Project 4: Event Poster Series
You will create a series of three posters that visually introduce three different arts events (e.g. musical concerts, plays, film festivals, dance performances, etc.) held within a single arts festival. The three posters will be distinctive, yet use some common design elements to unite the collection to communicate that the events relate to each other.
Each poster needed to have the following content:
The Name of the Arts Festival (this should be on all three posters and may be fictional.
The Title of the Event
The name(s) of the artist(s)
One (short) paragraph about the artist(s) or the event itself
The date, time, and location of the event (fictional)
Relevant ticket information including price, how to purchase, etc. (fictional)
I: Ideation & Hand Sketches
Given AI’s rapid rate of development today, a future where an AI is a standalone creative entity (as opposed to only being a tool used by people) isn’t too far off. This is precisely what I wanted to explore in this poster series.
What year would an art show with AI-conceived art take place? What would their (it’s?) art look like? Who would host such an art festival? These were all questions that I began to ponder as I sketched out basic ideas for the posters.
I quickly realized that some of these questions would have no answers, so I began to fill in the gaps. As for what the art would look like and who the “artists” would be, I decided to reference today’s ML models by using network topologies as the names and their outputs as the basis for the art. Currently, I’m working on an installation using a Composition Pattern Producing Network with ResNet, DenseNet, and Perceptron topologies. Here is what the outputs for those look like:
However, these images are quite hard to replicate using the Pen tool in Illustrator, so I modified the networks to produce very simple outputs or ones with only geometric shapes. Here were some of the results from this exploration:
On the left, I tried producing only outlines of black or white shapes using the algorithm. On the right, I had only shapes in gray scale, from which I made a video exploring the three-dimensional space generated by the algorithm. Finally, I created some simple gradients, and more geometries. Color was omitted from these images, since I felt that should be motivated by the design (not arbitrarily, or by the algorithm).
II: Digital Iterations
First, I began exploring different layouts and graphic ideas separately:
Above are some of the ideas I had for the layout of the text on the poster. Note that the ‘ists’ is in white, which is only barely legible because the grid is being displayed. This was one idea I had for the event title in my sketches that carried through to my initial posters, except instead of using the grid I decided to add a light drop shadow behind the letters. This way it might blend in to the background at a passing glance, but would be noticeable to a viewer up close. With this idea, I began combining the text with graphics I created based on the algorithm’s output.
My idea for incorporating colors was to create a contrast between the mechanically-generated images and a color palette with natural/human tones (skin colors, tree colors, beige, brown, green, ect). However, I found that (like that iteration above-right) these muted tones did not contrast well with the art. I also felt that the graphic by the different topologies looked far more different than I originally imagined, which made the posters seem visually unrelated. One solution to this I came up with was to make base the shapes for all three posters on one image, so the edges would like up nicely. Below are some of the very wide test images I created for this:
Another problem I was having was displaying the description text in a way that incorporated it into the graphics cohesively. To solve this, I decided to abandon my strict adherence to formatting it in a big chunk that didn’t quite fit within the art. Instead, I could place it along the art to mimic it’s patterns an shapes.
After adding bright orange to my color palette to create more contrast, removing some of the clutter the complex graphics were adding, and breaking up the text in a playful way, I arrived at this iteration:
Note that I removed the drop shadow from ‘ist’, and opted to color it similarly to the background to achieve the same effect. After taking a step back, I felt that a lot of the changes were moving the poster in a good direction. However, I also felt that my original idea had been lost somewhere along the way. I believe my strict adherence to the original graphics was what held my design back in other areas (like color palette, hierarchy, ect). So, I decided to go with a style that was still referential to the original idea but wasn’t strictly derivative of it.
III: Back to the Drawing Board
After taking some time to think, I remembered Julia had suggested making each poster monochromatic and using the three colors in that way. So, I decided I could invert my idea for the poster by creating graphics that looked like nature and using natural colors to convey this. Then, I could use the words and ideas in the poster to contrast with the background. To create this effect and still reference the original idea, I used the same shapes the algorithm had produced before and layered them on top of each other. In this way, I created a gradient of one color that began to look like the geometric landscape I was hoping for:
After some iterations exploring various levels of irregularity and different colors, I arrived at the final geometric landscapes I would use:
Then, I began adding text text back into the mix:
At this point I realized that I should rethink my use of typography, since I felt that using Helvetica and Garamond was a bit generic and not related to the poster idea. After some experimentation, I also decided to go back to vertical poster so that I would have more negative space to work with. Lastly, I refined the placement and hierarchy of the text to achieve the final result.
IV: Final Iteration
Overall, I learned a lot about my limitations as a designer through this project. By failing (miserably) with my first iterations, I was able to make something I would have never imagined at the beginning of the project. I believe this discovery made me a much better designer, but also made me realize that I still have a lot to learn.
