Free the Poor Image!

graphic design, glitch, video

Design Brief:

Select a text that has personal significance to you. Compose a text-based response to your original text. Share your point-of-view. What is the dialogue between the two texts? Sketch the relationship between your texts. Combine the original text and your response into a finished whole so that a person can read all texts, leaving with a clear idea of the argument that you have made. Your final piece can use any medium. The selection of format, form, genre, technology, and context should be intentional, related to the overall concept of the piece.


I began this process by choosing the text In Defense of the Poor Image, by Hito Steyerl. The text describes poor images - those that are in constant motion, displaced from capitalist contexts and thrown into the world as “errant ideas”. This text has been a significant influence on my work to date, and it felt like a good text to respond to.

I edited the text down to the introduction and conclusion, as I wanted to respond to the statements in those specific sections. To start experimenting, I chose to the mediums of an animated GIF, Google Picture Translate, and mobile chat, experiment with formatting your text in each medium.

In an attempt to speak to the theme of the text, the GIF was made using 3 frames at a time, each one compressed more than the last. Once I started using Google Picture translate, I took screenshots of each translation and ran them through their own translator over and over, until the result was a heavily degraded translation.

Translating these texts into different media helped me better understand both the nature of the original text as well as ways in which I could treat the original text as a non-text based response. Following these experiments, I put the text into various text analysis software to get a better idea of how software encountered the words and paragraphs within the text. I used a mood emoji analyzer, a word cloud generator, and a paragraph separator.

Following this analysis, I selected specific parts of the text that I wanted to respond to and used those as a jumping-off point for generating my responses.

Text A:
“The poor image is a rag or a rip; an AVI or a JPEG, a lumpen proletarian in the class society of appearances, ranked and valued according to its resolution. The poor image has been uploaded, downloaded, shared, reformatted, and reedited. It transforms quality into accessibility, exhibition value into cult value, films into clips, contemplation into distraction. The image is liberated from the vaults of cinemas and archives and thrust into digital uncertainty, at the expense of its own substance. The poor image tends towards abstraction: it is a visual idea in its very becoming.”
Text B:
“Poor images are the contemporary Wretched of the Screen, the debris of audiovisual production, the trash that washes up on the digital economies’ shores. They testify to the violent dislocation, transferrals, and displacement of images—their acceleration and circulation within the vicious cycles of audiovisual capitalism.”
Text C:
“The poor image is no longer about the real thing—the originary original. Instead, it is about its own real conditions of existence: about swarm circulation, digital dispersion, fractured and flexible temporalities. It is about defiance and appropriation just as it is about conformism and exploitation.”

I responded to these sections of the texts with different iterations of responses and formats.

Following these iterations and critique, I narrowed down further areas that were interesting to me. The most interesting was a process different from the others (bottom right image below). This process involved opening a JPG of the text in TextEdit and responding within the text, creating a glitched, degraded image.

I felt that glitching the text, turning the text into the very poor image it talked about, accomplished my response in the most cohesive way. I used this method as the basis for narrowing down my response iterations. I drafted my full response, and incorporated the text glitching from Google Picture translate as an experiment. Following this, I experimented with displaying my response in a more traditional desktop format next to the original text. In this response, I copy/pasted each paragraph of my response and screenshotted it next to the glitching text.

I felt that the desktop format best communicated my response in a way that also gave contextual clues to the process. Following critique, I made further iterations as a video that depicted me typing each of the three paragraphs of my response in real-time as the text glitched with each response. The video, however, made the original and response text difficult to read, and was very slow.

Following critique, I redesigned the original text to improve readability direction and highlight parts of the text that I was responding to. I also broke down my responses into one sentence at a time, pasted in. Because I exported the JPG of the original redesigned text in Illustrator, the glitching effect changed to colorful lines, which inadvertently improved the readability by glitching line-by-lie instead of on the whole image. I combined a GIF of the glitching with the response window in an effort to mix the speeds.

This iteration posed more challenges: with the inclusion of the desktop came questions of performance, where the image came from, and where the image went. The response was also more difficult to read because it wasn’t very large on the screen, and didn’t correspond directly with glitching moments on the original text image. The original text redesign was also difficult to read, given its one column format. For the final version, I redesigned the original text once more to include elements of the original formatting for ease of read, while maximizing the response screen and typing my response in real-time. I types from the bottom, which created a bottom-up glitch on the original text, degrading it with my response until it was impossible to read. 


Throughout the course of this project, I felt like each iteration took me two steps forward and one step backward. This approach was incredibly helpful in getting me to the point where I felt my response was having a conversation with the original text - not just aesthetically, but conceptually as well. By turning the text into a poor image using my segmented response, I created a new poor image. In the future, I would really like to turn this project into an interactive desktop, where users can interact with both the image and response in a more cohesive way.