AI and music

Creative artificial intelligence can be the result of training neural nets to generate content based on the analysis of a large number of pieces of music. If the computer program is taught the rules of a certain tonal system it can rearrange sounds and create music. Music theory is based on rules which structure the simultaneous or the successive sound of frequencies. It is possible to determine statistical values for the coordination of sounds and the arrangement of scales.

Tones or harmonies can be arranged according to logical concepts based on axioms and theorems. The program can also compose a piece of music applying reinforcement learning. Recurrent neural networks (RNN) can use musical phrases as input and generate an output in form of a sequence, a variation, a diminuition or an augmentation. Time plays an important role concerning the programming of this network.

Artificial intelligence can be applied to the realm of digital sound processing. Pitch and time of audio files can be formalized and the sound can be optimized, edited and corrected.

AI is appropriate for the nature of music which is based on numbers. Psychoacoustic research shapes the development of the technology of music. The effect of the connection of certain harmonies and sequences of tones on the listener is predictable.


No comments:

Post a Comment

Note: Only a member of this blog may post a comment.