Introduction

Diatoms are microscopic unicellular algae and one of the most abundant groups of phytoplankton. It is estimated that they’re responsible for up to a quarter of the world’s oxygen production and almost half of the ocean’s primary organic production. Besides this biological importance, diatoms are most famous for their magnificent geometric outer skeleton made out of silica, known as a frustule. These outer shells come in a wide variety of shapes, from squares and triangles to stars and more complex geometrical shapes.

These biological shapes are a testimony to nature’s mathematical way of working that becomes more apparent when looking at her simplest creations. In fact a mathematical equation known as the superformula is able to produce many of these shapes found in nature, and when extended into 3D can produce forms that are strikingly similar to diatoms and other simple organisms. Although the relevance of the superformula is disputed with the argument that any sufficiently complicated computational model can produce such simple organic forms, it is still interesting that it is possible for us to formulate models that can mimic nature’s mechanisms. The mathematical model is but an abstraction of the naturally occuring forms, so it can tell us how existing diatoms look like but it can also give us clues to how potentially existing diatoms would look like.

Microscopes give us only a limited understanding as we can only be mere observers that cannot really interact with this world. This world as it appears under our microscopes and under our own perception of time and space is very still and silent. Even using the latest technology in electron microscopes, images take days to produce, rendering the information we receive almost static. In the digital world, we can bring these organisms into life by setting them into motion, manipulating them and changing them. As virtual digital objects we can inject them with data such as sound frequencies and watch them react or use data as virtual forces, such as wind and gravity, that can move, reproduce or “kill” these objects. In a way, we become digital “genetic engineers” – merging the digital and the biological. Using the digital tool we get a deeper experience of the phenomenon that no microscope could ever give us. We are no longer observers, but makers, architects, where we can apply our own rules but also analyse at a greater depth the inner workings of nature, and where we can immerse our selves in this invisible world of organic simplicity taking it away from the scientific context and into a more abstract/metaphysical context. Back in the Victorian era microscopists would painstakingly arrange diatoms on slides into elaborate forms and shapes, creating microscopic works of art. A digital, virtual representation of diatoms in a way is bringing the now defunct “diatom art” into the 21st century.

