Brief
When we received an initial enquiry to create an illustration for a trade show wall, we were only 11 days away from the final delivery deadline. The client - MIM Solutions - focuses on machine learning and the application of artificial intelligence algorithms in medicine.
Because we found the topic very interesting, we decided to take on the challenge despite the crazy deadline. However, instead of an illustration, we decided to propose a 3D visualisation, which we felt would be more reflective of the modern nature of the client's services.
In addition to the dimensions of the visualisation and a general description of what the company does, we were given a great deal of creative freedom.
Day 1-2 - Exploring the idea
Conversations with the client and brainstorming within the company resulted in several ideas and preliminary concepts.
Concept 1
The initial idea for the graphics referred to embryo testing and DNA testing using AI. As these are two of MIM Solutions' most important products, we wanted to make them particularly visible on the wall. The DNA helix was meant to suggest the different types of analysis carried out with AI, while the fetus was meant to be an element that worked on emotions and attracted attention. The photo of the doctor or doctor and patient were to refer to the medical application.
However, after discussion with the client, we decided to abandon this concept altogether, as such an approach would have heavily categorised the company in fertility-related research, and MIM Solutions is working to apply AI to many other areas of medicine as well.
Concept 2
The second idea is a graphic consisting of several smaller visuals. In the central part would be the logo of the form in 3D, shining on a dark blue background, and from it would go paths/connections referring in appearance to the paths of an integrated circuit.
The other visualisations would in turn refer to the various medical fields in which AI can be successfully used, such as DNA research, cancer cell research, medical record keeping, embryo research or virology.
All visualisations would be separated by diagonal lines in MIM Solutions colours.
Concept 3
The third idea was for one visualisation for the whole wall. The idea was to present the whole thing as a computer board with a processor in the centre. Above it would be a luminous company logo, which would also be the AI symbol.
From the processor, there would be pathways leading to various integrated circuits, which would have miniature models displayed above them, referring to themes that AI can analyse and that are associated with the medical industry.
Day 3-6 - Modelling
Of the ideas presented, the last one turned out to be the best and it went into production. At the same time, however, it was the most difficult concept to execute, as there was very little time to produce it and there were a lot of elements to prepare, of which the visualisation was to consist. In addition, due to the large format, the models and materials had to be perfected in every detail. It was therefore very important to stick strictly to the schedule.
On the first day of modelling, we prepared the general scene as well as all the elements relating to the construction of the computer, such as fans, resistors and integrated circuits. However, we didn't want these to be exact representations of the actual circuits, but rather some idea of them. They simply had to look cool and give the impression of components that could be found, for example, on a computer motherboard.
We spent the entire next three days preparing the remaining models, which were to be the main elements in the scene. Below is their appearance without the target materials.
And here the next stages of modelling, or rather sculpting, one of the models...
... and the process of arranging the various elements in the scene.
Day 7-9 - Materials and lighting
After modelling, it was time to prepare the lighting and materials/shaders to give the final look to the models in the scene.
When choosing a renderer, we hesitated between Redshift, which we normally use for animation, and Corona, which we have so far used mainly for visualisation. In the end, due to the short time frame and our better knowledge of Redshift, we decided to use this renderer.
A very important element that we had to pay special attention to was the colour range of the scene. As the graphic was to be printed, there was a risk that colours that look great on the computer screen (in RGB space) would come out pale and flat when converted to CMYK space. In order to prevent this, we additionally checked each element already at the stage of preparing materials and lighting. In this way, we reduced the unpleasant surprises associated with converting colours for print to practically zero.
Below is a screen shot of us working on the materials:
Day 10-11 - Rendering and post-production
The target format of the work was very large (32510x6555px) and on our local computers the rendering would have taken tens of hours, and we didn't have that kind of time. Therefore, we knew we would have to use a render farm.
Having learnt from experience, we always leave ourselves a larger margin of time for rendering - that way, in case of problems, there is time for testing and necessary changes. This project was not without its minor perturbations either.
During rendering, the image was split into 24 parts, which greatly speeded up rendering, but also had its drawbacks. Among other things, we were not able to apply certain post-processing effects (such as the Bloom effect, i.e. lighting up the lights) directly during rendering. Fortunately, we were able to add these effects later, in post-production.
Below is the final look of the visualisation:
And a few fragments in slightly more close-up:
Project completion
Despite a very tight deadline, we delivered the project on time.
Below are photos of the finished wall at Medica 2022, which we received from our very satisfied customer.
Our only regret is that we could not see our work live ;)
The project in numbers
Hours of operation: 98
Number of people involved: 2
Number of models prepared: 19
Number of lights in the scene: 32 + HDR
Number of materials prepared: 97
Project dimensions: 32510x6555px
Software used: Cinema 4D, Redshift, Photoshop
Rendering time: 7 hours
Litres of coffee drunk: a lot ;)