23 Jul Machines Making Music

Today is an interesting time in the world of music.

Music has always been a powerful means of expression – throughout all of human history, and through many ground-breaking societies, music has been a way to engage and express and reflect the vibrancy of life. It reveals a lot about human thought and human character – and it also helps us to entertain, relax and enjoy life.

Now music is taking a huge step forward along with all of the other parts of our lives, as artificial intelligence and machine learning progress changes how we live. In the world of computing, engineers are discovering machine learning, which works by imitating the thought processes of the human brain.

How does this relate to music? New projects like Google’s NSynth are creating “computer-written” music – tunes that are made by both humans and computers, and some that are made without any real human involvement at all!

Music Machines: Built on Artificial Intelligence Advancements

In many ways, bringing automation and machine learning to music has evolved from other models that predated this kind of innovative musical work.

In the past few years, scientists have been taking huge strides in creating convolutional neural networks or CNNs. These CNNs perform sophisticated image processing and implement computer vision activities. Other types of neural networks have been able to replicate other types of learning tasks. For example, recurrent neural networks use artificial “memory” to help with tasks like analyzing speech.

Music is important, too! Now, engineers are deploying the power of machine learning and artificial intelligence to create new possibilities based on traditional music theory.

How it Works

In new types of music machine learning projects, engineers load existing sounds and samples. They may also provide clips of existing music to a data set known as a “training set.” Then the machine takes that training set, and it learns how to make its own music!

The result is prolific improvisation – in drumbeats, melodies, jazz and nearly any kind of music. For example, humans can work with machines in a ‘call and response’ program, where the computer bases its composition on what the human musician has given. A musician enters a sequence on a keyboard – and the computer “riffs” back.

Making Music Exciting

Will music machines make human musicians obsolete? That’s unlikely.

Instead, think of it as “jamming” with your computer. Many musicians have already experimented with the kinds of interesting studio recording that the personal computer era brought us. Now, they can actually riff with sentient technologies to produce new tunes and think outside the box when it comes to composition and songwriting.

Talk to Rochester Conservatory about starting a new musical career, or to learn about music, to get closer to some of these exciting opportunities. We help students of all ages start out on their journey toward a musical career or a life of fun and productive music-making. Ask about our learning programs and what we offer our customers to get closer to becoming a master musician.