Tiankai Li is a music producer, songwriter, and researcher in music cognition. A graduate of New York University’s Music Technology program, he has learned to combine multiple disciplines from art and science, namely music, audio, computer science, and psychology. He is looking to pursue his passion: through combined fields of study, he will try to understand how music is heard in people’s minds and use it to improve music recommendations for people who need it. He currently resides in China.
Projects
My Master Thesis “Emotional Method For Music Recommendation”
A study on the emotional impact of listening to music could be a good indicator for further assisting music recommendation algorithms. Through online surveys and qualitative interviews based on existing Music Emotion/Cognition literature, this investigation revealed modern music listening patterns with different levels of emotional engagement. It brought up essential suggestions for current music streaming algorithms to implement.
The file is an excerpt of my master’s thesis.
Deep Learning for Gesture-Based Music Control
In this project, I helped develop a deep learning-based system that integrates TensorFlow’s 3D hand pose detection model with a custom neural network to control musical instruments using hand gestures. My role explicitly is to adjust dense layers to improve the recognition accuracy and design the music instrument concept. The system processes real-time hand pose data to map gestures to instrument control, enabling dynamic interaction with the instrument’s pitch and loudness. With over 600,000 training samples, the model achieved high accuracy (88%) and recall rates (up to 99%) for detecting specific gestures. The project also featured innovative integrations with Max/MSP for real-time music performance and instrument design.
The training framework and design are shown above (left), the screenshot is displayed on the right, and a potential instrumental design is illustrated in Max MSP (bottom).
Max Patch Music Visual Project
I created a jitter visual in Max MSP for my Creative With Interactive Media class—a particle visual that syncs with the music. The patch operates with any music featuring a kick drum, which is detected and used to trigger the particle system in real time. Here’s a video of me demonstrating how the max patch works.
My Original Music
Here are samples of my original music, Forget Troubles (忘记烦恼), Just Perfect (完美的刚好), and Baby Please Baby, from my recent EP; you can check them out on Spotify if you like! My NetEase Music (网易云音乐) account is 骨碌; check it out if you like!
Follow My Blog!!
Made with w.link
Tiankai Li 