A MAGICAL JOURNEY THROUGH NeRF:02
FuturistoGraphy Glimpse 👀️
Last month, I embarked on an experimental journey with NeRF, using NerfStudio and Unreal Engine to create a mesmerizing flower garden scene. At the time, there was no direct way to import NeRF into Unreal Engine, which required me to render the NeRF background separately and composite it together with the other elements.
However, just four days ago, Luma AI released an Unreal Engine plugin, enabling seamless NeRF integration within the engine. Excited by the possibilities and curious to see the performance of their NeRF model in real-time, I couldn't resist revisiting my flower garden project.
The process was incredibly straightforward. I simply dropped my footage on the Luma AI website to generate a blueprint of my NeRF model. Surprisingly, the NeRF BP file size was a mere 240KB and 59MB of texture, thanks to the efficient volumetric representation learned by the neural network. This compact format allows for high-quality 3D content with minimal storage and bandwidth requirements.
I then reused the camera and miniature CG animations of people and animals that I had set up previously in Unreal Engine. After importing my NeRF model, I took the opportunity to experiment with lighting. Not only was Luma AI's NeRF model visually stunning and rich in detail, but the real-time performance was also nothing short of fantastic. This was the first time I had witnessed high-resolution NeRF in real-time without rendering.
![]()
![]()
![]()
![]()
![]()
Additionally, the NeRF model's ability to react to lights within Unreal Engine allowed me to transform the flower garden into a nighttime scene, complete with vivid and colorful lighting. This new experience was truly awe-inspiring. Last month, I wished for everyone to have access to these incredible assets quickly and easily – and now it's a reality.
As Luma AI's NeRF technology continues to evolve, I eagerly anticipate what the future holds. In the meantime, I'll be exploring the ever-expanding potential of this groundbreaking integration between NeRF and Unreal Engine.
However, just four days ago, Luma AI released an Unreal Engine plugin, enabling seamless NeRF integration within the engine. Excited by the possibilities and curious to see the performance of their NeRF model in real-time, I couldn't resist revisiting my flower garden project.
The process was incredibly straightforward. I simply dropped my footage on the Luma AI website to generate a blueprint of my NeRF model. Surprisingly, the NeRF BP file size was a mere 240KB and 59MB of texture, thanks to the efficient volumetric representation learned by the neural network. This compact format allows for high-quality 3D content with minimal storage and bandwidth requirements.
I then reused the camera and miniature CG animations of people and animals that I had set up previously in Unreal Engine. After importing my NeRF model, I took the opportunity to experiment with lighting. Not only was Luma AI's NeRF model visually stunning and rich in detail, but the real-time performance was also nothing short of fantastic. This was the first time I had witnessed high-resolution NeRF in real-time without rendering.
Additionally, the NeRF model's ability to react to lights within Unreal Engine allowed me to transform the flower garden into a nighttime scene, complete with vivid and colorful lighting. This new experience was truly awe-inspiring. Last month, I wished for everyone to have access to these incredible assets quickly and easily – and now it's a reality.
As Luma AI's NeRF technology continues to evolve, I eagerly anticipate what the future holds. In the meantime, I'll be exploring the ever-expanding potential of this groundbreaking integration between NeRF and Unreal Engine.
Related Projects:
NeRF01
NeRF03
Imagined by Masaki Yokochi