I spent a majority of the time between Summer 2023 and Summer 2024 learning how to code, train and use generative ai models.
As my art practice has become primarily traditional, I became interested in using algorithms to turn my drawings and paintings into moving images.
The following works are a collection of artistic experiments, geared towards deepening my understanding of generative filmmaking and its intricacies.
The works are ordered from new to old and span over about a year.
Dip:
The idea behind ‘Dip’ was to explore frame interpolation, trying to understand how algorithms create new frames based on given ones, especially when handling transitions.
The frames were purposefully kept simple, and the narrative gaps between images wide. This was done to push the limits of the tool and see how much work is necessary.
Virginia Creepers & Polyester Triplets:
Both ‘Virginia Creepers’ and ‘Polyester Triplets’ are moving image works based on image to video tools which use an image as an input and create motion from it. The input images, both mixed media works I created in the past show, led to fascinating results.
Treadmill:
‘Treadmill’ is a moving image piece based on a custom ai model I trained, aiming to replicate the style of my drawings and applying them to input footage in the form of a style transfer.
Though the colour-palette seemed to translate to the input video, the line-art was still too close to the base model I used as a training reference. Due to the immense resources required to train custom models at that point in time, I decided to move on to swifter tools. Thus resulting in ‘Virginia Creepers’ and “Polyester Triplets’ soon after.
Chanelling Red & Blue:
‘Chanelling Red’ & ‘Chanelling Blue’ are my first ever completely rendered generative films. Both works serve as predecessors to treadmill, as I have attempted to follow the same idea, training a model on my art and applying it to reference footage. Even more visible here, is the influence of the base reference model I used, thus unfortunately deviating from my personal style.
Ultimately, though abstract in nature, these experiments provided an immense level of new found knowledge and understanding. They allowed me to begin working on novel pieces of work.