The light position and amount (and any other qualities of the light sources) may have a crucial impact on the final shadow on the canvas. Therefore some experiments were necessary:

Experiment 1: “Lightdome”

= Sphere with 70 randomly positioned lights. (shadow type: area)

resulting look in the machine:

Conclusion: Not feasible.


Experiment 2: “Lightring”

= Ring with 25 randomly positioned lights. (shadow type: area, intensity: 25%)

resulting look in the machine:

Conclusion: Not feasible.


Interim status so far: It takes a few strongly directed light sources to create sufficiently “interesting” shadows. Above a critical mass, the resulting light is too diffuse to be processed by the machine.


Experiment 3: “HDRI”

= Using an HDRI-globe as only lightsource. The HDRI was created two years ago as an experiment on its own. (Very dark conditions with a few tiny lightsources.)

resulting look in the machine (with exposure compensation of +4 EV!):

this looks promising! So let’s render this image as 32bit greyscale EXR to use it as a displacement map:

Resulting Image when rendered with Global Illumination (Cinema 4D rasterization on CPU) in mode “HDRI Image, high quality”

Resulting Image when rendered with Global Illumination (Cinema 4D rasterization on CPU) in mode “default”

 

When this displacement map is used and rendered in 3D geometry it could look like this:

Quick viewport render (Blender Cycles) of the displacement map rendered in “default” settings

Quick viewport render (Blender Cycles) when using the displacement map generated with “HDRI-high-quality”

 

Quick conclusion

this approach could introduce some very distinct qualities to the scene. BUT: the HDRI is a static image – the position of the individual lightsources can not be altered/adapted in any way.