ADAD1001

my final work was a computer program which generates an image and reads out instructions for a human to follow to produce the image. the above video shows me carrying out this process.

I can’t think of a good name for it and I don’t like bad names so at the moment it’s nameless.

here is a photo of the finished image. I made a mistake in following the instructions but it’s hard to tell.

after doing a demo with only around 50 lines I realised I would not be able to have more than ~200 lines in the final image because it just takes too long. I tried simplifying the abstract landscapes but they don’t look as nice:

so I tried a different approach for generating the image:

took way to long to write the code (it always does) but I like how these turned out. each time the program is run, it checks the state of the computer (how many programs are running, memory usage, etc.) and the geometry that’s generated is roughly determined by those statistics. trying to get at this idea of the machine representing its own experience.

at the same time I also like the idea of the image being something recognisable, so you’re drawing out this thing line by line and you have no idea what it is and suddenly it comes together and becomes a face, or a house, or whatever. still thinking about 3d printers and this weird liminal state where half of the object exists there on the platform but the other half exists only in some obscure digital representation.

but I think this kind of works against the core thing of trying to reframe the human/computer dynamic. at least, I can’t really see a way to resolve the two.

demo run using a simple scene. object is generated and rendered in Blender (with some custom Python):

decomposed into a series of instructions

and manually transcribed to graph paper:

I made a mistake :(

it took a lot longer than I’d expected to draw all the lines, even for this simple scene (56 lines). for my final work I want to video the entire process, and given memory limitations on the video camera, I think I will have to do something relatively simple so it doesn’t take me too long to draw.

finally figured out how to convert a 3d render into a series of lines, using the ‘export to SVG’ feature that just got added in one of the newest versions of Blender. unfortunately it’s not perfect, which means some of the lines get screwed up when quantizing to a grid, e.g.

experimenting with rendering randomly generated 3d landscapes as line drawings

just messing around, made with Python.

fleshing out the idea I had:

I think I want to have the computer generate an image (according to some procedure which I haven’t yet figured out), decompose the image into a bunch of lines, and then print out (or 'speak’?) instructions to draw each line. and then I’ll follow the instructions to draw out the image line-by-line on graph paper.

I think the best way to present this would be video. I want to try and capture the feeling of watching a 3d printer where you have this object being constructed by a machine that has no understanding of the object’s meaning, only its form. I think what I will do is video myself drawing out the image line-by-line, so that you can see the process involved as well as the incremental construction of the image. would probably have to speed this video up for presentation in class.

what I am not sure about is what the nature of the image itself should be. this detail wasn’t a central part of the concept I had but ideally it should resonate with the theme. I want the image to use some degree of procedural generation, since it’s supposed to be at least nominally the computer’s creation. but to what extent should the image reflect a human aesthetic as opposed to a machine one? [for example, a photograph is very much a human artifact, whereas the wi-fi data flowing in and out of my laptop lies much closer to what we might call the machine’s experience. Ian Bogost on Alien Aesthetics: http://www.theatlantic.com/technology/archive/2012/04/the-new-aesthetic-needs-to-get-weirder/255838/]

on one hand creating a totally non-human image would be a more thorough inversion of the typical human/computer relationship (as in 3d printing etc): computer uses human to produce object which is meaningful only to the computer.

on the other hand, this would result in imagery which by construction would be alienating to human viewers, and given that ultimately the work is for a human audience perhaps this would be unwise.

I had 2 ideas I was thinking about, the first one was looking at how google search results and suggestions reflect (and promote) racial prejudice: e.g. if you search ‘Congolese’ on google images the first suggestion is ‘cannibals’. just looking at the first page of image results for different searches (’Sudanese’, ‘American’, ‘Italian’, etc) paints a pretty striking picture of the hegemonic lens through which we view different groups of people. I thought this was a good idea at first but now I’m not so sure it’s really that interesting.

2nd idea comes from thinking about how much people use computers now in art-making, how computers are getting better at making art ‘on their own’, and the increasingly complex power dynamics between people and computers. what are the implications of ‘outsourcing’ much of our art production to computers? what does it mean for humans if we can one day (soon?) create machines that are better at making art than us?

I want to invert this relationship, i.e. frame the computer as the designer and the human as labourer. so something like having the computer generate an image and then reproducing this image manually. still thinking about how to do this in an effective way.

inspiration:  - the genre of algorithmic / computer-generated art  - watching 3d printers build stuff. the dissonance between the object’s *meaning* and the mechanical, indifferent way in which it is constructed is interesting to me.  - American TV studios outsourcing animation work to Asia; art production as labour  - semantic style transfer: people have taught computers to recognise ‘style’ in paintings, something perhaps previously seen as fundamentally non-algorithmic.  - Bruce Sterling’s idea of ‘alien aesthetics’: what kind of art are alien consciousnesses (like AI) interested in creating? Google’s DeepDream images fit into this.  - more generally, the idea of deprivileging human existence as opposed to that of nonhuman objects is one that I find deeply fascinating, for reasons I can’t really articulate.

hmm idk which theme this would fit into best. I guess ways of seeing in that it might suggest a different framing of the relationship between humans and computers.

obFuscator is a virtual reality experience taking place in a small bedroom, around which the player can look, but not move. Over the five-minute duration, the forms and textures in the room gradually degrade, stripping away the illusion and revealing the underlying fabricated nature of the experience.

