MAKING-OF-RED-TOWER

MAKING OF “THE RED TOWER” | DESIGN BY MECANOO ARCHITECTS | TAIWAN

 

LOOKING FOR SOMETHING FRESH…

Mecanoo Architects approached us to create fresh visualizations for a tower project in Taiwan, seeking an external perspective beyond their in-house images. With no fixed brief or set number of views, we explored various camera angles so they could later send a photographer to capture the selected shots. Our process began with quick, low-detail raw renders to identify the strongest viewpoints and whether each should be shown in day or night conditions. We also analyzed sun position to determine the most photogenic time of day, guiding the photographer for both daylight and dusk shots.

After receiving the site photos, we screened all 65 shots, hand-picked the strongest ones, and quickly matched the building to check if those viewpoints held up. These are the ones that made the cut.

After some back-and-forth with the architects, we pushed the images toward realism to see what truly stood out. This project worked differently than usual because we were coordinating closely with the photographer, and a strong background photo basically guaranteed strong final images. The architects had to choose which of his shots we would use so they could request the high-res versions. They also wanted a closer viewpoint, which had been a point of debate from the start.

IS HAPPINESS OVERRATED?

From all the images, the architects picked five for the final set, but then came an unusual request. For the first time in our studio’s history, a client asked us to remove the happy mood of an image and turn it foggy and rainy. We were all for creating a foggy scene, it’s a fun challenge, but the problem was that the original high-res background was a perfect sunny shot, and transforming that into a convincing rainy atmosphere is notoriously messy. A few years ago it would’ve been almost impossible. Now, after discussing it with the client, we agreed to try out the possibilities with AI.

THEY USED NANO BANANA, REALLY?

Using the infamous Nano Banana, and after a few attempts, we finally got something usable. Once we upscaled the image, a few odd bugs and artifacts showed up, some fixable in Photoshop, others not so much. Still, the architects were open-minded; they were after a specific mood and cared more about the overall atmosphere than technical perfection.

And bellow you can appreciate the five final images. As the building was a tower we decided to crop all the images vertically to reinforce the sense of height and for a full immersion.

As mentioned earlier, because the background image was AI-generated, some artifacts are visible in high resolution (you can see it on the buildings in the back of the scene). We tried cleaning up a few neighboring buildings in the foreground, but in the end the client was satisfied with the result and actually encouraged us to lean into the AI approach.

This second image also came out strong. The client specifically asked for more street-level activity, so we added the scooters in the foreground with AI. It gives the scene a bit more local texture and scale without distracting from the tower and the additions helped anchor the composition and emphasize the everyday urban context the architects wanted to highlight.

This image also uses a photo backplate, and on the left you can see those lovely existing sheds, the kind of thing that normally screams “please don’t put me in the foreground.” So instead of pretending they weren’t there, we leaned in and slapped a massive graffiti piece on top to turn the problem into a statement. If something is ugly, might as well make it intentional…

They also asked for a close up view just to show some detail on the facade when you approach the building.

THE UNEXPECTED HERO

And the final image, this one wasn’t even supposed to be part of the bunch, we just made some last minute test and the architects loved it and called it the Hero shot!


If you made it this far, I’d like to thank you for reading till the end! The point of this page was to give you a peek behind the curtain and show how we actually work. We’ll probably post more things like this in the future so follow us.

When we start a project, even if the client only needs a few images, we like to push the exploration as far as time allows. That means lots of tests, lots of viewpoints. Every project is unique, and we approach each one with a mix of tools and techniques that allow us to explore multiple directions. Our experience helps us navigate these possibilities with purpose. This balance between experimentation and expertise is what lets us uncover unexpected perspectives and deliver visuals that not only look stunning but also communicate the essence of the project. This project had its own twist, AI played a much bigger role than usual. Normally we use it in a subtle way at the end just to enhance details, but here we had to go full power to transform an image. It was impressive, it’s also not perfect, it still required a ton of human cleanup, it’s good we were there to do it! AI will keep evolving, maybe steal our jobs, who knows? For now, you’re stuck with us humans!

If you’re into our work, send us an email. If you have questions, feel free to ask!

Chris