What I did:

Rather than spend weeks modelling period-accurate cars from scratch (which isn’t what I’m being assessed on), I sourced free 3D models for the four vehicles John drives across the years. Getting them to work together as a cohesive set turned out to be quite a bit of work.

The models:

I found Creative Commons licensed models on Sketchfab:

These were the closest matches to the actual cars my dad drove. The files came as a mix of OBJ and FBX formats, so I imported each into Blender individually to check them over, then saved them as blend files for easier workflow.

Getting them to work together:

I created a new blend file, brought in my John character for scale reference, added a ground plane, and appended each car. Immediately obvious: massive scale discrepancies. The cars were all wildly different sizes relative to one other, and the level of detail varied.

The Peugeot 2008 was by far the most detailed – probably meant for close-up renders. I decimated the geometry as much as possible to bring it closer to the other models’ poly counts. The Rover SD1 was the lowest poly of the lot as it was meant for 3D printing. Most meshes were single pieces of geometry, so I separated out components like windows to give myself more material control.

Interiors:

Neither Rover had any interior whatsoever. The P6 had almost black windows which looked inconsistent with the other cars. Since John needs to be visible driving, I needed at least basic interiors.

I duplicated the steering wheel from the Peugeot 308 and placed it in each Rover. For the P6, I selected a section of the car body, duplicated it, scaled and manipulated it to sit just inside the outer body as an interior shell. Cut away the bits I didn’t need, then extruded basic geometry to represent seats. To save time and potential errors, I cut away half the mesh and added a mirror modifier rather than modelling both sides.

Interior in the making:

Extruding seat shapes:

Cut in half and mirror modifier added:

Glass separated from main mesh, helped me to see where the dashboard should be positioned:

I duplicated the P6 interior for the SD1, separating the front seats and adding headrests to match the later period:

Again, separating the glass from the main mesh helped me to position the dash:

The sourced Peugeot models were left-hand drive, so I mirrored the interiors to fix that – the story takes place in England, so they needed to be right-hand drive.

Flipping the steering wheel and column:

Right-hand drive:

Whole dashboard flipped:

The new Rover interiors:

Materials:

The P6 had been UV unwrapped with everything on a single texture. To change materials, I had to open the texture in Photoshop and find the relevant sections. The paint was super shiny and the chrome highly reflective – too photorealistic for my stylised aesthetic, so I tweaked the appearance to take the shine off.

Polished chrome:

Shiny paint and blacked out windows:

The 2008 had loads of different materials, all labelled in Italian (I think), which made them tricky to follow. I worked through systematically, adjusting roughness values to tone down the shine and bring everything more matte.

I changed each car’s colour to match what my dad actually had; I also added number plates to each, apart from the P6 as I can’t remember what that registration was.

What I learned:

Sourced assets are never production-ready. Different artists work differently – some UV unwrap everything onto one texture, others create dozens of separate materials. Some model interiors, others don’t. Understanding how different people structure their files and approach materials is a valuable skill.

The time saved by not modelling cars from scratch was significant, but there’s still substantial work in bringing disparate assets to a more consistent standard. It’s not just downloading and dropping them into your scene.

Next:

Now that the cars are sorted, my next two blog posts will cover the production infrastructure I’ve set up to manage this project.

The first will document my technical setup – hardware, software, file organisation, backup strategy, and how I’m using Trello to manage the workload alongside my freelance work commitments.

The second will be a research update covering my literature sources, how they connect to my evaluation framework, and stakeholder progress.

I had considered adding research for some of my earlier posts so you’d know why I’m doing what it is that I’m doing, but ultimately decided to keep this blog as a technical reference of my progress. These new posts are now required for my upcoming WIP assessment. They’ll be useful documentation for my thesis reflection later.