Lab 2: Public Spaces

Welcome to my Lab 2 assignment. The instructions were: create something in OpenFrameworks that has something to do with the concept of "space". I was unfamiliar to OpenFrameworks and the C++ language, so I first spend a few days exploring the possibilities. I really liked the artworks that Zach Lieberman makes with OF (https://www.instagram.com/zach.lieberman/?hl=nl). I tried to make some myself, but found out that I needed get a better understanding of the workings and possibilities first. I found that there is a free book on the OF website, that helped me with the core concepts. 

Suddenly, after having spent a lot of time thinking about it, I knew how I wanted to understand "space". The public space. These times have drastically changed the organisation of the public space. During the strict lockdowns in some countries, people were no longer allowed on the streets. In other countries, like the Netherlands, being outside was still allowed, but not in groups, and always with 1.5 meter distance between each other. I wanted to do something with this notion.

Live cameras provide a glimpse into many places around the world. I tried to find a way to grab the live footage from such camera's, and play them as video files in OF. This proved to be beyond my capabilities at this point in time. I therefore settled for a screen recording of such live cam footage. On the website webcamtaxi.com , I found plenty of live camera's, even from places in the Netherlands, like the Dam Square and the beach in Domburg. I liked the bird-eye footage from the Grote Markt in Groningen, so screen recorded two minutes and brought it up in OF.

Ideally, I would have had the people traced by some sort of tracking algorithm, and then draw a 1.5 meter circle around them, which would do something if collided with another such circle. Soon I found out that for me it would be challenging enough to get the people tracked nicely, so I first decided to dedicate my time on this. The options were OpenCV and ofxCv, as I learned from this tutorial. As I followed the tutorial, and tried to implement it in my OF file, it resulted in the unwanted situation where seemingly random objects were traced, but not the people walking or cycling:

I didn't know what I was doing wrong exactly, but decided to move on. Via the Discord Channel of the MT group, I found out that in the ofxCv add-on, the Flow example is provided. With the Flow example that is provided with the OFxCv add-on, I was able to get a nice trace of the moving subjects. However, if I were to trace them, the video would slow down to 1 frame per second:

Now, the moving objects were traced nicely and a group of white small rectangles was clouded over them. However, I didn't understand how the add-on worked , and couldn't find more information about it. That meant that I couldn't tweak the add-on to my desires. I decided to see if I could work with different desires, and thought further about what I would like to do with the traced subjects. I found a tutorial  about another OFxCv possibility that dealt with the difference between images. I tried it out. Unfortunately, it didn’t yield good results, as the moving objects were too small compared to the stationary parts. However, I found another source using the difference effect and gave it another try. 

The difference effect hardly gave any result for the Groningen footage. I figured I might try that with footage from a different Live Camera. So I went back on the live cam website and found that in Prague, a camera was low at street level, resulting in close up footage from pedestrians. Unfortunately, I had troubles with the size of the original footage, which was rather big (1800 x 1000). I knew I could resize the footage in OF, but then I didn’t know how to get the other layer (the diff in this case) to follow that resized footage. Therefore, I decided to get around the problem by resizing my footage in Premiere Pro (other online tools took too long). This also significantly lowered the file size of the video. With this footage then, the difference function gave a cool effect ( a bit hard to see, but the white lines are the outlines of people passing on the street):

Although I liked how the contours of the people were now visible, it was a bit too subtle to me. I decided to add polygon lines to the frames, such that they would be disturbed at the points where the difference lay. This resulted in the following:

Especially the moving waves made a nice effect. I liked how the pedestrians were creating disturbance in the lines. Moving at different speeds, in their own spaces. At this point I realised that I didn't need the (to me) complicated Flow example anymore, and could just base my project around the difference function.

I thought it would be nice to present two spaces at the same time, to get a more complete experience. I found a live cam directed at the Times Square in NY, which looked like good footage too, especially because here it was not so much the people moving in public space, but the advertisements. Again, I screen recorded the footage, resized it in Premiere Pro and took it into OF, where I applied the same functions to it (but now made the polygon lines blueish). Moreover, I inserted a small GUI to toggle on and off between showing the original footage:

All in all this is quite different than what I had in mind at first, but I like the result. To me it shows how (1) public spaces are accessible from an online space, and (2) how moving objects (even advertisements) can be a disturbance, but part of to the public space. 

Future directions with this project would be to, for example, explore the differences in "disturbance" between a date during the lockdown and one from the same day a year ago. I tried to find this, but it seems hard to download live cam footage that is archived. Another cool feature would still be to implement a circle around the moving objects to show how private spaces in the public space can collide. Yet another feature would be to have the moving objects leave their traces, a bit similar to the final project of a previous NMNT student: https://sites.google.com/site/nmnt2014/portfolios/rinus/final-project---trails .