“Hi, I’m Oscar.
I have been taken from the streets of Montmartre, Paris to draw your portrait.
thanks to my keen eye and an excellent memory, you don’t even have to be around while I draw. You can just come back and collect your portrait when I’m finished.
More high-res images can be found here: http://sdrv.ms/I6tjN7
We have come to the end of the Oscar project. While we aimed for the stars, we landed somewhere on a cloud, much closer to earth. The process however has been full of learnings and we have created Oscar who truly has shown us that he has his own aesthetics and own ideas despite us wanting to control him.
Final changes and improvements
Because of the challenges and tangents experienced during various iterations in both coding and construction, we have only reached the stage in which Oscar can draw an image line by line rather than more free-flowing movement. The only interaction we were able to successfully achieve has been recognition of person in frame, facial recognition and facial capture. We were not able to achieve distance to speed of drawing interactive element. We tested many different drawing methods, image capture methods and interaction elements, however we found through user testing that often the simplest functionality worked most reliably and could complete an image as well as have it a recognisable image. Hence, we often reverted back to previous code with our focus being on optimising the basics rather than adding in much new functionality.
Our latest Oscar iteration is the most stable yet. It can complete a whole drawing and the portrait is somewhat recognisable to the user. We have incorporated the Kinect to create a level of interaction, where it responds to a user walking up to it and “performs” some sizing up of the user (something similar to how an artist would hold out a pencil to obtain proportions of an objects). We have also added in the drawing of a frame if a person is not in the Kinect’s realm of vision. This was particularly important as we thought that in an exhibition space a still object would not attract as much attention as a moving one.
We created some fun cable management systems by using the measuring tape retraction mechanism to hide surplus wires. We have found this to work very well. Below you can see our initial crude implementation of it, we have neatened it’s appearance.
Initially we wanted to hide all technological aspects of Oscar under wooden panels, however what we found is that the very mechanisms and movement were of interest to users and so we made a decision to leave it exposed.
To be consistent with the aesthetics of Oscar, our poster for the end of year student exhibition will be a series of real works by Oscar which are framed and displayed immediately above the installation. We believe this is more relevant to our style of drawing machine rather than creating a poster. See below for a photo on how we will arrange these artworks. We have also added in some commentary to support user interaction.
In the future, we hope to enhance the skill with which Oscar draws and processes the image. Some ideas would use frame differencing and tracing lines, or using random walkers to create a more random, humanlike, scribble drawing.
Critical evaluation of design process
Thinking about the process, result and experience critically, we have extracted the following learning.
1. Have a working physical prototype sooner would be valuable so we could get started on the coding part earlier and do more testing, enhancements and refining. I’ve often wondered how valuable our focus on the aesthetics is in an interaction course. Perhaps a partnership with more aesthetics/building-focused students would be valuable so that more of the team could focus on the code and improve that part. Conversely, what we create is a play between technical and aesthetic/physical, so both are important. It’s a difficult balance to strike.
2. Reality is hard to achieve through drawing aesthetics, and gets in the way of artistic endeavours and creative thinking. There were times when we felt that creating Oscar was like creating a printer. Thinking back on the initial ideas now, we would’ve opted for more abstract aesthetics. It is also difficult to measure up to reality when the picture being drawn is so obviously trying to replicate it, hence it adds an additional layer of judgement from the audience.
3. User testing was difficult and slow due to the many challenges we had with code and physical machine. This machine was unlike other projects. There was something to change literally after each user test. The process was very slow, but a lot was learnt. Because we focused so much on the technology and build of Oscar due to the challenges we had, we didn’t question the users to the degree we would’ve wanted to. As such, we may have missed opportunities for learning or may have gone down incorrect tangents. At the end of the day, we are making this drawing machine for users and should listen more to create a device which they want.
4. Rely on easy to obtain standards. Our starting point to the project build was the paper size and the paper was BIG. It affected everything. We could not get timer belts which were long enough, we had to custom build tracks and axes. Because everything was custom built it took time. For future projects, we would consider using more standardised elements to alleviate design and build pressures so that more time could be focused on developing amazing interactions.
5. It is enormously helpful to have the university supply the required equipment. Not having the financial limitation on what we could buy, it was of huge benefit that we could stretch our imaginations with equipment which the university already had: motors, power supplies, Open Beam, timer belts etc. We very much appreciated this support towards our learning.
Complete design process can be followed in the posts below:
Paul HOC – Drawing Machine https://idea9202studio.wordpress.com/2013/09/12/paul-hoc-drawing-machine/
Phase 1 – Individual Ideation and Grounding Research https://idea9202studio.wordpress.com/2013/09/08/phase-1-individual-ideation/
Phase 2 – Concept Proposal – Oscar https://idea9202studio.wordpress.com/2013/09/15/phase-2-oscar/
Oscar – Phase 3 presentation https://idea9202studio.wordpress.com/2013/10/13/oscar-phase-3-presentation/
Oscar – Getting ready for user testing https://idea9202studio.wordpress.com/2013/10/27/oscar-getting-ready-for-user-testing/
Image Selection & Processing https://idea9202studio.wordpress.com/2013/10/04/image-selection-processing-for-drawing/
Oscar – Phase 3 – Technical Journey driven by User Testing https://idea9202studio.wordpress.com/2013/11/17/oscar-phase-3-technical-journey-driven-by-user-testing/
Oscar – Phase 4 Video https://idea9202studio.wordpress.com/2013/11/15/oscar-phase-4-video/
Oscar – Final Documentation https://idea9202studio.wordpress.com/2013/11/17/oscar-final-documentation/