Music and AI final project
“Escape from the Turing trap”
Hello everyone! I am currently in France, and may not have access to wifi while the presentations are going on, but I want to say thank you for such an unforgettable quarter of AI and mayhem, and I hope you all have a lovely spring break!
My video submission will include both projects, feel free to skip through as needed. The first project is at the beginning and the second starts at [timestamp 2:11]!
My video demo: https://youtu.be/QzOnG5EEv78
The new project: LaughTrack
Previously Made Project: A Choral Audio Mosaic
Choosing between the two projects was a difficult task I must admit. I am both fond of and frustrated by everything I have done so far. I am proud of the little bits of good that came from them and frustrated by how much more I wanted to do. For example, the audio mosaic is fairly simple in its execution and I wanted to be able to provide more functionality to it initially and overall. Seeing how it turned out so well I want to continue with that idea and advance it to have more functionality for affecting the sound. It would be nice to continue it even further to build a program like it that harmonizes with you as you sing, making copies of your own voice like a vocoder, perhaps that analyzes the harmonic structures of the songs and uses that to create automatic harmonies in the style of whatever music you put in. Of course, time and stress did not allow for this to be done, but you can see my frustration at what I have achieved given the ideas I tend to aim for. The second project was a little more lighthearted for me, which was to create a program that laughs at your jokes which I dubbed, LaughTrack. I initially wanted it to be a program that determined when you told a joke based on the tone of voice, however, I had some difficulties as I realized that speaking for different words makes the reading of tone highly unpredictable even when said in a similar cadence. I think this still could be done using a more advanced program that is trained on the jokes in text and uses a speech-to-text as well as the actual mic input. However, for the sake of this class and this project, it was clear that limiting its input to face recognition would be the best choice.
The laugh track may be one of the more immediately entertaining projects that I have ever done. At first, I thought it was because laughter (even recorded) can be contagious. Then I thought it was because it seemed to have a mind of its own, laughing at inopportune and sometimes inappropriate times. But ultimately I realized it was entertaining because it was because of the way other people seemed to laugh right alongside it. Everyone I have told about this project or shown it to has enjoyed it or the concept and I am humbled by the joy that something so simple has been able to bring to me and others. For it to not only be an odd tool or toy, but entertainment that brings people together, that actually does get people to laugh (even indirectly) at my stupid jokes.
Final Project Milestone
Here is a link for my milestone video, showing my first check-in video in the development