We went viral with an AI generated launch video

I got advice from https://www.youtube.com/@Putchuon (amazing veo 3 videos). They said, "Oh, the easiest and best advice I can give is to understand camera lingo. Wide shot, close up, medium. Or. Crane shot, dolly shot, tilt shot. Or Crash zoom, crash pan. Or Tracking shot, rotating shot, rising shot. Learning how cameras move is the most important thing. Second thing is understanding and describing the environment (location and lighting) separately. And then lastly, describing the subject. What do they look like what are they doing? If you separate all of those things into tiers BEFORE you tell Veo or the character to say something, it’ll get better."

And then I made this prompt (and a ton of other prompts and then put it into ChatGPT and Claude)

Describe the shot. Ex. Wide shot, close up, medium. Or. Crane shot, dolly shot, tilt shot. Or Crash zoom, crash pan. Or Tracking shot, rotating shot, rising shot.

Describe the scene

Describe the lighting

Describe the subjects

Describe what they are saying

Then I yelled at my computer and tweaked prompts for hours.