can provide explanations, maybe I can come up with ideas to improve the integration of foreign content in Sozi.