Workflow - Commercials
Est. Time: 1 Hour
Steps to Completion
- Write script
- Generate Images
- Generate video
- Composition (Video Editing)
Write A Script
Here is the script result, I actually ended up just writing this myself, I was going to use GPT but then it all just kind of flowed out as I was contemplating what to ask the GPT to get the output I wanted, but here it is:
Starting a brand... it is like growing a tree.
You start with a seed... It's merely the idea of a tree, with maybe a little bit of instructions.
You plant the seed, and water it into a healthy sapling. There is only a single branch above the ground with maybe a few leaves. It has no fruit to bear, yet.
You water it diligently. Ensure it gets the right amount of sunlight. You bring your family, your friends, even other townspeople to provide the right energy and nutrients your growing sapling needs.
After months of love, many storms, and several rainbows, your sapling has begun to spring fruit. This is a good first step. You have a plant that creates fruit, but you're not done here.
To grow the tree from here, you must bring on more allies, more team members, to nurture more branches than you alone can support... creating enough fruit to share with all.
The best brands, the best fruit, leave a lasting flavor in the mouth. They are sought after by the townspeople and returned to time and time again, because they trust the fruit will return them to that feeling of their first taste.
Maintaining this tree that bears this fruit, is a labor of trust and love. This quest should not be taken on lightly, for it is with great burden that great trees grow to great heights.
We like to use DALL-E-3 via ChatGPT. So this is what we used.
We took each image over to RunwayML, and used their Image2Generator model.
Effectively we're coming up with scenes that are slighly more interesting in movement compared to the KenBurns animation.
Each video slice is then taken over to Capcut. We load up the audio first, and match the timing of the audio to the animations generated from RunwayML, extending and sliding the 4 second animation to the correct length of the vocal segment it represents.
Once we've downloaded the video, we can now post to our socials.
Tiktok, Youtube, Twitter, wherever.