Animating Lana

I recently wrote a fantasy story called A Rearrangement of Terms and generated its cover image using Chat GPT Images 2. I wrote the tale prior to conjuring up the image, so I was thrilled with the result on the first try because it captured the essence of the action and story environment.

After posting the story, I was curious about how different models would handle animating this particular scene. So, I used two different AI art generators to create the two animated videos for comparison. I anonymized the name of the software and just called them Model 1 and 2.

As a prompt, I copied and pasted a paragraph from the story that gave rise to the idea in the first place. I copied the text word for word without any alterations. This is the prompt:

Lana pulled the arrow on the string, her breath even, and at its rising zenith, she flicked her fingers and released the arrow. The projectile whizzed through the air like a hot metal rod, alight in flames, it struck the giant spider on the mandibles, exploding in a shower of sparks that sent the foul creature scurrying and screeching down the dark mountain tunnel.

It looks like Model 1 took more creative liberties with the image and the prompt. It focused on the flow of events including the shower of sparks. It did not keep character consistency per se, but it created a rather explosive action-packed scene with the appropriate sound effects.

Model 2 maintained character and scene consistency that was visually breathtaking and suspenseful but not as active as the scene generated by Model 1. It added artifacts like the fire engulfing the bow, which is actually not a bad idea if you want to show that the character is firing magic arrows.

The software packages that I used are not the most powerful out there, and I don't have the most convincing AI skills in town, but I'm sure that an AI cowboy or cowgirl with the right aim can fire up the prompts and create an incredible story.

Thank you for exploring this dungeon 🕷️

X	InLeo	Instagram	NFT Showroom