Sarah Youen

I found this project very personal to me as it helped me understand my disease much better. By visualizing my blood sugars in a more simulating way than just numbers on a machine, I have learnt further how to keep them under control. Seeing my results as an interactive video experience, I could understand that for two days my blood sugar has been high and thus correct it.

Even though I understood what was going on I decided to ask some people if they thought this project helped them to understand their blood sugars some more. Here is what they had to say:

“I felt quite ill when the red video was playing, so now I know what it is like for Sarah to be in a hypo state.”

“It was a very clever video and to know that it is based on real time data is very cool and dynamic.”

“I felt like I went on a journey and helped me to understand diabetics a lot better.”

There were errors with the program, one being that you do need a powerful computer to run the sketch. If it is run off a memory stick, the videos lag and don’t play around the 4th one. However on the right computer, it works perfectly. Here are videos of the full and quarter versions running in the dome.

.

So where can this project go from here. These are some the ideas I have come up with as an extension to this project.

• Make an online based version of the program so people can view the dynamic video around the world.
• Allow people to integrate their own twitter feeds into the program and see their own blood sugar journeys.
• Eventually allow the user to record and download the videos to show a particular journey or sequences

I am very happy with the outcome of this project, as it has helped to explain to people what it feels like to be a diabetic. I hope to expand upon the idea in the future and try and get more people to collaborate with me.

Video Lag

When I came to run the full resolution version of the program, I found that the video lagged on my machine. I thought a more powerful processer would help this issue. However when testing it in the inflatable dome, I found the computer still wasn’t powerful enough to play the videos through processing.

.

After a discussion with Simon Lock, we managed to come up with an alternative idea still using the code from my processing sketch. Every time the If statement is called by the code, rather than triggering a video, a key press is generated. For example when the sketch registers a low blood sugar level, it will give the impression of someone pressing 1 on the keyboard. It is achieved by putting an object known as a robot in, and in this case he is called ‘Robbie’. Here is example of the program working below:

.

.

So how we can use this sketch in collaboration with another program e.g. Adobe Flash or Quartz Composer, as they are better than handling high quality video. It would work by having these programs listening for the key presses and reacting to it. I will look into creating it in Quartz Composer as I would like to become more familiar with this program, and I am already confident in Flash should it not work.

.

Lack of Internet 

Another flaw in my program is that it will not work without the computer having internet access. This is because it reads in a real-time feed into the program. There are two ways around this. The first being a back-up video on hand to play if I can not connect to the Internet. Secondly, to create a version that I control, so it is like a live-performance done by me, rather than the RSS. I like the idea of the second idea as it still make the experience interactive for the viewers and still has that live element in it.

.

So here is a video the alternative version I have created and me controlling it.

.

.

The final step is to test it in the main dome.

Below are the following dome corrected videos, now with sound. They are 1400 by 1050 to match the dome’s resolution. I slowed down the spin on the low video as people felt too much vection, so now it makes them feel mildly uncomfortable.

.

.

.

.

Next step is to test the program in the inflatable dome, and move on to the Immersive Vision Theatre. 

So after editing the videos into an .avi format, I decided to shrink the size down while doing the demo program. When it comes to the dome, I will create high quality videos and sketches.

The next problem was trying to get three videos playing when triggered, without freezing or reloading. The sketch kept redrawing the videos so much, they couldn’t play. In the end I created a variable called wait_count to delay the feed.

As for the videos, the videos are fed into a playlist. Then called into the variable currently_playing, each when they fit the criteria set by the if statements. Here is my code for the processing patch:

After a couple of problems with video playback, I finally have a working versions of my program ‘Journey of a Diabetic’. An example of it working can be seen below:

.

.

Next is to create dome-corrected videos and test them first in the inflatable dome and finally in the Immersive Vision Theatre.

So when I came to replacing the background colours with videos, this was and still is the hardest bit of the program. When researching videos into processing, they use the format .mov in their examples. However when I rendered out my videos in this format, they either wouldnt play nor have very bad lack.

.

After a discuss with Simon Lock about my program, I thought it was my code that was wrong but it was in fact the video were in the wrong format. He showed me an example of a video player that triggered different video on different key presses. In this, he had used .mp4 and .avi, which played back perfectly like the example below shows:

.

.

So now I have to go back and re-render my videos into avi, as they play better than .mp4 and then try again to integrate them.

In the first version of my project, I started first getting the twitter feed into processing. Here is how the feed looks on the twitter page itself:

.

.

Then this is how the RRS Feed looks:

http://twitter.com/statuses/user_timeline/201393384.rss

.

As you can, it doesn’t just give you the integer numbers, instead the whole text status. So in processing I had to use a sub-string method to extract the numbers, and then convert them into an integer I could use to evaluate with.

The next part in the program was to write the argument, which would trigger the different videos. They were as the following:

  if (DiabeticNumber <= 4) {
        Playthelowvideo();
      }   

      if (DiabeticNumber > 4 && DiabeticNumber <= 7) {
        Playtheperfectvideo();
      }
      if (DiabeticNumber >= 8) {
        Playthehighvideo();
      }

.

The next part was to put a delay on the feed reading in, in order to get the videos time to play. This is very tricky as the sketch is constantly re-drawing itself, so I couldn’t use the sleep method as it would kill the whole program, including playing the videos. So instead I had to use delay, causing a problem that the program doesn’t start up for 8 seconds.

To test the program initially, I dynamically changed the background colour of the sketch to see if it would work. Here is a video of the program at work:

.

.

So now this part is working, time to implement the videos.

To understand how I might be able to extend this using real time data feeds, rather than tweeting them, I attended an workshop that taught us how to build and use arduino based objects known as Ecoids.

.

Luis Girao (http://www.artshare.com.pt/) ran the Remote Control Workshop on Wednesday 9th and continued onto Thursday 10th March. It was an optional workshop and we worked in small groups to develop the ecoid and work out the circuit board..The first day was a little slow, as we had a lot of introductions and presentations to watch before we got down to looking at the Ecoid themselves. We spent over 2 hours working out the circuit rather than building them. However by the end of the workshop, we had 2 working Ecoids and a few that needed to be tweaked. .The parts I found useful from this workshop was:

• The Xbee shields being able to transmit data across an area
• Pulling the data into processing and manipulating it (Similar to my 4D piece)
• Learning more about electronics and building circuits
• Learning about more about the different sensors available to us

.

After working on the program for a month, we managed to the get the Arduino to react to the sound of a car. As the car gets closer and closer, the light gets brighter until it is at full capacity. Then as the car gets further away, the light begins to dim until it turns off. 

.

We didn’t use the broker provided by IBM to talk to the system as the Arduino could cope with the system by itself. If we were to extend the project, we would indeed link it up to a server so people could monitor all the street lamp and make a smarter city.

Here are some photos of the street lamp at work:

.

.

So all we needed to do was to make a dummy street and put the Arduino into it. Unfortunately, my partner Jay went away the week before submission date without telling me and didn’t return till the day before. It’s a shame after all the work we put in for it, for it not to be seen.

However I did learn a lot about IBM and their initiatives, and how I could use their thinking in this project. They do their best to educate people about the world we live in, where I want to educate people more in this disease. I could ask more people to contribute to the twitter data feed and make this project more global and interactive, having it based not just in the dome, but online too.

.

Next is to get a working example of the program working in the inflatable dome.