I used Rhubarb and referred to these handy instructions here. Keep in mind you’ll need to use DaVinci Resolve which is a paid product that lets you use its older versions for free.
The steps to take:
- Download a release from Rhubarb’s Github Page (search for the “Release” tab on that page)
- Be sure to grab the phoneme images from here.
- Situate yourself with a
cd
into the rhubarb directory. - Copy your
sound.wav
audio file there. - Create a
transcript.txt
file there. - Run
rhubarb
with the exported file and the transcript file using:./rhubarb -o output.txt sound.wav -d transcript.txt
- Use DaVinci Resolve according to Jackall’s instructions which includes: 1/ Dragging all the mouth image files into a Resolve timeline, 2/ Exporting the timeline as a
mouth.fcpxml
file. - Run jackall’s Python program with:
python RhubarbDR.py
give it themouth.fcpxml
file, theoutput.txt
file, and any offsets you want to make happen. Provide the destination.fcpxml
file name. - Now in DaVinci Resolve import the timeline file (give it a new name instead of the default).
- Drag the
sound.wav
file *below* the video file because audio tracks sit under video tracks in Resolve. - Switch to Deliver tab (bottom of screen) and use codec Apple ProRes > Apple ProRes 444 to be able to “Export Alpha” …
You did it!