BECAUSE ART: CAN MACHINES BE CREATIVE?

You’re walking through a gallery and stop to take in two seemingly unrelated pieces hanging side-by-side. one of them is a drawing of a bird, rendered with such precision its feathers could easily pop off the paper. The other is a sketch of what seems to be the same bird, however it’s nearly unrecognizable due to inconsistent line quality and parts that are entirely missing.

This article was written for the Omnibus vol #02 order yours now
In staring at the photo-real drawing of the perfect bird, you marvel over the technical ability required to produce it. You also study the sloppy sketch just as long, picking out each one of its flaws, yet decide you like the image of the strange bird because the errors are interesting to you.

When you lean forward to read the title card posted on the wall between them, you’re shocked to learn that the two considerably different images were made by the same artist; not the person them self, but a machine they built to create both drawings in two different styles.

As an illustrator, I’m fascinated by drawing machines because their purpose is to emulate an act which has always been a highly personal form of self expression for me. drawing machines and their creators are in a sense my peers.

Usually machines or robots that draw, regardless of type, are bound by the human influence of their creator. This causes most of us to see them as complex tools rather than collaborators… even though the human is more than likely just as dependent on the ability of the machine for the production of the work.

Really, the machine is a liaison between their own functional capacity to create images and the programming of their maker; a sort of co-dependent relation of ability between the two. like with the perfect bird, the machine doesn’t have to develop a greater sense of spatial acuity to draw photorealistic images like its maker would have to. It relies on programming which leverages the strengths of a computer mind, acting as a shortcut to producing the sort of realism us humans struggle with.

The trade off is in the subtlety. Machines are rock-stars at executing precision, however the important quality of imperfection found in drawings made by humans is a lot harder to forge. just like in painting programs, the simulated sable brush is only as believable as the software writer’s attention to how each fiber could potentially shift and splay in response to drag and pressure. Without these characteristics taken into account, every stroke would be a flat line, lacking what you’d probably define as feeling or expression; attributes that funnel into larger less easily definable ideas like creativity.

As the mind behind your drawing machine, how do you rectify something like feeling or gesture to produce works of expression like the strange bird?

The machine as a Tool

Along with fellow robot builders of the world, I ventured to the San Mateo maker Faire last year. Amidst the chaos I found myself lured into a booth which had a wall-hanging plotter pinned on the chain link fence in back. This particular type of robot tugged on memories from my distant past and got me talking to its creator, [Dan Royer].

I learned that [Dan] was in the business of designing robots and releasing them into the world as kits. By allowing the majority of his work to be open source, [Dan] hopes to start a collaboration with the world that will result in more capable iterations of his robots down the road. His means to an end being that these future renditions of his machines land themselves on the moon, either as capable tools or whatever they’ve evolved into by that point. having lofty goals of my own, I respect this sort of ambition.

In order to design robots for a living, one must start out selling robots for a living. [Dan] leverages the sales of his self engineered creations to afford a life style of perpetual development. because of this however, the element of production has a heavy influence on the way he designs his machines. Where 3D printing is a wonderful tool for rapid prototyping, it no longer becomes viable when you have to create a hundred of the same part.

The solution for [Dan] in handling all of his own manufacturing was to invest in a laser cutter. With this tool he can quickly create complex and dense shapes by stacking layers of thinner material together… like a robot sandwich.

This method is implemented expertly with his 3-axis arm kit. dozens of detailed cross sections pile together to create a robot that does more than just pick things up and place them somewhere. in spite of operating radially, it can produce accurate drawings too.

The arm rotates on a circular platform, so the use of X and Y coordinates would result in an image that looks like its been mashed into a funnel. To remedy this, [Dan] implemented inverse kinematics for the jointed robotic arm in order to create images that are proportionate.

