# Chicago in Five Breaths

Source: https://dinand.com/side-projects/chicago-in-five-breaths/
Author: Dinand Tinholt

> A 51-second watercolor film of Chicago, from first light over Belmont Harbor to Navy Pier after dark, made with AI for $11.65 in video-generation charges.

*51 seconds · narrated, with music. Best with the sound on.*

First published as a post on [LinkedIn ↗](https://www.linkedin.com/posts/tinholt_i-wanted-to-see-how-far-i-could-take-a-creative-activity-7519983040866209792-BN72)

I wanted to see how far I could take a creative idea with AI. A short film about Chicago, moving from the first light over Belmont Harbor to the glow of Navy Pier after dark, inspired by @wecraveai. The result is Chicago in Five Breaths, a 51-second watercolor journey through a city I love.

I wanted Chicago rendered in ink and watercolor, with the light changing as we traveled through the city. Our journey would pass the Water Tower and follow the river, reaching Cloud Gate at dusk before Navy Pier lit the lake. We (the AI and AI) developed five master images and agreed on the visual direction before generating video. Planned camera movements and generated transitions helped the separate locations flow together. ByteDance Seedance 2.5 generated the footage through OpenRouter. Google Lyria supplied the music. We tried different voice treatments before settling on Google Gemini's Aoede voice. ChatGPT coordinated the process, including the editing needed to bring those outputs together.

The video-generation charges, including an initial test, came to $11.65. Audio generation and the time spent shaping the film were additional. At that price, an idea becomes something you can afford to explore. You can watch a first version and discover what deserves more attention. An alternative treatment becomes a practical next step, with enough room to follow your curiosity. I've tried video generation with local models as well but haven't reached the same level of quality I wanted but hopefully soon that will be possible as well.

The narration showed me how much that freedom matters. The early script described the scenes individually. I wanted a story that carried us through Chicago, so we rewrote it around the hope of the people who built the city and the possibility of helping shape its future. We refined the voice until it drew the listener closer.

That ability to keep shaping the work has consequences for business. A leadership team can explore how a message will feel to an audience before committing to a larger production. Learning teams can test where an explanation holds people's attention. For client work, a film can make an imagined future easier to discuss. People can respond to a shared experience and explain what needs to change.

As the ability to produce convincing material becomes more accessible, the brief deserves more thought. What should someone feel when they watch? Which details will help them understand? A clear intention gives the technology useful direction, and judgment shapes what reaches the audience.

I began to imagine what people across an organization could make with room to experiment. Someone with a strong idea can carry it further and build confidence through the work. I began with the feeling of Chicago beside the lake. By the end, that feeling had become a film I could watch and share from my phone.

First published as a post on [LinkedIn ↗](https://www.linkedin.com/posts/tinholt_i-wanted-to-see-how-far-i-could-take-a-creative-activity-7519983040866209792-BN72)

[All side projects ](https://dinand.com/side-projects/)[Back home ](https://dinand.com/)

---
Site guide for AI agents: https://dinand.com/llms.txt