A brief chronology

Following my last year’s project that focused on organic 3d forms and manipulation of image through sound, I was looking into ways of mapping sound directly on 3-dimensional forms and creating a more intuitive motion instead of plain random deformations. I started looking at sound visualisations and how people tackle the connection between sound and image. At this point I was also exploring the possibility of “freezing” time so that the 4th dimension is always visible in these visualisations which led me to the slit-scan technique which I tried to reproduce using After Effects.
In my early experiments I used simple computer-generated geometric primitives – such as spheres, cubes and toruses – that became almost alive with the input of sound. The sound used was extracts from Mozart’s Requiem because as classical music it would give more fluid motion instead of the more “spastic” results one would get from electronic music. Some of them just looked like blobs of mucous bubbling to the sound, others grew into forms of crystals before collapsing into their primitive object while others were reminiscent of microscope images of parasites and biological molecules.
This resulted in research on biological microscopy where I visited the Institute of Neuro-science to take a look at their hi-tech microscopes. Their top microscope occupied a whole room on its own and it could record video of intracellular processes like DNA-splitting or render 3D animations based on light or chemical parameters on cells. This was an amazing piece of equipment except from the fact that it would have to work for weeks on end on its own in order to get any decent results to work with. For the purposes of this project I knew it would be very counterproductive to rely on real microscopes and furthermore the people at the institute didn’t seem too keen to help.
Continuing my research in microscopy I was intrigued by the treatment of the images that were created through the various microscopy techniques. I liked how low-res and full of noise and artifacts these pictures were, and the unnatural colouring that would either result from added dye or the actual sensor of the microscope. I worked on reproducing in 3d a microscopy technique called darkfield illumination where essentially a single beam of light is shot at an angle at the object causing it to reflect details that could not be seen with normal lighting.
During my research I finally stumbled upon diatoms; these fascinating “jewels of the sea” are famous amongst biologists for their highly geometrical beauty. I also came across the Superformula, a mathematical equation that can produce 3d objects that resemble astonishingly to diatoms and other microscopic organisms. Using this formula and primitive shapes as before, I started constructing diatoms in 3d where I found some interesting parallels between their natural construction and the digital construction. For example the way we arrange textures in 3d programs mimic the way that nature arranges its textures because both are innately mathematical.
I went on to animate these diatoms using sound input by assigning different points of their body to different points of sound frequency. The result was a rather fluid and organic movement that brought these otherwise still micro-organisms to life. Although successful, these experiments decontextualised the diatoms making them look as just abstract shapes that happen to move to sound. As a solution I went on to create environments for them to “live” in. Here the diatoms were produced by a particle system that distributed them in the world in thousands with a random scale, position and rotation (within some defining parameters). Adding depth, fog, dust and terrain to the scene created an atmosphere where the virtual diatoms could become a part of.
The diatoms here became swarms that not only morphed individually according to sound but also followed the music together as a group. This was done by assigning sound frequency points to each of them and then a general sound effect to the whole swarm.
In parallel to these experiments I also worked on texturing as a big part of the diatom world is their complex geometric textures (upon their geometric bodies). Getting the textures right was very tricky as I had to use live shaders instead of pictures to avoid pixelation. This meant that I would have to create them from scratch which gave me an interesting perspective on the mechanics of nature. Texturing these organisms was all about finding a balance between perfect geometry and random organicity.
Getting feedback from my peers was important at this point to give my project a perspective and a direction. People urged me to further contextualise the whole concept and relate it to something more recognisable and emotional. The solution was drawing a parallel with outer space, comparing these miniscule organisms to the vastness of celestial bodies, comparing the far and unknown with the very near yet totally foreign. To help me with the treatment I looked at known space movies to give me a sense of how people have dealt with this vastness and scale – this resulted in experiments where I storyboarded and subsequently recreated scenes from 2001: Space Odyssey but replacing the planets with diatoms. These looked too sci-fi but served in helping me see the diatoms from a different perspective than what I was used to.
These scenes were ultimately flat lacking any engagement with the viewer which made me start thinking about the relationship with the viewer and hence the very experience of the film I was producing. In order to portray a feeling of immersion I was looking into ways of having a more active, life-like camera that would follow these organisms. The solution was to treat the camera as a particle itself reacting to the sounds and to the various forces that affect the particles. To do this I had to restructure the whole model of the environment and the particles using a programming language (Xpresso) and a module called “Thinking particles” that allows particles to communicate and interact with each other in more complicated ways. I worked with a programmer to assign the camera to a particle getting a direct view from a diatom’s perspective, furthermore using the programming language I was able to “jump” the camera from particle to particle getting totally different perspectives.
In the mean time I started collecting various water sounds from sound banks and worked on manipulating them in a sound program by layering them, removing treble and adding various effects such as wah and slow resonance trying to create an underwater effect in combination with the classical music. The different layers of sounds were eventually mapped to different parameters within the animation, for example the sound of crushing waves was translated into a wind effect that pushed the diatoms towards a set direction.
The next step was production. This involved: 1) detailed modelling and texturing of different species of diatoms , 2) assigning specific points of their body to sound frequency to create a smooth movement, 3) creating various environmental conditions for swarms of the virtual organisms to live in and assigning sound frequency to the parameters of these virtual forces, 4) creating granular sounds based on evolutionary algorithms that are used as the life force of the system, 5) assigning cameras to various particles within the environment and have them record randomly 6) using a render farm to renderhigh quality videos of these recordings.
Post-production involved the creation of "maps" and information graphics that explain what happens in each system. The structure of the final animation was divided into 5 parts or "colonies" where different interactions occur, while overview maps of the system as a whole were created to show how the system evolves.