Surefire mechanism

The University of the Aegean was home during studies. The 3D graphics lab assigned us the task of developing a 3D animation. In this video, a 3D model from the real world, a Japanese ball bearing, is typing the word deliverance on a cement floor. As a young child, I always received little gifts, promotional keyrings of ball bearings. I loved mechanisms from an early age. Growing older and living a motorist life taught me all kinds of lessons.

At the time of the development of this video (more or less 2004), I bought my first motorcycle. A good friend of mine bought a new bike too. We had a blast until October 2005, when I experienced my first serious accident. Succeeding the same month, I lost my close friend in a motorcycle accident. This experience influenced a lot my perspective of motors. A rider must be cautious. Any mistake he pays. I didn’t know that I had to predict others behaviour since driving two wheels means paying mistakes of others too. During my recovery, I took the time to repolish the project and add some extras to the 3D scene.

Technically this project brought many difficulties. Processing power at the time was expensive. The application required a powerful graphic card using OpenGL to deliver the results. There was always the possibility of network rendering to get faster results. However, it’s more fun to see what you can achieve with your custom build. I did whatever possible to exploit my custom machine properties. Nowadays, components have overwhelming computing power compared to their cost. Back then, I had different tools available. I took the time to assemble custom components with whatever I could afford at the time. Building it myself would save me some money.

Anything related to audio and visual production comes highly expensive. In the end, it depends on how you make the best of it. I had to prepare the scene carefully since it contains a lot of reflections. Two mirrors, a metallic surfaced ball bearing in the middle receiving light from a high-dynamic-range image. Casting light inside the scene was consuming too much computing power. I had to select and prioritize surfaces receiving light and remove unnecessary ones. It was also a prerequisite to prepare carefully scene properties. I had to deal with frame lag and many other technical difficulties. Inside the application, you learn once more the value of prioritizing resources.

It took thorough research to select the components to build the machine. The magic in making your production machine lies in learning more about technical specifications and building a balanced machine. After careful setup of the scene and mindful render settings, I have managed to render this video in a resolution of 1920 X 1080 and 29 frames per second with progressive scanning in 96 hours. It felt great to develop a small render station and manage available computing power. I have added the shadow of the ball bearing, and colour graded the video in post-production.The sound selection was also a task. In the beginning, I was searching for ball-bearing sounds. I have tried different sounds of mechanical components, but nothing fitted the image. The truth is that a ball bearing makes no sound. Ball bearings produce friction sounds if there is something wrong with their mechanism. It would be more evident that the circle’s fire would make a sound. The metal paving the cement floor as it types the word would also produce sound. The mechanical clicks should also be audible since the metal plate with the typographical elements locks and follows the path of the ball bearing to type the word on the cement floor. After rendering the picture, I asked a friend permission to use an old composition he made. I think it fits well on these images. Reverb increases the feeling of space to the viewer.

Another challenge was mixing all of these sounds. According to how close the camera is to the ball bearing, there are differences in distances inside the scene. I have used sound effects in this video from freesound. A great project which I would recommend to anyone involved in production.

Computers cards follow Moore’s law. They get smaller and faster. According to big tech companies, they will follow this path until the end of 2025. Management and prioritizing resources is crucial in any case. Even if they get faster, the necessity of moderation will be present.