By Matt Jones | February 15, 2021 | 0 Comment
I made some videos for my Troop as we worked on the Programming Merit Badge. It seems the effect was lost on Zoom, so I am posting the videos of the guest speakers here.
These were generated using a service from Amazon, called AWS Polly to generate the audio. The audio was then lip-synced to the videos using an open source python library called Wav2Lip.
Here’s the Colaboratory Jupyter notebooks used to create the videos if you want to look through the source or use it for your own projects. I did not completely author this, but I cannot find the original source at this moment. It looks like the description on this video has links to the most up to date info on Wav2Lip:
(518) Deepfake With Audio Only [Wav2Lip] – YouTube .
In order to run this, you’ll need both a Google Drive with a bunch of stuff downloaded and an AWS account with Polly enabled.