The first 3D software I downloaded was Blender.
I was in my final year at university and eager for any open-source 3D software that came without a price tag. By the time it was unzipped and installed, I had watched about two YouTube tutorials and was feeling pretty confident.
I opened my first project, played with the little default cube Blender starts with, and . . . realised I had absolutely no idea what to do next. How do you actually create a scene in 3D?
So, from initial blockout to final render, let me walk you through the 3D modelling pipeline, sharing all the tips, resources and mistakes I’ve learnt along the way.
Phase 1: References
It’s tempting to jump straight into the 3D workspace and unleash your inner creative beast.
Unfortunately, to avoid mistakes down the line, you’re going to have to gather some visual references first. These images should cover everything from model and texture examples to lighting and shot composition.
Most 3D software allows you to import references as image planes within your scene. This is super helpful for modelling down the line.
Phase 2: Modelling
Modelling begins with a blockout. This is basically a 3D sketch for your project that breaks the scene down into basic shapes.
It’s important to set up a virtual camera at this juncture to ensure shot composition is proportionally balanced (remember the big, medium and small design principle – a good scene will have objects of varying sizes).
A low-polygon (poly) model can then be built from the blockout. This is a more detailed model but retains a low polygon count, allowing easy manipulation of the geometry and quick render times. If you’re rendering in real-time (i.e. a game engine), this will be your final model with high-poly details ‘baked’ onto texture maps and applied to the low-poly.
The next step is to build a high-poly model from the low-poly. This model will be your final appearance and should feature all the intricate details you want, regardless of large poly-counts.
Make sure you keep copies of each modelling stage (trust me, I learnt the hard way)!
Phase 3: Lighting
Remember those virtual cameras we just set up? Now’s the time to look through the viewport and check your scene is clean, assembled, and ready for lights.
Virtual lighting is very similar to live photography lighting. This means a three-point setup is usually a good way to go, complete with key light, fill light and rim light. Also, be sure to experiment with light temperature and colour to refine the mood of your scene.
Of course, we’ll fiddle more with lighting after texturing is completed, but it’s a good idea to set up lighting now and define where the focal points of your piece are (i.e. which objects are brightest lit).
Phase 4: UV Unwrapping
Let’s take a quick intermission and unpack 3D texture theory.
Unfortunately (or not), it’s far easier to work with high-resolution images in two dimensions than in 3D space. This means 2D textures can be far more detailed and realistic when the complexities of 3D geometry are removed from the equation.
So, how do we apply a 2D texture to a 3D object?
The solution is UV unwrapping, a tedious yet crucial step that tells the 3D software where the object’s ‘seams’ should go using UV coordinates. Basically, you are ‘unwrapping’ your object and laying it flat into 2D space.
This allows the 2D texture to be applied to the 3D object without any distortion, like an Easter egg wrapper perfectly wrapping around the egg.
Phase 5: Texturing
Welcome to the ‘fun’ stage. Texturing is where your model comes to life. It’s also (at least for me) one of the most confusing stages.
That’s because texturing isn’t just about choosing what colour to use; it’s also about creating shaders to determine how the object interacts with light.
A common shading technique is physically based rendering (PBR), which comes with a host of affordances, including roughness, IOR, transmission, specular, and more. This is where your object’s material is defined.
For instance, a high roughness will make a surface appear matte. Decrease specular roughness, and it’s a mirror. Now turn on subsurface scattering, and you have skin.
These textures should also come with several texture maps (including normal and bump maps). These maps will project extra detail onto your model, like the illusion of height, without requiring additional modelling.
Phase 6: Rendering
So, you’ve created your model, it’s textured, it’s lit, you think you’ve finished . . . but you’re not.
Rendering is a little more complicated than heading up to the top menu bar and clicking ‘render’.
To begin with, which render engine are you using? Each engine (including Arnold, Redshift, Octane Render, etc.) will use different rendering algorithms and produce different render results.
For instance, a ray tracing engine will produce highly accurate light simulations, while engines prioritising rasterization will reflect light faster but to a lower quality.
There are also many settings you’ll want to choose yourself. Do you want depth of field? Or global illumination? What about post-processing effects, like colour grading and glare?
Just remember that rendering can take a long, long time to complete, so be sure to schedule your time accordingly.
Conclusion
3D modelling is by no means a simple process. It’s time-consuming and confusing and can lead to eye strain and a sore back.
However, once you wrap your head around it, I promise the process is a lot of fun. It’s super rewarding and highly creative. If you’ve got the skills and the dreams, there’s nothing you can’t create in 3D.
So, good luck and enjoy it!
