The "AlphabetsMagic" proyect is a series of experiments to explore and conquer the new multiverse of Typographic Artificial Intelligence led by Pablo Impallari and open to the participation of those who wish to collaborate.
My first econter with AI applied to Typeface Design was at the Atypi 2016 conferece in Warsaw. There I learned from the Morisawa guys that a Japanese typefaces sometimes has more that 26.000 glyphs, and as you can imagine, making a Japanese typeface is a massive, huge amont of work. So they experimented with a Deep Learning Netwrok to assist them, and explaimed the main aspects of it.
My mind was blown away! I wanted to do the same for Latin! But at that time it was impossible for me. Now in 2023 (7 years later) its finally possible. So I can't wait to start experimenting with it.
Here is a link to their presentation, Deep Neural Network in Typeface if you want to learn too.
The aim of the experiments is to put into practice what was previously only theory or speculation. The process of experimentation allows us to assess what is possible to do today and define the way forward. It presents us with challenges and motivates us to overcome them. In this way we can improve our skills and gradually overcome more complex objectives.
The methodology consists of defining a specific objective in each experiment, so that we can evaluate it as objectively as possible and provide feedback for the next experiment.
Here we keep track of the experiments and we host the resultsing alphabets
The objective of this particular experiment is to generate exiting, cool, flamboyant and exotic alphabets. We want to test if we can generate creative, eye catching and excentric lettershapes while maintaining consistency with the entire alphabet. In this experiment we are looking for alphabets that can be used in short words or sentences, we are specificaly avoiding the design charasteristic of typefaces designed for long paraghaps of text. In fact if we get the result looking like Arial or Times New Roman we will consider it as a epic fail.
Here are a few results. There are many more alphabets inside the Experiment folder
Jasper the Waard suggested that "I'd be interested to see more 'boring' fonts!", so here we are. This time we will do the opposite of experiment 1, We will try to generatde some standard lookiing Sans and Serif alphabets, the more boring the better. we now want restrained and legible alphabets.
This time instead of presenting the typical A-Z a-z 1-9 lettering sheet, I wanted to present something lika a mini-specimen image, so we can see how the alphabet performs when sed in Sentencecase and in all uppercase. Easier to evaluate than in the lettering sheet style presentation.
Experiment 2 is for Sans. I will show here 2 alphabets. there are 30 or so inside the experiment folder. Download it and browse freely
Same goal as experiment 2, but for serif alphabets. 19 serif alphabets in the experiment folder.
Ok.. so far, so good! By the way... you may have noticed that some glyphs got drunk? isn't it? Dont worry about it, its not relevant at this stage.
What if we have a very expresivve alphabet and we want to tone it down? Can we restraine a design without losing its essence? Let's try! Our goal for this experiment is: "Make Me Boring!"
Lets start with the typeface called Chantelli Antiqua.
Its a nice callygraphic design, quite expressive.
Lets Make it Boring!
This time the font will be Great Victorian, full of decoration.
Let make it boring!
Now its Jester turn, a casual brsh serif. Lets Try it!
Let make it boring!
Lets try Spedball, just becase I love this font so much!
Let make it boring!
I have stopped updating this readme.md file, now I'am posting and discussing all the new experiment at the TypeDrawers forum. See you there!
Cheers! Pablo Impallari