Cycle 1: It Takes Two Magic Mirror
Posted: April 1, 2025 Filed under: Uncategorized | Tags: cycle 1, Interactove Media, Isadora, magic mirror Leave a comment »My project is a magic mirror of sorts that allows for interaction via an XBox One Kinect depth sensor. The project is called “It Takes Two”, because it takes two people to activate. In its single-user state, the background and user’s shadow are desaturated with a desaturation actor linked to the “Bodies” output of the OpenNI Tracker BETA actor. When the sensor only detects 1 body (via an Inside Range actor), it puts the saturation value at 0. When a second body is detected, it sets the saturation value at 100. I have utilized envelope generators to ensure a smooth fade in and fade out of saturation.
The above patch was added onto the shadow mechanism I created. I did some research on how to achieve this, and experimented with a few different actors before concluding that I needed an Alpha Mask. The LumaKey actor was one I played with briefly but it did not do what I needed. I found a tutorial by Mark Coniglio, which is how I ended up in alpha-land, and it worked beautifully. I still had to tinker with the specific settings within the OpenNI Tracker (and there is still more to be fine-tuned), but I had a functional shadow.
My goal with Cycle 1 was to establish the base for the rest of my project so I could continue building off it. I sectioned off my time to take full advantage of lab time to get the majority of my work done. I stuck to this schedule well and ended Cycle 1 in a good position, ready to take on Cycle 2. I gave myself a lot of time for troubleshooting and fine-tuning, which allowed me to work at a steady, low-stress pace.
I did not anticipate having so much difficulty finding colorscape videos that would provide texture and contrast without being overwhelming or distracting. I spent about 45 minutes of my time looking for videos and found a few. I also ended up reusing some video from Pressure Project 2 that worked nicely as a placeholder and resulted in some creative insight from a peer during presentations. I will have to continue searching for videos, and I am also considering creating colored backdrops and experimenting with background noise. I spent about 20 minutes of my time searching for a sound effect to play during the saturation of the media. I wanted a sound to draw the users’ attention to the changes that are happening.
Overall, the reactions from my peers were joyful. They were very curious to discover how my project worked (there was admittedly not much to discover at this point as I only have the base mechanisms done). They seemed excited to see the next iteration and had some helpful ideas for me. One idea was to lean into the ocean video I borrowed from PP2, which they recognized, causing them to expect a certain interaction to occur. I could have different background themes that have corresponding effects, such as a ripple effect on the ocean background. This would be a fun idea to play with for Cycle 2 or 3.
The other suggestions matched closely to my plans for the next cycles. I did not present on a projector because my project is so small at the moment, but they suggested a bigger display would better the experience (I agree). My goal is to devise a setup that would fit my project. In doing so, I need to keep in mind the robustness of my sensor. I needed a very plain background, as it liked to read parts of a busy background as a body, and occasionally refused to see a body. Currently, I think the white cyc in the MOLA would be my best bet because it is plain and flat.
The other major suggestion was to add more things to interact with. This is also part of my plan and I have a few ideas that I want to implement. These ‘easter eggs’, we’ll call them, will also be attached to a sound (likely the same magical shimmer). Part of the feedback I received is that the sound was a nice addition to the experience. Adding a sonic element helped extend the experience beyond my computer screen and immerse the user into the experience.
This is a screen recording I took, and it does a great job demonstrating some of the above issues. I included the outputs of the OpenNI Tracker actor specifically to show the body counter (the lowest number output). I am the only person in the sensor, but it is reading something behind me as a body, so I adjusted the sensor to get rid of that so I could demonstrate the desaturation. Because it saw the object behind me as a body, Isadora responded as such and saturated the image. The video also shows how the video resaturates breifly before desaturating when I step out and step back in, which is a result of the envelope generator. (The sound was not recording properly, please see above for sound sample.)
My score was my best friend during this project. I had it open any time I was working on this project. I ended up adding to it regularly throughout the process. It became a place where I collected my research via saved links and tracked my progress with screenshots of my Isadora stage. It helped me know where I was at with my progress so I knew what to work on next time I picked it up and how to pace myself across this cycle by itself and all three cycles together. I even used to to store ideas I had for this or a future cycle. I will continue to build on this document in future cycles, as it was incredibly helpful it keeping my work organized.