The audio consists of two parts. The first is a composed musical soundtrack which increases intensity as the environment collapses. The other is a recording of ambient noises on campus, which becomes gradually more noticeable over time. The ambient noise contrasts with the virtual environment and serves to further disassociate the player from the constructed reality.

Towards the end of the experience, some of the textures switch to a live webcam feed of the actual human player wearing the headset. This creates dissonance by forcing the player to acknowledge their simultaneous occupation of both the physical world and the digital one.

The videos that display on the television were chosen simply to suit the atmosphere that we were aiming for. They are:

1990s KCRA news broadcast

https://www.youtube.com/watch?v=RYkTFhYUtgg

Collection of 90s TV ads

https://www.youtube.com/watch?v=fAMwriy2OAI

The Land of Pleasant Dreams: A Fence Too High. Bridgestone Multimedia Group 1991.

https://www.youtube.com/watch?v=UchOmI7O2Z4

Beyond The Mind’s Eye 01: ‘Virtual Reality’. Odyssey / Miramar 1992.

https://www.youtube.com/watch?v=mdFfVx-LkN8

A few words about experimental process I followed, and the end result, my game 'And when your goal is nearest’. Some spoilers – play the game first! (https://xq.itch.io/and-when-your-goal-is-nearest)

I wasn't happy with many of my experiments because they weren't honest – I was transforming boxes, but not communicating anything that I actually wanted to communicate. Considering this realisation to be the most valuable outcome of the experimentation, I wanted my final work to actually reflect my own creative vision.

For my final work I created a game I’d had in my head for a while, and which I partially implemented in my third experiment. The game places the player in a surreal landscape, dominated by a mountain with a large white box at the peak. Upon ascending the mountain the player can transform – improve? – the box by plastering it with symbols either of freedom or of imperialism, depending on the player.

I did most of the work – lighting, texturing, environment design, mechanics / programming – in Unity and a little bit of Blender. The trees are modified slightly from the ones that come with Unity. I made the music with my guitar and Ableton Live and this sick ‘sound effects’ LP I found at a garage sale once.

building on the results of experiment #3, I made a short game/dream thing, called ‘And when your goal is nearest’. you can download it here for Mac OS X or Windows: https://xq.itch.io/and-when-your-goal-is-nearest

if you really don’t want to download it I made a video of me playing through it. but don’t watch this, just play the game.

https://vimeo.com/160085567

experiment no. 4 ‘übermensch’

human figure in MakeHuman, animation + render in Blender (Cycles)

REFLECTION

I wanted to do a more literal transformation of a box than previous experiments. I started off just with a box morphing into a man, but I discovered that the transformation factor could be extended beyond 100%, taking the form ‘beyond’ human. I thought this was a cool concept and a nice outcome of practice-led inquiry.

since I spent longer than I would have liked on previous experiments, with this one I experimented with my own creating process by limiting myself to 1 hour. this worked out pretty well, there were lots of things I didn’t have time to try, and I couldn’t give much attention to presentation, but it meant I was more focused and worked more efficiently.

experiment number 3

I stuck freedom flags on a box and recorded it, it’s called Resurrection

world made in Unity, sound in Ableton

REFLECTION

I was pretty happy with this, it’s the only one that was actually something I wanted to make rather than just making something for the sake of the assignment. my original vision was a bit more extensive so I think I will expand on this for my final work. but there’s also one other box-related thing I want to try making first.

to me the phenomenon of unboxing videos is kind of absurd, and a little disturbing. it suggests we’ve raised the status of the product beyond something to be used, to something to be revered.

I tried to reflect these feelings, but I’m not sure I succeeded.

PROCESS

first thing I did was to download the first page of youtube results for ‘unboxing’. I found a tool called youtube-dl which makes it easy to do this from the command line.

to do the actual video manipulation I used a command-line tool called ffmpeg, controlled from a script. this approach gives much more flexibility than using a graphical editor like Final Cut or whatever, because you’re not limited to performing manipulations that are built into the editor.

I experimented with a bunch of different ideas – presenting the videos simultaneously in a grid, or overlaying them. In the end I went with the following process:

 - scale all the images to the same length (260s)  - alternately take 40ms chunks from each video to construct a new 260s video

REFLECTION

good things:

 - I thought it was interesting that although the video is incomprehensible, the brand iconography is hard to miss – and isn’t that really the whole point of the videos? so in a way mashing all the videos together distills their real essence – brand promotion – and makes it more obvious. this wasn’t something I was deliberately trying to achieve, but it was a nice outcome.

 - it was fun experimenting with different ways of manipulating the video without having any clear aim in mind. one thing I like about computer programming as a creative tool is that you can make modifications to code without having a clear idea of what the resulting change will be, sometimes leading to unexpected discoveries.

bad things:

 - I don’t think the video really does a great job of communicating what I had hoped to.