This is the 4K/60FPS version of the video I submitted for Runway Gen:48, which is a 48 hour competition where participants receive the design brief and have to generate and assemble their final piece in literally two days. I had a really big event yesterday which took up most of the weekend, so it was a mad scramble.
We're given a choice of three groups of subjects to include, mine were: Heartbreak (defining situation), A Dark Entity (archetype), Limbo (location).
In the story, our main character is trying to work through his inability to manage his emotions - he regularly dissociates, and can't inerpret the signals his brain and body are sending. He's been trying to communicate with his girlfriend about the experience, but she finds it hard to relate - outwardly, he's not emotionally responsive when he dissociates, and she just feels he's using his mental health diagnosis as a crutch for avoiding responsibility for himself. "His Supreme Obstinance" is the "dark entity" archetype. This is the personification of his autistic stubborness (I am writing about this sensation very literally) - and who presents as the dictatorial ruler of his emotions.
Making the video quickly like this was really tough, but also a great challenge to work to a super tight schedule!
I wanted to try to get characters as consistent as possible, using Midjourney's character reference function to create image start points, then animating each 10s segment using those static images. I found that using text + image prompts worked worse in Runway than simply loading the images without text guidance.
Reflection: I should have invested more time in developing the storyline more clearly, and defined some clear visual scenes and stronger imagery which would have been easier to build out as a basis for more varied plot development. On reflection, the video comes across as someone doing a lot of intense staring (which is on the money in terms of what the song's about) with a pretty vague sense of the underlying story or plot movement.
I also felt I needed to use a lot of visual effects to combine videos together - normally I'd have put in an additional 8 hours of storyline development and image generation, and I think the outputs from Runway would have been better working with a stronger set of starter content. I ended up with a much smaller set of baseline images, and that is reflected by the narrower set of video outputs.
On the other hand, it's been pretty cool to play with those effects and develop some ideas I've wanted to experiment with in overlaying videos with effects, and I will definitely be doing more with very synchronised music to video effect sequencing for future videos!