Erik Natzke‘s generative drawing in Flash

Game Is Over from Erik Natzke on Vimeo.

manual mouse input + flash actionscript

Craig Mullins‘ digital painting

CCTV Ink TV Commercial – Directed by Niko Tziopanos

CCTV Ink TV Commercial – Directed by Niko Tziopanos from Troublemakers.tv on Vimeo.

proposal

I would like to create a generative piece that would be able to compose something similar to Natzke’s gesture drawing but with limited manual input. The goal is not to imitate traditional drawing in textures and strokes, but I do think I can learn from the lighting technique we see in a lot of Mullins’ works. It has to be apparently digital however I would like to try coding emotions into the piece. It would be a too ambitious project for one semester, but I have some smaller starting points to work from.

The final form of the machine should be an automated computer program living on the web. It feeds on its own observation and produces one drawing everyday and publishes it to the web. This process is not to mimick evolution nor to develop artificial intelligence, it would be just another representation of myself, since I would code it not to be a general purpose drawing machine that draws everything perfectly, but to draw in a style similar to myself.

It contains apparently a lot of different components requiring totally different background knowledge. I guess I will starting with building a custom stroke module that is more responsive to the gesture, the ink ad could be a nice example of such behavior. I’m not crazy about this ink effect, but it serves a good reference for the fine controls that I would probably need to build into the system.

So, yes I’m taking three design classes this semester, really pushing myself hardest ever. As a semester long project for Mainstreaming Information class, we were required to look for “jaw-dropping facts” and create visualizations to articulate. As a kickoff I created this poster to pitch the main topic of my project. I would like to look into the history of US nation building attempts and to try creating more insight by combining more historical events over the timeline.

The poster itself was a challenge for me. I have always been working on screen based media, to create a 24″ x 36″ poster for real was challenging and attempting. However on the weblog I can only show it back on screen…

Anyway, I’ve been collecting data for weeks, varying from key historical events of different nations, US intervention timeline, military spending data, and war casualty data in the last two century. It’s amazing to see so much data so easily accessible online however I’m still not an expertise in international relationship and trimming down and interpreting those fact would be most challenging for me. Also for the aesthetics of the visualization, timeline seems most promising but a little boring. I came across the code swarm project by Michael Ogawa and found it very inspiring for these kind of data. But in my case it’s harder to have well formed data with accurate time stamps. And what would be emphasized on the main stage is still under evaluation.

Will update more about this project when I make any progress.

I meant to focus more on the design and implementation aspect of the product, however there are foreseeable technical obstacles to overcome. A simplified todo list to start with

• Get the cameras work
• Pattern recognition probably featuring ARToolkit
• Hardware for audio recording/playback (optional if processing audio on a computer)
• Create toy for record/playback interface
• Figure out wireless communication for the hardware

Several factors that determine the game play:

- the stickers, cheap vs costy determines quantity. record over time vs overwrite. customizable vs preprint barcodes, or combined. RFID vs piece of paper.

- the reader & the speaker, separated gadgets for group play, more speaker than reader? more reader than the speaker, if the sticker is overwrite-able, probably a 2-in-1 would be more favorable.

- the system. on board recorder and audio player, or transferred the identified barcode to networked system (bringing more restrictions to game play indoor? )

UPDATE

Seems that RFID tags might be most suitable for the sticker implementation. Still need to find out if it’d be too sensitive to scan other RFID tags nearby? To learn more about RFID would be my next step and hopefully I could have a proof-of-concept working prototype.

Thanks Alex for the RFID book!

Prosopagnosia, also known as P.A. or face-blindness, is a symptom with which people lose the ability of identifying people by their faces. In my understanding the optimization module in the brain taking care of the facial recognition is not functioning, so that people can no longer associate the visually seen faces with the memory of people. Recognizing people in the social life becomes as hard as looking for one specific stone among thousands of them.

My proposed project is to recreate this experience for normal people by obscuring faces on the screen. Meanwhile the faces will be replaced by the texture of cloud. The idea is to create a mapping function between human faces and cloud shapes/patterns. Ideally I would be able to find a consistent conversion so that two similar faces should generate similar clouds.

some early sketches:

special thanks to “social genius”, images courtesy of ITP community

Proposal 1: The Sphere

I’m taking the visualizing the five senses class this semester and am always curious about extending the senses by means of physical computing. I couldn’t be happier if I’m able to combine these two courses to visualize the sight data and give it a more visually appealing presentation. That’s how I came up with the idea of “the sphere” (sounds like it goes well with my midterm cube).

The sphere is a toy ball mainly with a camera wrapped inside of elastic and transparent material (I don’t know what’s that yet). It should be able to be thrown between people and to bounce against walls and floors. It could also serve as a general purpose remote control.

The core of the sphere is composed of following:

• CMOS camera (which takes pictures!)
• microcontroller (controlling everything and serving the data to xbee)
• xbee module (talking wirelessly to another peer xbee connected to the computer)
• battery (power supply for everything)
• suspension system (maybe? foam?)
• accelerometer and other potential sensors that are required to fulfill the visualization
• wires … and stuff

The ideal size will be a little bigger than the tennis ball and smaller than a volleyball. It’s encouraged that people can play with the ball in pair or group. Due to the nature of the xbee communication and the fact that it does not have any external memory so it’s bounded to the computer, it’s not supposed to be a totally outdoor sport but I think apprently it’ll take much better pictures with the sunshines outdoor. Anyway the intention is to make the ball as durable as possible and involve as much people interaction as possible. So we’ll see.

The wrapping material should be:

• soft
• elastic?
• transparent (or having a channel for the camera to take pictures)
• using force/flex sensor to log enormous shape change

Basically I’m looking for something like rubber but transparent. Not sure if I can find one. If not I will also consider creating a woven structure by vines or bamboo. (pic.)

On the other side (the visualization):

• another xbee module is reading the data in real time to a processing sketch, or whatever client application that logs the data to the file system;
• ideally the data would be visualized in real time;
• the images taken are arranged in a collage in a virtual 3d space according to the position and orientation while it’s taken.
• data logged with different groups of people playing with the ball could be visualized separately and compared side by side
• other 2D visualizations are open options.

Since the sphere is designed for general purpose use, the post process could also be extended to music composing, gaming or any other applications that require multidemensional inputs.

Proposal 2: Physical Interface for a 16×16 (or less) pixel matrix

I don’t have a very nice name for this piece yet. Basically I’m inspired by a project called Shade Pixel. It looks definitely inspired by Rozin’s wooden mirror though. Similar project called SnOil is accomplished with oil and electromagnets. I would like to look for a right fabric to build this surface. Instead of the solenoids I wish to try the electromagnets to pull down the fabric surface, I really don’t appreciate that cone look of the holes. (link to yet another similar project.)

On the input side, I wish to make a website that allows the visitors to create whatever pattern within that 16×16 matrix in black and white. Whenever there is an update on the website, the physical interface morphs to the new pattern and so on. The inspiration is from one of my favorite simple site tiny.

The biggest challenge is, pretty much the same as previous one, to find out a right material, in this case, for the surface. This point is to break the boundary of the online space and make a change happened in the cyberspace directly impacts the real world. Of course if someone wishes to make it a one way instant message system I’m still more than happy to see that.