SOME RANDOM DUDE

A couple weeks ago, my wife and I took a drive through San Francisco. With us, we brought two laptops that recorded the drive through their webcams. Using Processing, the webcam video data was analyzed and reinterpreted into abstract visualizations. This project was a test-run of a larger project myself and a few friends are working on, but I thought it would be fun to show you what has been done thus far.

The concept behind this project was to take ordinary scenes from everyday life and turn them into something that would look completely alien, yet familiar enough to have the faintest semblance of normality in order for the viewer to digest what they saw. The data stream was too great for the application to read/analyze/interpret/output in absolute real time, so the videos appear to be sped up due to dropped frames. The size of the objects were based on the brightness in the relative area of the video, the colors of each shape mirrors the general color in that portion of the video. The captures were taken at night which is the reason for the strong contrast and bright colors.

Unfortunately, the encoding of the videos really hurts the aesthetics of the project, although this should not be an issue with the final product. Below are the sample test runs from each laptop’s capture.

Video From Passenger Side

A Drive Down California Street (Passenger’s Side View) from P.J. Onori on Vimeo

The video above was the main video - I personally held the laptop up to the window while my partner in crime drove around downtown. The somewhat persistent blue-gray circles are the reflective glare of the laptop’s monitor against the side window. This definitely distracts from the overall aesthetics of the visualization, but considering how often I got confronted at stop lights, I think it was a good idea that I kept the window up.

A Drive Down California Street (Driver’s Side View) from P.J. Onori on Vimeo.

The driver side video was given less attention - I essentially set another laptop down next to me and pointed the webcam towards the window. Since it was not right up to the window, the amount of light/form it pulled in was much less. I decided to give it a slightly different visual style than the passenger side visualizer.

Process

As the photos below show, the process was extremely make-shift with just simply putting the laptop’s iSight next to the window. This worked fine for a test run, although it undoubtedly looked bizarre to onlookers. This laptop was secured by myself while the other laptop was next to my lap pointed in the other direction.

Issues like excessive shaking and shifting during the braking/acceleration process created a distraction in the video. Additionally, I had to constantly shift the iSight to point outside the window. The process was sub-par for quality input/output to say the least - digital video cameras proved to be much easier to control. That being said, the ability to capture and produce a visualization in real time with only one device and no post-processing is quite powerful and shows the ease/speed at which projects like this could be done.

Look for the follow-up of this project in the coming weeks.abstract visualizations, aesthetics, bright colors, data stream, drive, everyday life, laptops, real time, san francisco, video data webcams