Ee arm was doodling away while I talked to [Dan], but my attention was glued on the wall plotter a few feet away. This particular robot was his first success, dubbed the “Makelangelo”. Its creation was an indirect result of [Dan] teaching himself how to control stepper motors. In order to gauge accuracy, he would compare the axis of one motor against the other to see if they could consistently land at the correct spot. naturally these sort of tests evolved into the production of drawings. If the motors could plot a series of coordinates over time and turn out the expected image rather than a rat’s nest of scribbles, then he knew he was making progress.

From the milestone of accuracy, [Dan] continued to develop other drawing styles for both the Makelangelo and the ARM. like filters in Photoshop, these styles can be applied to any image and the plotter will act on code to re-skin the outcome.

The goal for much of [Dan’s] work is to create a platform capable of producing works like the perfect bird, so that others can use this conventional as a starting point. like crafting an optimal paintbrush that the owner can then pluck bristles from to create an alternate stroke.

When designing the machine, what if perfection was never the point to begin with? What bountiful coolness could the pursuit of the unexpected yield?

The machine as an Artist

At some point during my expedition in art school years ago, I diverted from illustration and threw myself into a robotics class. This being my first run-in with electronics ever, it felt a lot like taking high diving instructions without knowing how to swim.

To stay afloat, I channeled inspiration from the other seasoned tech veterans. For instance, the guy who sat across from me, [Harvey Moon], already had a reputation for designing and fabricating his own drawing machines. His work became the object of my fascination because I saw him as an inverse of myself; I being an artist that drew pictures of robots, he being an artist who made robots that drew pictures.

[Harvey’s] drawing machines were in a league all their own because they seemed to effortlessly pull off a sense of human-like feeling and style, which was a quality granted by the imperfections in the images they produced… imperfections that appeared genuine, not canned or determined ahead of time.

By surrendering the outcome of the drawing over to the machine itself, [Harvey] gives the machine permission to have its own creativity. Ouais. let me tell you how.

The Wall-Hanger

The project [Harvey] was known for by the time I landed myself in his presence was a wall-mounted robot. It was deceptively simple, but the elements were expertly developed to work exactly as he wanted them to.

This little machine created lines inherent of the delicate tension between pen and paper like a seismograph. One continuous trail meandered in and around itself, becoming an image that lacked the precision indicative of machines while also remaining impossible to replicate by human hand.

Almost everyone is familiar with plotters of this flavor. The kind that tether a pen on a string between two stepper motors… acting in cooperation with gravity, like a spider weaving its web. They’re fairly common these days, but [Harvey] started building his first prototype at a time before they were as heavily documented on the Internet as they are now.

Like the rest of us students in fine art, [Harvey’s] original passion was in a more traditional medium, photography. Without anyone else’s prior experience to use as a starting point, he had to begin from scratch, teaching himself the ins and outs of stepper motors by salvaging them from old printers.

This being his first project involving microcontrollers and programming, it would be years before he figured out how to coax a stepper into producing a drawing. Along the way, the dialogue of trial and error between [Harvey] and his work became integral to the meaning behind it.

He came to realize that he was more pleased when his machine failed than when it worked as he expected. If the output was always unknown, there was no way to grow bored of it. due to this attraction to chance, it was never his intention to produce a machine that drew perfectly. He set his conventional elsewhere.

But alas, you can’t really force a “happy accident”. The better [Harvey] became at his craft, the more work he had to put into engineering unpredictability. His way of pulling off this trick is through the use of clever algorithms.

Where most wall hanging plotters are given a starting image to copy, resulting in a more or less predetermined output, [Harvey’s] algorithms allow the machine to carve the path as it’s being walked.

To give an example of how this could work, the machine is similarly provided with a source image, however depending on what pixel it’s told to begin the drawing at, the result will be different each and every time. To pull this trick off, a program in Processing converts all of the pixels de l’image en pourcentages de gris. Du point d’origine choisi, la machine se poursuivra vers le pixel le plus sombre de la proximité, tout de suite, le fouettez-le à l’arrivée afin qu’il ne trachait pas une seconde fois.

