²

Reflective Statement

Over the course of this semester Tash and I have pursued quite a long line of enquiry to arrive at the point we are at now.

Initially we intended to pursue a multi-touch system capable of tracking both multiple finger inputs and fiducal markers, using this system we intended to introduce physical input devices such as dials that could be recognised and calibrated by the fiducal software on the fly. on the software end of this design we explored concepts such as cloud computing, virtual scrapbooking and the desktop metaphor. However after doing some more in-depth research it became painfully apparent that we would need an expensive short range projector for the setup we wanted. On top of this we found that every idea we were coming up with was already being explored simultaneously in different parts of the world. Rather than compromise we decided to follow our line of questioning into more rewarding territory.

Augmented reality was our general area of interest so we decided to look at a low-end solution that could be easily demonstrated. We started by looking at the marker based toolkit in processing but quickly realised it wasnt effective for what we wanted to do. We decided to use Flash in conjunction with Papervision 3D and a FLAR library because they are relatively well documented and easy to use (apart from some japanese commenting in FLARtoolkit).The final result was functional but slow due to the use of flash and much feedback data, along with multiple marker tracking. The reasons for the overdose of feedback is the use of a different transformation matrix between the single and multiple marker trackers. To avoid these problems in the final design we decided to track a a single marker and use it to control music. To do this I learned to use the internal sound class to play and pause a song depending on whether the marker was being tracked or not. An interesting side effect of this is marker_added is triggered more than marker_removed, resulting in the song being played over itself. The final design used a collada file designed in 3d studio max with textures from photoshop. aliasing created the necessity to hold the marker quite close to the webcam to make the album list playable but apart from that it was quite a robust solution.

The final design is meant to be an example of a bigger system. In theory if you are standing outside a gig or in the album racks at JB hifi you should be able to take a card (or a pxt) with a logo on it and the name of the album/band. By holding the card infront of the webcam and visiting a website you can hear a sample of the music streamed remotely with additional information shown in an interesting way. This idea was more effective in our minds than cloud computing portfolios because it used the cards to present someone elses information in a creative and new way.

The final result was successful in that it applied a theoretical idea in an effective and robust way, obviously the idea could have been pursued further (ad infinitum) but the fact that it worked well enough to clearly demonstrate the idea we were trying to present makes me happy with the final product. I am also happy with the teamwork and the clear process we went through to get a final product, even if it was hard won in the end!