Tuesday 31 May 2011 - Filed under Uncategorized
GML Recording Machine is an ongoing project to build a low cost hardware and software system to record graffiti motion data during graffiti writer’s tag writing process in the city. I started to this project after the announcement of GML Field Recorder Challenge and with the motivation to contribute the GML community. This post will briefly describe the project and present a demonstration video of the device in streets of Vienna.
GML RECORDING MACHINE
GML (Graffiti Markup Language) Recording Machine is a DIY device that is capable of capturing and writing a spray can’s movement to a microSD card. With two separate strings attached to each foot of the graffiti writer -writer needs to keep his/her feet still in this case- (also can be attached to the wall, to keep the writer’s feet free), position of the spray can is tracked during the tag writing process. These string’s length are measured by a simple ball mouse where these strings turns the X and Y rollers inside the mouse and the strings are retracted by two retractable tape measures. These measurements are processed by an Arduino and stored in an SD card to be converted to GML later on with a simple script. This device can be constructed using and modifying everyday found materials such as:
- Arduino microSD Shield
- Retractable tape measure x 2
- Ball mouse
- Water tube to attach the spray can and mount the mechanism inside
- A nut and a bolt (too keep the can attached)
For the ones asking why didn’t I just use an accelerometer and a gyroscope:
I tried to use an accelerometer in the beginning. The data was not that clean and I had some problems with processing the accelerometer output. After discussing with some friends, I decided to find a mechanical solution just to see if I can find something that will work properly. It was of course a bit more challenging but much more exciting to work on. Creating a small and compact enough system that will consists of the all required elements was not easy. For now, the last (and the third) prototype is the one that you can see here is the best I could so far (visit here to see the second prototype):
I will try to release a howto guide soon for the ones who would like to understand how the mechanism works.
The solution that I came up with, has some advantages and disadvantages. This seems the first system that can record GML with multistroke with the help of the button attached to the spray cap. And also writer do not need to carry a computer with himself/herself. There is not much time (actually not at all) needed for calibration and setting up the device. So it is easy to run away right after drawing the tag.
As a disadvantage, graffiti writer cannot move his/her feet during graffiti writing if the strings are attached to the feet. Another problem about the mechanism is that the can must be hold vertical to the ground to have a clean data since the mechanism actually measures how far the spray can’s “bottom” away from feet of the writer. If the can is not hold properly, data capturing goes wrong ending up with partly distorted tags.
EXHIBITION & THE REAL ACTION ON THE STREETS
First public experiment with this device has been done during the Playface Intercult exhibition which was organized in collaboration with amberPlatform from Istanbul, Turkey and Interface Culture Programme of Linz University of Arts Department of Media, Linz Austria. During the exhibition I used an XBee to make the communication between the computer and the device. For visualizing the captured GML data, i used GLM4U library for Processing. Thanks to Jerome for his help for providing me the sketch that animates the drawing of GML nicely.
I went to visit the exhibition to Vienna for the last days of the exhibition and with help of Evan, had the chance to met with an anonymous graffiti writer there.
The following video (thanks to DRAWVOLUTION) shows some real action done on the streets of Vienna with GML Recording Machine in right hands:
Tags in the video are visualized by Graffiti Analysis 2.0. For some reason, some parts of the tags (usually stroke beginnings and endings) do not look good. However, you can see the raw png output below. You can see that we actually captured the data but could not visualized it properly using GA2.0. The PNG files are generated using GMLImageRenderer, nothing fancy. You can also see the GA2.0 output next to the raw PNG files. You can download GML data and the PNG’s here.
In the demonstration video, data is captured to the computer for debugging purposes, since that was my last night in Vienna and we wouldn’t want to risk not being able to see if something goes wrong or not with the machine even though there is an LED indicator on the device.
2011-05-31 » muharrem yildirim