personal blog

GML Recording Machine @ Brussels

Monday 6 February 2012 - Filed under Uncategorized

I was in Brussels, Belgium for Verbindingen/Jonctions 13 to give a workshop on how to build the GML Recording Machine in collaboration with Evan Roth and Wendy Van Wynsberghe. As you might already know, I started this project with the intention to submit to GML Field Recorder Challenge but couldn’t manage to bring a decent documentation together for the challenge for this actually working device before the deadline. So I did not submit it to the challenge at all. But thanks to Constant for this workshop, there were plenty of people to help for the documentation and thanks to Evan for the great pictures that he took through the building process.

The purpose of the workshop was more creating a documentation for the device than building the device and it took us around 5 hours to build it from scratch. At the end it worked surprisingly in spite of our beliefs that “Electronics projects do not work at first attempt”.

Here is the video by Evan that you can see all the images and the video of the device in action this time in Brussels, (Last time was in Vienna, check it here) :

Here is the complete documentation that you can download in pdf format which is the official how to guide of the device that I finished a bit while after the workshop.

I’ll go through the steps briefly in this post anyway:

The parts that we used to make the device were mostly from Home Depot, and the PS2 ball mouse was from Ebay since it’s not easy to find a ball mouse in stores anymore. There are tape measures, washers, a snap fastener kit, 20ft/6m string, some steel wire, Arduino, Arduino microSD Shield, some resistors, a 9V battery, buttons and some regular tools to make stuff, such as pliers, screwdrivers, soldering iron and solder, Dremmel etc.


On the software side, we need Arduino, Arduino PS2 library to interface the Arduino with PS2 mouse, SDFatLib for interfacing with the Arduino microSD shield, and some programs that I wrote.

We started by modifying the tape measures, we removed the tape from inside and replaced it with a string for each of the measures:


Second and the most time consuming part was modifying the mouse and attach the tape measures and the mouse all together. We cut some parts inside the mouse, drilled holes to put the string through to turn the encoders inside the mouse, modified the discs of encoders, attached some washers to keep the strings on the mechanism. Here are some snapshots from the documentation but be sure to check out the pdf documentation for more details.

The mouse on the right will look like the mouse on the left at the end of this process

Getting rid of some extra parts from inside the mouse.

Cutting some parts of the encoder’s disc in order to decrease the resolution and slow it down since the mouse could not handle the fast rotation by the string.

Fixing some parts from snap fastener kit to smooth the edges of the plastic and  prevent the string cut the plastic as it goes back and forth.

Attaching the mouse and the tape measures all together.

Extending the buttons of the mouse with wires so we can use them as inputs for stroke and calibration.

Then we soldered everything on microSD shield, diagrams are on the pdf documentation, but it’s basically a combination of basic sketches like connecting an LED, connecting a button and tutorial of Arduino PS2 library.

We glued a button to the spray cap to capture the stroke.


Finally we made some modifications on our shoes using the steel wire to be able to attach the strings on our shoes:


And tape everything together make everything stable and colorful. Instead of using tape to attach everything together you can create a case and attach the case to the spray can to be able to replace the can easily.


Briefly, these are the steps, as I said before you can check the pdf documentation for more detailed description of each step.

And again, thanks to Constant and Evan Roth for inviting me to Verbindingen/Jonctions 13 for the workshop and all the participants attended to the workshop so we could document this hard work properly.



 ::  Share or discuss  ::  2012-02-06  ::  muharrem yildirim

GML Recording Machine @ Vienna

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 (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
- Arduino microSD Shield
- Retractable tape measure x 2
- Ball mouse
- Strings
- 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.


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:

GML Recording Machine from muharrem yildirim on Vimeo.

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.

2 comments  ::  Share or discuss  ::  2011-05-31  ::  muharrem yildirim