Music generater with Denoising Auto encoder
Project name: An Odd Music Generator
the project has 4 main parts which should be done in a loop to generate the output song:
-
Denoising: During the performance, the musician must pay full attention to the sound of the music and must not allow noise and the sound of crowd cheering (!) destroys his concentration.
-
Note Recognition: During the performance, the musician must remember in his mind the series of notes he performed To know what notes he has hit in the performance of a piece up to every moment.
-
Note Prediction: During the performance, the musician must keep in mind the series of notes that are being performed up to the moment predict the next note to sound it.
-
Noise maker: Naturally, since the crowd in the hall is not covered in pleasure, during the performance classical music with hands and whistles encourage your music production.
the dataset contains three parts:
- notes
- noises
- songs (sequence of note names)
Neural Network models implemented by Keras in Tensorflow library.
- Amir Kasaei
- Amir Ezzati
This project was a pair work that we wrote as final project of Deep Learning course at University of Guilan in May 2022.