Who doesn't want to be a princess?
What it does
As you sing, we parse the lyrics of the song and match them to a Disney princess's song. The LEDs then take on the pattern of the dress of the princess you are singing.
How we built it
Magic and Fairy Dust!
Kidding. The LEDs are individually programmable Neopixels (by Adafruit), and what we've done to dressify them is map them into different regions on the body-- so we control bust, waist, and sections of the skirt. We can customize per princess following that, but we assign colors to each section once the princess has been identified.
To parse the song lyrics, we use an API called DeepGram.
Challenges we ran into
Customizing Elsa's dress, which doesn't follow standard princess dress configuration. We coded a gradient into the dress, by taking the initial RGB value and ending RGB value and having each R, G, and B increase proportionally. We discovered, however, that they can't linearly change or the colors in the middle of the dress will be on a completely different spectrum since the ratios are off.
Accomplishments that we're proud of
We SANG in front of the entire closing ceremonies at MHacks: Refactor. Don't believe me? ("Ask the dishes!" Kidding.) Check it out here. If that isn't every singing averse engineer's nightmare, I don't know what is.
Also, we built a dress that parsed out lyrics and displayed a princess dress in real time. None of us had experience with LEDs or lyric parsing before this project.
What we learned
40A of power wrapped around your body gets a little toasty. We also learned a lot about each other, our love for Disney tunes, and code! We worked heavily with the Edison trying to integrate it with the Arduino IDE, and although it didn't end up working, guess what group of girls knows about back end Linux now? We also learned about waveforms, phonetic parsing, and how much of a headache you can get from staring at LEDs too long.