13 Music generation with MuseGAN

Up to now, we have successfully generated shapes, numbers, images, and text. In this chapter and the next, we will explore two different ways of generating lifelike music. This chapter will apply the techniques from image GANs, treating a piece of music as a multidimensional object akin to an image. The generator will produce a complete piece of music and submit it to the critic (serving as the discriminator because we use the Wasserstein distance with gradient penalty, as discussed in chapter 5) for evaluation. The generator will then modify the music based on the critic’s feedback until it closely resembles real music from the training dataset. In the next chapter, we will treat music as a sequence of musical events, employing natural language processing (NLP) techniques. We will use a GPT-style Transformer to predict the most probable musical event in a sequence based on previous events. This Transformer will generate a long sequence of musical events that can be converted into realistic-sounding music.

13.1 Digital music representation

13.1.1 Musical notes, octave, and pitch

13.1.2 An introduction to multitrack music

13.1.3 Digitally represent music: Piano rolls

13.2 A blueprint for music generation

13.2.1 Constructing music with chords, style, melody, and groove

13.2.2 A blueprint to train a MuseGAN

13.3 Preparing the training data for MuseGAN

13.3.1 Downloading the training data

13.3.2 Converting multidimensional objects to music pieces

13.4 Building a MuseGAN

13.4.1 A critic in MuseGAN

13.4.2 A generator in MuseGAN

13.4.3 Optimizers and the loss function

13.5 Training the MuseGAN to generate music

13.5.1 Training the MuseGAN

13.5.2 Generating music with the trained MuseGAN

Summary