As you may know I have been working with Luma AI's Dream Machine, a freemium online application that makes 5 second clips from a written prompt and/ or photo. I have found the experience very interesting and encouraging to see the progress of. In case you don't know it is free to anyone with a Google account with 30 generations monthly with the option of a paid subscription for more. In light of that and my own generation quota being renewed I've decided to give some things I've noticed while using it.
Using a photo is the way to go. You already have an idea of what you want to see, why not give it something to start with? Make sure you have a good, clear photo to start with. It doesn't have to be too large. I used 1024 X 1024 as a basis.
The content of your image dictates how it is animated. I probably won't post it but there was a very interesting mistake with one of my generations. The image I used was not touched up, it was an image I took straight from Ideogram (another Freemium