Forget-Me-Nots
The story behind the work
The idea for the project
Before I discovered generative art, I was already a full-time artist and created works with oil paint, gouache and charcoal.
With these works, I participated in some group exhibitions in my country (Brazil), Portugal, and the United States.
So after some time already introduced to generative art, I began to be interested in creating an algorithm that simulated brushstrokes, combining traditional and digital.
It was then that my first project with a traditional painting theme was born, called 'Water Lilies', where I represented Monet's famous Water Lilies. After this project, I started several tests looking to visually improve the idea of brushstrokes, and little by little I got to where I am with 'Forget-Me-Nots'.
Representing flowers in my work is not something new for me, in addition to the portraits I did, I have always liked painting outdoors and I love representing nature in my own way, as I see it.
'Flores no quintal', 2020, oil on canvas
'Water Lilies', 2023, p5js
Building the algorithm
I started the project by testing a way to simulate a single brushstroke at a time, so that I could use the same technique in future works, in addition, I also created it in a way that I could reproduce some desired outputs in physical form in future projects, using a brush, paint, and paper, on a homemade penplotter that I created following an internet project! This is still in testing at the moment, but with great progress!
One of the first outputs of Forget-Me-Nots, and next to it, my penplotter
Here things started to take a better shape, the brushwork no longer seemed as artificial as before, which really encouraged me, as I was on the right path. I also started testing light/shadow to give a more realistic effect, but I ended up changing this first method throughout the project.
I started testing new shapes for the flowers, as well as starting to create a vase to give more vision to the composition.
Throughout the project I decided to make the brushstrokes that would be for the vases more uniform, without so much use of random() in their shapes, this brought more fidelity to what it would be like in a painting.
In this image I'm still using a lot of random() in the vase's brushstrokes.
Starting the position of the flowers to put together a composition, also starting the planning and testing of the backgrounds. Was one of the most complicated parts as it is a space that would have less details, where the brushstrokes appeared a lot, and as they were still very raw, the background ended up bothering me, but this was solved over time and with the refinement of the algorithm, the brushstrokes even had a transparency that during the course of the project was removed. I also started creating some new palettes to change a little, I was tired of seeing only pink haha.
On some days the changes in the outputs were not that huge, as they were just hours of refining what was already done, always trying to achieve the best result.
I also started some new shapes for the vases and flowers, in addition, I started implementing shadows, and also a texture using noise() to simulate the effect of painting paper.
In the first image, I'm using new color palettes and adding leaves to the flowers, and in the second image, starting the background variations. Here a new variable was also inserted into the project, which are the petals falling on the table, which may or may not appear in 1 or more petals. Its color is influenced by the shadow of the vase, and if the vase is between the petal and the light point, it will be darker.
One of the tests trying to give a more realistic look to the brushstrokes during the project progress, I used the color black during the tests to clearly visualize the shape it was taking, after that I changed the color tone to something more natural. The idea was to bring volume to the brushstrokes, using light and shadow.
Notice the difference in the orange flowers in the image above, and the yellow ones in the image below, with this new method included.
During the process I tested new ways of aligning the brushstrokes, I tried something more symmetrical, but obviously this didn't fit with the theme, the brushstrokes needed to be looser, just as they are naturally. The algorithm would still undergo many refinements, as the shadow of the brush strokes was not yet as good as I know it could be.
As time went by, I continued to improve the algorithm, I managed to make it look more natural than the initial outputs, and that simpler and empty background, which I didn't like so much before, ended up becoming one of my favorites, as it shows many details of the brushstrokes, which are the soul of this project, in addition to the beautiful flowers, of course.
This was one of my longest projects to date, I intend to further improve this algorithm to use it in future projects. I will continue testing with paint and paper on my penplotter to complement new work that is yet to come.
It's been a long and hard journey so far, but very rewarding, and I hope this work touches the hearts of everyone who sees it and brings some good feelings to the viewer, I hope you like it! :)
Final thoughts
'Forget-me-nots' is an important part of my development as a generative artist, and it will have a strong connection with my next work, which will be released on BTC by Genify. With the same style seeking to simulate brushstrokes, I will bring a work even more personal and deeper, with one of my favorite themes.
If you still don't follow me on X (formerly Twitter), consider following to stay up to date with the latest news about future projects!