Tweet
Hi there.
An extremely large and time consuming part of my workflow. Be it architectural design, product design or interior design, is modelling, texturing, finetuning and detailing the environment the object has to be rendered in.
Modelling the subject of the scene is sometimes just 50% or less of the entire job to get the desired end result renders. Placing the subject of the project in an environment that is detailed and realistic is sometimes a daunting task but necessary to convince the clients with renders that are complete and accurate.
The Chaos cosmos browser, HDRI's and scatter materials are a great help filling in this work but are still very labor intensive and not always produce the desired results. A lot of scenes in specific locations need heaps of manual 3D modelling, texturing and detailing to complete the picture.
Now if a guy is able to bring a 360 camera or smartphone to the location in which the project has to be rendered. Capture footage of the environment, load in a 3D NERF studio/ Gaussian Splatting self learning model and is able to incorporate this pointcloud with Gaussian Splatting into a vray scene. A lot (a whole lot) of the manual work in my 3D rendering projects will get automated and this will improve my efficiency by several orders of magnitude. If I could use a Gaussian Splatting trained Pointcloud as a "3D HDRI" as it were it would vastly, vastly speed up my render projects.
So this 3D rendering technique is still in it's early stage but any render engine compatible with this technique in the near future will be on the top of my list of tools to use. Just because of massive time savings alone.
An extremely large and time consuming part of my workflow. Be it architectural design, product design or interior design, is modelling, texturing, finetuning and detailing the environment the object has to be rendered in.
Modelling the subject of the scene is sometimes just 50% or less of the entire job to get the desired end result renders. Placing the subject of the project in an environment that is detailed and realistic is sometimes a daunting task but necessary to convince the clients with renders that are complete and accurate.
The Chaos cosmos browser, HDRI's and scatter materials are a great help filling in this work but are still very labor intensive and not always produce the desired results. A lot of scenes in specific locations need heaps of manual 3D modelling, texturing and detailing to complete the picture.
Now if a guy is able to bring a 360 camera or smartphone to the location in which the project has to be rendered. Capture footage of the environment, load in a 3D NERF studio/ Gaussian Splatting self learning model and is able to incorporate this pointcloud with Gaussian Splatting into a vray scene. A lot (a whole lot) of the manual work in my 3D rendering projects will get automated and this will improve my efficiency by several orders of magnitude. If I could use a Gaussian Splatting trained Pointcloud as a "3D HDRI" as it were it would vastly, vastly speed up my render projects.
So this 3D rendering technique is still in it's early stage but any render engine compatible with this technique in the near future will be on the top of my list of tools to use. Just because of massive time savings alone.
