Posted February 11, 2023 by droid productions
ElevenLabs then narrate that; they do a significantly better job at getting emphesis right than other text-2-speech systems I've looked at; likely it's doing sentiment analysis first, and using that to generate better text prompts
I initially planned to use some other AI tool for the voice over, but then I decided to go oldskool.
And ran that through Rhubarb Lip Synch system for phoneme extraction.
Keyframes were then imported into Spine2D, and the whole thing rendered out as a .mov file for After Effects to composite together.
You can read more about the process here:
https://www.patreon.com/posts/talk-to-me-like-78533157