Sportscar motion graphics

Light version.

Light version.

First draft.

First draft.

Source mesh with gradient applied.

Source mesh with gradient applied.

InstaLOD reduction comparisons.

InstaLOD reduction comparisons.

Lighting setup.

Lighting setup.

Desintegration effect and ground fog added.

Desintegration effect and ground fog added.

Light version first draft.

Light version first draft.

Going a bit crazy and adjusted focal length. First experiment with motion blur.

Going a bit crazy and adjusted focal length. First experiment with motion blur.

Color adjustments, background refinements and (still) too much noise in the rear. Also lighting and volumetrics adjustments. No motion blur but depth of field added.

Color adjustments, background refinements and (still) too much noise in the rear. Also lighting and volumetrics adjustments. No motion blur but depth of field added.

Obligatory node tree p*rn. A compilation of various Blender geometry nodes and internal compositor.

Obligatory node tree p*rn. A compilation of various Blender geometry nodes and internal compositor.

My final project for Paradox Cat. It's for internal marketing/branding purposes. The request was to make the car less dense from front to back and then dissolve the rear into a cloud of triangles. The light and dark versions are nearly identical but in terms of geometry but differ in certain aspects for visibility and readability purposes.

modo was used to clean up the innitial triangulated source mesh and to use it's falloff tools to apply a vertex color gradient that can then drive reduction via InstaLOD.

Blender was used for animation and rendering and geometry nodes are the basis for the fx. Eventually the scene got so heavy that working with it slowed down to a crawl. To navigate, let alone animate, many things had to be turned on and off on a constant basis. That being said, the scene was surprisingly stable and crashes were few and far between.

Also for the animation a lot had to be simplified, like applying some geometry nodes, disable some fx to save on render times and also use the original, untriangulated car.

Because this project is self contained and time was short the Blender internal compositor was used instead of using a traditional post process pipeline.

A lot of time went into dealing with performance issues, finding workarounds for Blender's quirks and bugs and of course limited project time. All in all I am very happy how things turned out.