writes:As I mentioned in the previous article, our staff editors have been involved making some promotional videos for social media. When 
Editor Gwen 
saw the first video, she expressed an interest in being involved in the project too!
Again, the first step was to work with her selected self-portrait. This time, the picture was almost perfect as it was, except that it needed enlarging – which AI does quite well. I also asked the AI to remove the lettering on her outfit.
Note that we didn't change anything about her appearance. That's really Gwen, just the photo quality has been improved.
Our approach with the previous video was to do something light-hearted. This time, we decided to take a more serious tone, and help explain the hobby to newbies. Once again, we wrote a script, and used low-budget AI to make three 10-second videos, which we strung together to make a 30-second video.
Again, the voice is what the AI thinks Gwen should sound like. There seems no way to modify this (using this AI), and I don't know how it decides what voice to use – but it used the same voice in all three segments, so it's apparently based on the AI evaluation of the photo.
All of the gestures were 'invented' by the AI. We did specify to end the video with a smile, but the AI came up with the idea of the ending close-up. Also, to set the tone, we told the AI that Gwen was 'serious'.
What would we do differently next time?
We made this video before we understood about the preferred size and dimensions of the original photo, which is why you may see black strips at the top and bottom of the video.
So that's how Editor Gwen made her first TMP video.
I've also posted the video as a reel on the TMP Facebook page, where it has been one of our top videos. It is also reaching a slightly younger demographic than our usual audience, but it still runs older than Katie's video.