-
Winning Scene
Inspiration
We get inspired by a very common scene in our everyday life. Sometimes when we listen to some fantastic music we would like to do some gestures to conduct it(especially for a symphony). So, we want to apply machine learning to the gesture information and generate the according music. Just imagine when you wave your hands and you will get the music which is very suitable for your current mood. (You can easily show your feelings from gestures.). No more switching and finding music. You are the music itself.
How it works
We use a sensor from our sponsors and connect it to Arduino. Then we get the sensors' information from Arduino to the computer. After this, We use hidden Markov model to inference the "real" note from the sensors information. Next, We use the note to compute the according rhythm, tune, chord, melody. At last, producing the music file.
My part
Brucefeynman:
I was in charge of the software part. I use HMM to get the according note from the gesture information, then use the FM synthesis to produce music and generate the music file.
demerzel1:
I was in charge of the software and hardware interface. I use Python's multithreading to transfer the data from our Arduino to the computer.
Difficulty
HMM and FM synthesis algorithm. I build everything from scratch except for the FM synthesis part.
Accomplishments that I'm proud of
It must be the first music we produce. That's so wonderful.
What I learned
Team work. Without my teammates, I won't be able to make it happen.
What's next for 3D Gesture Music.
Maybe an app :)
Built With
- ardiuno
- hmm
- ml
- python
