Daniel Statham Studio sponsored 3D visualiser to participate in artificial intelligence workshop, exploring new approaches in architectural rendering
6th June, 2024
Daniel Statham Studio sponsored 3D visualiser to participate in artificial intelligence workshop, exploring new approaches in architectural rendering

In Spring 2024, Daniel Statham Studio sponsored our 3D visualiser Fang Cui to join HyperPerceptions, a comprehensive workshop hosted by Chantal Matal, a worldly celebrated multidisciplinary architectural and generative designer and lecturer of UCL the Bartlett School of Architecture, whose work have been widely exhibited in international exhibitions such as Galerie Fractal in Paris and the Venice Biennale.

The workshop covered the revolutionary Artificial Intelligence (AI) platforms used to generate hyper-realistic visuals depicting architectural design spaces, Warp Diffusion to create immersive scenes and animations, and Neural Radiance Field (NeRF) to produce novel views of complex 3D scenes and 3D models.

As a research-driven design studio that champions innovative minds and exploratory spirit, we’d like to share Fang’s research and study from this workshop, as well as an interview in which she discussed her work, professional growth, and envision to incorporate the newly gained skillset into the Studio’s future production.

DSS: Let’s go back to the very beginning – could you tell us about how you joined Chantal’s workshop?

Fang: Actually, it was Daniel. I think Daniel has always been devoted and supportive in exploring and sharing the latest technology, software, fashion shows, art, and many more, which also greatly influenced me and people around to keep our eyes and minds open. Many architects who’s been in the field for a significant period of time lose the spirit of exploration and innovation, and tend to repeat themselves. But Daniel is on the opposite; he likes to experiment, and he enjoys experimenting with us. So one day I walked into the Studio and he looked super excited and said, “Fang I found a workshop that you might be interested in.” We reviewed Chantal’s work together and both agreed that her approach would greatly enrich and expand our existing practice method. And Daniel signed up for me and paid for everything.


DSS: can you elaborate on the notion of “Enriching and expanding DSS’s existing practice method?”

Fang: Right. For quite a long time we have been keen to explore how to use artificial intelligence to generate – not just regular 2D imagery – but 3D videos, with the help of modeling and software such Unreal Engine 5, which is commonly used in the field of game production. With this workshop, Daniel and I were curious to see if we could incorporate this method into architectural rendering process to create something grander, like a world that you can actually walk into with a point of view (POV).


DSS: It’s interesting that you mentioned the term “POV,” because often time in architectural render, whether it’s a video clip or an image, you can see the building from all sorts of perspectives, but you don’t get to experience it from a first-person point of view.

Fang: Absolutely. Without a POV, the videos you make are a collection of clips; but with it, you’ll get a storyline, an individual walking in a 3D space, etc. The project then will be able to deliver emotions, heartbeat, and memory. The video you present to client will become a sequence of spaces that can be continuously experienced on a personal level – spaces that have a conversation among themselves. This project also encouraged me to contemplate upon the meaning of visual art: Not only should a 3D visualizer stylises a naked architecture and captures its beauty, but also acts like a director to narrate an architectural experience to the viewer.


DSS: With that in mind, can you share your work with us? What is the story you try to tell and the space you want us to experience?

Fang: The work is titled “The Civilising Process,” because its narrative was inspired by the theory in the groundbreaking work of German sociologist Norbert Elias, The Civilising Process. To summarise it in the most simplified way, it is about the awakening of an AI figure’s self-awareness. You can see the figure is running on a path – two paths actually: one is leading to the nature, snow mountains and wildness, while the other one is toward an urban scenery, a modern society bonded by laws, rules, and ethics. The production process was a struggling but also fun one. Some of my colleagues at the Studio are present in this work too, thanks for their kind support of course.


DSS: Any scene or details you’d like to emphasise in particular?

Fang: My favorite one is no doubt the sea of flowers, sensational and relaxing. It’s also Chantal’s favorite. The image on the screen was inspired by Hiroshi Sugimoto, Ricardo Bofill, and Gasper Noé, and generated by AI. I think whether it is a creative project like this one or any architecture render, it is crucial to introduce a consistent aesthetic into the visual.


DSS: Totally agree. You touched upon the inclusion of AI-generated images – can you give us more detail regarding the software and technology you used for this project?

Fang: Unreal Engine 5 was the most important one. Mid Journey was used for generating 2D images, and Kaiber for 3D videos. Those two are prompt-based AI platforms. There was also another one I used – Warp Diffusion, which required not text but coding in order to generate videos. Coding was not frequently used in my previous work, but this whole experience was supposed to be about experimentation and stepping outside of my comfort zone, so I tried and managed to do so in the end, which is one of my proudest moments in this project. And Luma Field, which I used to create Neural Radiance Fields (NeRF) and generate 3D videos of the snow mountains, urban setting (in fact, New York City), and the images of my colleagues.


DSS: Any other challenges you encountered during this process?

Fang: Another one I think is worth mentioning is the diffusing particles you see at the beginning of the video, which took me a very long time to figure out how to import this design from Cinema 4D as geometric animation sequence into Unreal Engine 5. It was my personal attempt – this detailed design also added another layer of difficult to the final video. But I’m glad I was patient and persistent enough to finally work it out. It animates the overall visuals for sure.


DSS: I’m curious to learn what are the biggest differences between this project and other rendering work at DSS?

Fang: First of, I think it’s fair to say that I created a world. Instead of choosing camera angles and combining clips, this project allows one-take, which is something I’d like to try with future projects at DSS if possible. And as mentioned earlier, you see everything inside this world with a POV, so consequently a strengthened sense of immersive experience is delivered to the viewer. On a personal level, I am very grateful for this opportunity, because usually as a 3D visualizer, you’re subject to the project’s conditions, client’s requirements, architect’s design, and so many other things that your hands feel tightened up. But this project is completely about unleashing my creativity, experimenting in any direction I wish, and eventually producing an artwork that I enjoy. In other words, Daniel sponsored for my professional growth, which is something I’m very thankful for.


DSS: What is your biggest take-away from this project that you think can be transferred into your future work?

Fang: A refreshed understanding of software and AI of course, especially how to fuse their advantages together. Many have been arguing about how to use AI to deliver architectural designs, but I’m more interested in utilizing AI and incorporating it into the creative process. Interdisciplinary interaction among software and platforms. Even though the workshop already ended, I’m still pondering on the notion of using Unreal Engine 5 and Cinema 4D to present a game-like, immersive experience to client as deliverable, such that they can really explore the project and interact with in. With the advancing technology of VR and AR, maybe there could be even more possibilities. I’m willing to give it a shot, and so is Daniel and our Studio.


DSS: One last question – how do you see this workshop and your research reflect the value and practice approach of Daniel Statham Studio?

Fang: I believe they both share a great commonality in the pursuit of innovation and experimentation. Here at Daniel Statham Studio, we always embrace new technology.  We are curious, we are not afraid of change, we are active, and we look forward to creating something new, which in some way is the essence of design, because what is the point of self-repeating?