Closer to a Playable Build
Time has a way of moving faster than you think, and slower than it feels in the moment. It reminds me of my school days as a kid. Being in class seemed to drag out into an eternity, recess was gone in a blink, and if I ever got sick enough to stay at home… that day lasted about 1.5 hours in total.
I’ve now completed a decent draft of all the images for the first section of the game (the spreadsheet in the cover image is a look at the formatting behind the scenes). There’s still work to do in normalizing the sizes of the images, and the style. My partner also wants to run all images which include the face of the monk through an IPAdapter pass to make sure the character looks consistent in each image. It makes total sense (it would be confusing to play a Spider-Man game where he sometimes looks like Spider-Man, and sometimes like a hobo in a red hoodie and blue sweatpants), but I think it will require an unknown amount of time to achieve something of quality. Stable Diffusion tends to take a description and apply it to everything in the image. Imagine the Spider-Man game where you correct the image of Spider-Man only to notice that the crowd he was saving looks like a lot of Spider-Men.
I still need to create about 6 more scenarios for the final section of the game, and create a visual style for it that makes sense. It can’t be the same as the first acts style because it takes place 1,200 years in the future and technology has progressed. I want the look of the card/images to show that. My current idea is to show the scenario being described in the text as a bluish hologram floating above a sci-fi platform. Yes, I’m cribbing from Star Wars for this, but George Lucas created a language that is globally understood with some of this stuff.
As I said, it’s my current idea, but it might be biting off more than I can chew. This entire project has been about finding ways to reduce scope and implement AI into the workflow. This might not align with that thesis.
My partner has spent some time exploring various animation techniques to liven up the images and UI a bit, so that might make the entire experience feel a little more alive and polished. As of last week, Stable Diffusion 3 and Luma Dream Machine have dropped and they both threaten us with cool new ideas for how to make things better. But as we keep telling ourselves: “Done is better than perfect (or better).”
My aim is to be content complete by the end of the month, leaving July as a month for builds (and iteration). By the time my birthday rolls around, we should have a playable game we’re proud of. :)