Cintre mural Taiwan

Comme son cintre mural a changé d’itération à l’itération, [Harvey] a continué à expérimenter avec plus de moyens de rendre son contrôle sur le résultat du dessin. Lors d’une exposition plus récente au Musée national taïwanais de l’art, il a décidé de saisir l’image statique comme source et d’introduire l’utilisation de la vidéo en direct comme moyen de récolter une entrée imprévisible.

Pour cette installation particulière, [Harvey] localisée des caméras de surveillance en direct à l’avance dans toutes les taïwanes dans des zones de densité de trafic variable. Pendant trois mois, quatre des machines de dessin [Harvey] se sont lentement écoulées, essayant fidèlement de reproduire la vision toujours changeante vue par la caméra. L’image résultante était une représentation originale du passage du temps créé par le caprice de la chance.

À la troisième dimension!

En pensant à de nouvelles façons d’abandonner le contrôle du contenu produit par ses machines, [Harvey] considérés comme des moyens d’apporter une dimensionnalité dans l’équation de son travail. Ces pensées ont abouti à la construction d’un robot delta en tant que nouvelle plate-forme d’expérimentation.

Au lieu de se concentrer tellement sur la contribution de la production de hasard comme il avait avec son traceur mural, le robot delta a ouvert de nouveaux terrains à désordre avec différents médiums comme moyen de créer l’imprévu.

Bien que les robots Delta soient associés à une précision de qualité industrielle et utilisés comme imprimantes 3D pour cette raison même, [Harvey] voulait briser l’association commune.

Conformément à son désir de jouer avec de nouveaux médiums incontrôlables, il a construit une extrudeuse pour l’effecteur de fin qui vomirait la colle chaude au lieu de filament. Vraiment, il n’y a rien de moins contrôlable que gooey, à la ficelle, à des monticules complètement irréprodictables de plastique fondu refroidi.

Pour finir par ces piles intrigantes de Semitransparent GOO, [Harvey] programmée d’un générateur de forme aléatoire dans le traitement qui fonctionnerait en temps réel, produisant du code G de manière génératrice.

Alors, outil ou artiste?

Comme l’esprit derrière votre machine à dessin, si vous produisez un sentiment de sentiment ou de geste consiste à renoncer autant de votre propre contrôle que possible … à quel point votre machine devrait-elle recevoir un crédit pour ce qu’il crée sur vous?

Je pense à la situation hypothétique où [Harvey] s’éloigne de son installation à Taiwan pour ne jamais retourner, et comme un passage d’un conte de science-fiction, la machine se poursuit pour produire des dessins pendant des décennies maintenues par des personnes dans la communauté qui ne se souviennent plus lui.

Avec [Harvey] jusqu’à présent retiré de l’équation, qui produit l’art à ce stade? Est-ce la personne qui réapprovisionniste le papier? Ou les personnes sur l’alimentation de la caméra en direct fournissant l’entrée? Ou est-ce toujours [Harvey] parce qu’il doit être crédité pour l’idée originale?

Si la propriété appartient à la source?

Imaginez ce scénario. Si vous imprimez une copie d’une peinture que vous avez faite, vous attribuez un crédit pour l’art et l’impression à vous-même, pas le créateur de la machine à imprimer ou de la machine elle-même. Vous allez faire le tour et montrer aux gens la copie et se vanter, “hé regarder l’illustration que j’ai faite!”. Vous assumez la propriété de la créativité, de la capacité et de l’intention nécessaire à sa création, même si l’encre sur cette feuille de papier n’était pas appliquée techniquement par votre main.

Toutefois, si au cours du processus d’impression de la même image, le papier s’enfonce et frotte l’encre dans tout le lieu de manière intéressante, que vous accrochez ensuite sur votre mur parce que l’accident était plus intéressant que l’impression prévue … alors quoi?

Le mouillage barbouillé était basé sur l’entrée que vous avez fournie, sur une machine que quelqu’un d’autre a inventé … mais le produit était vraiment un résultat de son propre fait. Bien sûr, il n’y avait pas de conception intelligente derrière ce qu’elle a fait; Aucune – moins elle agissait sur des données fournies, sa propre capacité fonctionnelle et le RI