Generated Movements - Model Trained on Analytical Features
Songs From the Training Data
-
Chasing the Learning Rate
Video . BVH . Music . Beat Synchronization . Features Plot -
Inna
Video . BVH . Music . Beat Synchronization . Features Plot -
LikeMyJack
Video . BVH . Music . Beat Synchronization . Features Plot -
LM100 - Excerpt 1
Video . BVH . Music . Beat Synchronization . Features Plot -
Roucoule
Video . BVH . Music . Beat Synchronization . Features Plot
Songs Outside of the Training Data
-
LM97 - Excerpt 1
Video . BVH . Music . Beat Synchronization . Features Plot -
LM97 - Excerpt 2
Video . BVH . Music . Beat Synchronization . Features Plot -
PanSonic - Wanyugo
Video . BVH . Music . Beat Synchronization . Features Plot -
PenderStreetSteppers - BubbleWorld
Video . BVH . Music . Beat Synchronization . Features Plot -
Les Jours Tristes
Video . BVH . Music . Beat Synchronization . Features Plot -
Schubert - Serenade
Video . BVH . Music . Beat Synchronization . Features Plot
Generated Movements - Model Trained on WaveNet Embeddings
Songs From the Training Data
-
Chasing the Learning Rate
Video . BVH . Music . Beat Synchronization . Features Plot -
Inna
Video . BVH . Music . Beat Synchronization . Features Plot -
LikeMyJack
Video . BVH . Music . Beat Synchronization . Features Plot -
LM100 - Excerpt 1
Video . BVH . Music . Beat Synchronization . Features Plot -
Roucoule
Video . BVH . Music . Beat Synchronization . Features Plot
Songs Outside of the Training Data
-
LM97 - Excerpt 1
Video . BVH . Music . Beat Synchronization . Features Plot -
LM97 - Excerpt 2
Video . BVH . Music . Beat Synchronization . Features Plot -
PanSonic - Wanyugo
Video . BVH . Music . Beat Synchronization . Features Plot -
PenderStreetSteppers - BubbleWorld
Video . BVH . Music . Beat Synchronization . Features Plot -
Les Jours Tristes
Video . BVH . Music . Beat Synchronization . Features Plot -
Schubert - Serenade
Video . BVH . Music . Beat Synchronization . Features Plot
Real-Time Generation
Note: The model used in the real-time demo is different from the above models. It is trained on analytical audio features, but based on the data from a different performer. In the demo video, the first three songs are not in the training set.